Molecular genetic predictors of

progression after radical

Simon RJ Bott MB.BS, FRCS (Eng).

The Research Centre,

University College London.

Dissertation for the degree of Doctor of Medicine

University of London, 2005. UMI Number: U592749

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion.

Dissertation Publishing

UMI U592749 Published by ProQuest LLC 2013. Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 Acknowledgements

I would like to thank the following people for providing support, inspiration and guidance in the generation o f this thesis.

Professor Roger Kirby, Professor of Urology, St George’s Hospital, London.

Dr M Constance Parkinson, Consultant Histopathologist, UCL Hospitals Trust,

London.

Dr Joel Hooper, Senior Scientist, The Genome Centre, William Harvey Research

Institute, Queen Mary's School of Medicine, London.

Dr Charles Mein, Laboratory Manager, The Genome Centre, William Harvey

Research Institute, Queen Mary's School of Medicine, London.

Dr Chris Jones, Senior Lecturer, Chester Beatty Research Laboratories, London.

Dr Garreth Griffiths, Senior Medical Statistician, MRC Clinical Trials Unit, London.

Dr David Hudson, Lead Scientist, Prostate Stem Cell Laboratory, Institute of Cancer

Research, Surrey.

Dr Magali Williamson, Senior Lecturer, Prostate Centre, London.

Professor John Masters, Professor of Cellular , Prostate Cancer Research

Centre, London.

And thanks also for their financial support for this project.

British Urological Foundation and AstraZeneca.

The Prostate Research Campaign UK.

Everard Goodman.

2 To Emma, Freddie and Oscar

3 Contents

Abstract

Chapter 1 Introduction

1.1 Introduction

1.2 Incidence of prostate cancer

1.3 Prostate anatomy and physiology

1.4 Natural history of prostate cancer and active surveillance

1.5 Staging prostate cancer

1.6 prostate cancer

a) Gleason grade b) Tumour multifocality

1.7 Localised prostate cancer a) Presentation b) Patient selection for radical prostatectomy c) Preoperative predictors of outcome 1. Clinical stage 2. PSA 3. Gleason score 4. Biopsy features 5. Preoperative imaging 6. Circulating prostate cells 7. PSA density, PSA ffeeitotal ratio, PSMA

1.8 Post-operative predictors of outcome a) Radical prostatectomy Gleason score b) Pathological stage c) Positive surgical margins d) Tumour volume e) Lymphovascular f) DNA ploidy

1.9 Neoadjuvant androgen deprivation

1.10 Management of recurrent disease after radical prostatectomy a) Definition and timing of relapse b) Site of relapse c) Treatment of relapse

4 1.11 Molecular changes in prostate cancer a) Background 1. 2. Tumour suppressor genes b) Hereditary prostate cancer 1. Candidate genes 2. Low penetrance polymorphisms

c) Somatic changes 1. Early prostate cancer 2. Metastatic prostate cancer 3. Hormone refractory prostate cancer 4. Telomerase

1.12 Molecular markers in cancer a) DNA markers and Microarray technology b) Allelic imbalance c) Microsatellite instability d) Hypermethylation e) Mitochondrial DNA mutation f) Viral DNA g) RNA markers h) Protein markers

1.11 Summary

Chapter 2 Patient selection

Aim

2.1 Introduction

2.2 Methods a) The UCL Radical Prostatectomy Database b) Consent c) Identifying two groups of patient 1. Tumour stage 2. Exclsuion criteria 3. Biochemical outcome 4. Matching 2.3 Results a) The Radical Prostatectomy Database b) Patient matching c) Comparison between the groups d) The unmatched patients

2.4 Discussion

2.5 Summary

5 Chapter 3 Materials and Methods

Aim

3.1 The pathological specimen and tumour dissection

3.2 DNA extraction

3.3 DNA amplification a) Principles of the polymerase chain reaction b) Microsatellite marker selection c) PCR optimisation 1. DNA concentration 2. Marker concentration 3. Magnesium chloride concentration 3.4 Genotyping

3.5 Analysing the data

Chapter 4 Results

4.1 Allelic imbalance

4.2 Results a) Interpreting electropherograms b) Allelic imbalance and PSA outcome

4.3 Discussion

4.4 Study limitations

Chapter 5 The p21WAF1/CIP1 gene is inactivated in metastatic prostatic lines by promoter methylation

Aim

5.1 Introduction

5.2 Materials and Methods a) Cell culture and treatment b) Cell counting c) p 2i WAFI/CIPI gene expression 1. RNA extraction 2. cDNA 3. Primer design 4. Reverse Transcriptase-Polymerase Chain Reaction 5. Agarose gel electrophoresis

6 d) Promoter sequencing 1. DNA extraction 2. Sodium bisulphite modification 3. PCR 4. DNA extraction from agarose gel 5. Sequencing reaction 5.3 Results a) Cell culture b) RT-PCR c) Sequencing the p21HAF 1CIPI promoter

5.4 Discussion

5.5 Conclusions

Chapter 6 Conclusions and future directions

Appendix

Appendix 1. Power calculation: based on McNemar’s test Appendix 2. Preoperative features of matched patients Appendix 3. Radical prostatectomy pathological findings in matched patients Appendix 4. PSA follow-up in matched patients Appendix 5. Procedure for staining with haematoxylin and eosin Appendix 6. The ABgene™ DNA block 96well plates arranged for genotyping Appendix 7. Features of the polymorphic markers used in the allelic imbalance study

References

Abbreviations

7 Tables and Figures

Table 1.1. Outcome of patients undergoing active surveillance. Table 1.2. TNM classification of prostate cancer. Table 1.3. Comparison of body coil and endorectal coil magnetic resonance imaging at detecting extracapsular extension of prostatic cancer. Table 1.4. Risk factors for recurrence after radical prostatectomy. Table 1.5. Actuarial biochemical free survival and prostate tumour stage. Table 1.6. The site of non-iatrogenic (extraprostatic) solitary positive surgical margins after radical prostatectomy. Table 1.7. Biochemical recurrence and positive surgical margins. Table 1.8. Comparison of Gleason score, pathological stage and timing of PSA recurrence by site of recurrence after anatomic radical prostatectomy. Table 1.9. Chromosomal loci reported to contain hereditary prostate cancer genes. Table 2.1. Preoperative and pathological features of the UCL Radical Prostatectomy database. Table 2.2. The number and site of positive surgical margins from The UCL Radical Prostatectomy database. Table 2.3. Patient matching. Table 2.4. The site of solitary positive surgical margins. Table 2.5. Patients with 0, 1 or >1 positive surgical margins. Table 2.6. The stage, grade and preoperative PSA group for the unmatched patients. Table 3.1. Details of microsatellite markers. Table 4.1. Overall results - all patients. Table 4.2. Overall results - relapse group. Table 4.3. Overall results - non-relapse group. Table 4.4. Summary: rate of AI in relapse and non-relapse groups. Table 4.5. Status of 30 microsatellite loci in 31 matched pairs of radical prostatectomy patients with pT3N0 stage disease. Table 4.6. Incidence of microsatellite instability in the relapse and non-relapse groups. Table 5.1. Intragenic primers for p21 WAF1/C1PI and GAPDH. Table 5.2. PCR master mix for lx 50pl reaction. Table 5.3. p21 WAFI/CIPI primers.

8 Table 5.4. Results o f p 2 JlVAf,CIPI promoter sequencing after sodium bisulphite DNA modification.

Figure 1.1. Prostate cancer age-specific incidence in Great Britain (1979-2001). Figure 1.2. Genetic changes underlying the initiation and progression of prostate cancer. Figure 1.3. The of prostate cancer. Figure 1.4 a) Low power of radical prostatectomy wholemount section demonstrating organ confined prostate cancer, b) an intraprostatic positive surgical margin, c) extraprostatic with a positive surgical margin. Figure 1.5. Time to develop PSA recurrence (PSA >2) after adjuvant radiotherapy and salvage radiotherapy with a pre-treatment PSA <1 .Ong/ml and >1 .Ong/ml. Figure 1.6. The androgen receptor ligands and androgen responsive element. Figure 1.7. Demonstrates locus ‘a’ in a normal cell and a malignant cell containing microsatellite instability and allelic imbalance (LOH). Figure 2.1. Pathological stage over time in the UCL Radical Prostatectomy database. Figure 2.2. Assessment of margin status using the shave technique. Figure 3.1 a) Coronal block of prostate embedded in paraffin, b) master slide with area of prostate containing at least 70% malignant cells marked, c) 10pm section unstained, d) section stained, tumour area marked and excised with scalpel blade. Figure 3.2. Laser capture microdissection a) area of tumour marked, b) tumour captured on polymerised film, c) defect left on slide. Figure 3.3. Agarose gel of primer D7S523 at 2.5, 5 and 8 picomoles per 15pl reaction in DNA from unmatched patients. Figure 3.4. Agarose gel showing optimisation with DNA from 8 unmatched patients at marker D10S211. Figure 3.4. The electrophoresis circuit. Figure 3.5. Allelic imbalance is present where the amplitude of the smaller cancer peak divided by the larger cancer peak over the amplitude of the smaller control peak divided by the larger control peak is 0.7 or less. Figure 4.1. Shows characteristic cancer and control electropherograms for 5 markers, demonstrating stutter peaks preceding the allelic peak in a., b. and c.,

9 heterozygotes in a. and c.. a homozygote (uninformative) in b., allelic imbalance in d, and microsatellite instability in e. Figure 4.2. Electropherograms from patients 39 and 32. Patient 39 is homozygous at the D13S263 locus. Two stutter peaks precede the alleles. This appearance can be used to confirm that patient 32 is heterozygous at this location as there are four peaks, and the cancer sample shows allelic imbalance. Figure 5.1. Mechanism of DNA methylation. Figure 5.2. Sodium bisulphite treatment of unmethylated and methylated DNA. Figure 5.4 a) PC3 cells after 24 hours treatment with 5-Aza-CdR and 5 days growth in standard media, b) PC3 cells after 24 hours treatment with PBS and 5 days growth in standard media. Figure 5.5. Agarose gel showing RT-PCR of 5-Aza-CdR treated (+) and untreated (-) cell line DNA. Figure 5.6 shows part of the p21HAFI CIPI promoter region in modified PC3 cells. The cytosine residue is conserved at position -896 (a), -856 (b) and at the SIE-1 binding site at -692 (c), the latter shown here using the antisense primer. Figure 5.7. Positive and negative controls: a) Preservation of the cytosine residue in RD at -896 and b) conversion to thymine in NP.

10 ABSTRACT

Objectives

To identify genetic markers to predict biochemical outcome after radical prostatectomy and in a separate study to assess methylation, an epigenetic change, in metastatic prostate cancer cell lines.

Materials and Methods

One group of patients with PSA relapse and another non-relapse group were matched.

DNA was extracted from the radical prostatectomy specimens, amplified using thirty microsatellite markers and genotyping performed.

PC3, LNCaP and DU145 metastatic prostate cancer cell lines and control cell lines were cultured in the demethylating agent 5-Aza-2 deoxycytidine (5AzaCdR), in parallel with untreated cells. p21WAFI/CIP1 mRNA expression was analysed by RT-PCR. DNA from untreated cell lines was modified with sodium bisulphite and p2JHAf'ICIPI promoter sequencing performed.

Results

Allelic imbalance was significantly more common in the relapse than the non-relapse group at 10pl2.1(D10S211) (35% vrs 5%, p=0.03). Several markers showed a marked but not statistically significant difference between the two groups including: D1S158,

D1S422, D2S222, D5S500, D6S314, D8S136, NEFL, D11S2000, D12S89, D12S1697,

D13S269, D15S1232, D16S413 and D18S541.

RT-PCR demonstrated p 2 jnAf’1/CIPI was expressed at low or undetectable levels in metastatic prostate cancer cell lines but was reactivated by treatment with 5AzaCdR.

Sequence analysis of the promoter revealed several sites of methylation at the 5’ end of

11 a CpG island in the PC3, LNCaP and DU 145 cancer cell line DNA, including at a

STAT1 binding site.

Conclusions

A significant association between loss of a locus at chromosome 1 Opl2.1 and biochemical progression after radical prostatectomy was found. p21'y iW4FI CIPI expression • was low i in metastatic • prostate cancer cell If lines, i* but 1 was enhanced by demethylation. Several cytosine residues in the promoter region were methylated, in particular where STAT1 binds. The inhibition of the STAT1 signalling pathway by methylation of thep21nAFI CIPI promoter may inactivate the p21HAFI CIPI tumour suppressor gene in prostate cancer.

12 CHAPTER 1

1.1 Introduction

Prostate cancer is an enigma. This makes it a fascinating disease to study but a controversial and often difficult disease to treat. While still localised (confined to prostate), prostate cancer is curable by radical , radical external beam radiotherapy, brachytherapy or cryotherapy. Once the tumour has extended beyond the confines of the gland the chances of cure fall dramatically. For a patient with metastases curative therapy is not currently available and he has a median life expectancy of 18 months. Yet, the majority of prostate endemic in the male population will not cause any symptoms and will have no effect on the quality or quantity of that individual’s life. These tumours are termed ‘insignificant’ prostate cancers. To treat these patients would expose them to the inherent side effects and complications associated with radical therapies with no benefit in terms of their life expectancy.

To avoid treating patients inappropriately the decision to offer a patient radical surgery is based on the serum prostate specific antigen (PSA) level, a digital rectal examination

(DRE) of the prostate, and the results of prostatic biopsy sometimes augmented by radiological or radio-isotopic imaging. Even when the results of these investigations are summated, between a quarter and a half of patients undergoing radical prostatectomy have disease outside the prostate (Han et al. 2001; Partin et al. 1993). Yet, by no means all of these men will develop disease progression.

Currently the pathologist provides the best prognostic indicators by examining the radical prostatectomy surgical specimen and reporting the grade, stage and the status of the resection margins. However, in recent years attention has focused on newer

13 molecular markers, potentially applicable to prostate , to identify patients who are best suited for radical treatment.

14 1.2 Incidence of prostate cancer

The overall incidence of prostate cancer is rising in the UK. Europe and the United

States, such that prostate cancer is now the most commonly diagnosed life-threatening malignancy in men. In the UK, the overall incidence is currently 21,400 new cases per annum, with a rate of 74 cases per 100,000 men (National Statistics- http://www.statistics.gov.uk/ststbase/datasets2.asp). The age-matched incidence gives a better indication of the true incidence within an ageing population as it gives the number of sufferers at different age intervals per 100,000 head of population. It is not affected by the population size - so for prostate cancer although the overall incidence of the disease is rising rapidly, as more men are living longer, the age-matched incidence is t V l rising more slowly (Figure 1.1). > •

Figure 1.1. Prostate cancer age-specific incidence in Great Britain (1979-2001). 1200

1000 - c o co K 800 - © Q l o — 55-64 <=>. 600 - 65-74 o o 75-84

& 400 - — 85+ © CO QL 200 -

1975 1978 1981 1984 1987 1990 1993 1996 1999 Year of diagnosis

15 The age-matched incidence in the United States is over two fold higher than in the UK. despite similar mortality rates - this probably reflects the greater use of PSA as a screening tool in America. Stamey published his seminal paper on the potential use of

PSA in the diagnosis and management of prostate cancer in 1987 (Stamey et al. 1987).

The incidence rates in the US subsequently showed a sharp increase in 1990 and then a decline in 1993 until settling in 1995. to a level substantially higher than the 1986 level

(Shibata & Whittemore 2001). The UK witnessed a slower increase in age-matched incidence up to 1995. The incidence in men under 75 has continued to rise since 1995.

Whereas, in older men the incidence has levelled off or fallen, presumably as more men are diagnosed at a younger age.

The mortality rates have shown a downward trend in both countries since the mid

1990‘s, although the decrease is more dramatic in the US. Proponents of PSA screening have claimed these improved mortality rates reflect the rise in serum screening. Indeed the US and to a lesser extent the UK and Europe have seen a stage migration at diagnosis, where now a greater proportion of men are presenting with organ-confined disease (Bott et al. 2005c; Freedland et al. 2003; Han et al. 2001; Quinn et al. 2001).

However, for an affect on mortality there would be an expected lead-time of 10-20 years - i.e. we would expect to see a reduction in mortality over the next few years, rather than a decade ago. Furthermore the reduction in mortality witnessed across the

USA has been the same in regions of high and low levels of PSA screening (Brawley

1997). While PSA testing may play some role, improvements in treatment or some othei^ factor must also be considered (Shibata & Whittemore 2001). — (. ^

Autopsy and cystoprostatectomy studies looking at the incidence of clinically insignificant or latent prostate cancer in the population have revealed 30-50% of males

16 bom in the western world have a life-time risk of developing prostate cancer (Franks

1956; Villers et al. 1992). This figure varies depending on whether random or systematic whole-mount sections are examined (Lundberg & Berge 1970). When the entire gland is examined using whole-mount step sections at 4-5 mm intervals, foci of cancer are more commonly found and its prevalence increases with age, reaching 80% in men over 90 years old (Hirst & Bergman 1954). Hirst and Bergman showed this trend with advancing age was reflected in the incidence of clinical disease and mortality. They suggested there is a ‘latent* period of around 20 years during which such clinically unsuspected tumours progress and metastasise. It is estimated over an individuals lifetime 10% of men will develop symptoms and 3% will succumb to prostate cancer (Scardino, Weaver, & Hudson 1992).

17 1.3 Prostate anatomy and physiology

An understanding of the anatomy of the prostate is essential to the understanding of its

disease processes. McNeal defined the anatomical regions of the prostate: the peripheral

zone, the transition zone, the central zone and the anterior fibromuscular zone (McNeal

1972).-^

In a 20 year old male the peripheral zone makes up 75% of the glandular prostate and

forms the posterior, lateral and apical regions. It is composed of small acini, within a

fine fibro-muscular stroma, whose ducts open into the urethra at and distal to the veru

montanum. The peripheral zone is the site % o f prostatic cancers and the majority

occur posteriorly, this makes digital rectal examination important in the assessment of a

man with suspected prostate cancer. However, small posterior tumours may not be

palpable and 20% of prostate cancers occur anterior to the urethra (Bott et al. 2002),

posing difficulty not only in clinical staging but also achieving an adequate biopsy of

the anterior gland (Bott et al. 2002; Stamey et al. 1987).

The transition zone comprises 5% of the prostatic glandular tissue and in young men is

an inverted cone shape encircling the urethra with its apex at the veru montanum. In

older men with it may appear as lobes impinging on the prostatic urethra.

The acini of the transition zone are similar to those of the peripheral zone, and are

embedded in a dense stroma - more like the central zone. The ducts run parallel to the

urethra, into which they open at the distal margin of the smooth muscle sphincter.

Contained within the transition zone are the periurethral glands and their ducts open

directly into the urethra. The transition zone is the site of 20% jbf prostate cancers and

the site of benign prostatic hyperplasia development. Tumours arising in the transition

zone have distinct pathological features, which may reflect differences in their genotype. Transition zone tumours are reported to have a lower incidence of extracapsular and metastatic spread and to be of lower Gleason grade than peripheral zone tumours (Grignon & Sakr 1994; Noguchi et al. 2000). In contrast, they tend have greater tumour volume and have high PSA production when still confined to the prostate (Stamey. Dietrick. & Issa 1993). Several workers have reported significant differences in biochemical free survival in favour of transition zone tumours

(Erbersdobler et al. 2002b; Noguchi et al. 2000). though the question to what extent genetic variation between cancers arising from the different zones or the anatomical location dictates tumour behaviour remains to be answered .

The central zone, representing 25% of the glandular prostate, is an inverted pyramid, which sourrounds the ejaculatory ducts. Its base forms the prostatic base and its apex lies at the verumontanum. The acini are large and square in outline with marked infoldings and a sourrounding compact muscle component. The ducts of the vasa deferentia and enter the central zone and coalesce to form the ejaculatory ducts, which open into the urethra at the verumontanum. Less than 10% of prostatic originate in the central zone (McNeal et al. 1988).

The anterior fibromuscular stroma accounts for 30% of the prostate mass and forms the entire anterior surface of the prostate, extending as a continuum from the detrusor muscle of the bladder to the striated sphincter distal to the prostatic apex. It merges laterally with the ‘capsule’ and the glandular prostate is fused with its posterior surface.

This stroma was thought to be void of glands, however recently Tiguert et al. identified prostate acini within it (Tiguert et al. 1999). These acini have at least a theoretical risk of undergoing malignant transformation.

19 The prostate is described as having a ‘prostatic capsule’. This is important as tumour penetration through the capsule is one of the most fundamental prognostic features.

However, in reality this is a 2-3mm condensation of fibrous tissue with occasional muscle fibres, which separates the prostatic glandular tissue from the periprostatic connective tissue. It not a true capsule (like the renal capsule) as it is incomplete and is an intrinsic and inseparable part of the gland (Ayala et al. 1989). Although, where present, this condensation of fibrous tissue may offer an anatomical barrier against extraprostatic extension of tumour, it is absent at the prostatic apex. Here there is little or no barrier against local spread to adjacent structures.

The majority of tumours arise in the peripheral zone at the prostatic apex (Bott et al.

2002; Byar & Mostofi 1972). Byar and Mostofi showed that tumour was present in one or two of the apical sections in 75% of 208 men undergoing radical prostatectomy (Byar

& Mostofi 1972). The short inferior neurovascular pedicles enter the prostate at the apex and, as prostate cancer has a propensity to invade along the tissue planes surrounding nerves, this is a common site for extraprostatic disease. Furthermore, the surgical dissection at the apex is limited by the proximity of the external urethral sphincter and the neurovascular bundles, rendering the apex the most common site for positive surgical margins (Blute et al. 1997; Bott & Kirby 2002; Obek et al. 1999).

These observations suggest that the location of malignancy within the prostate may influence its presentation, its pathological behaviour and its treatment.

The function of the prostate is not fully understood. Secretions from the sexual accessory glands constitute 99% of the ejaculate, with 1% being spermatozoa. The ejaculatory fluid is produced by the seminal vesicles, the prostate and Cowper’s glands.

20 The prostatic contribution consists of citrate, zinc, polyamines (spermine, spermidine and putrescine) as well as antioxidants (superoxide dismutase) and prostate specific antigen (PSA) (Narayan et al. 2000). Citrate is thought to act as a buffer and as an energy source. Zinc is found in high levels in seminal fluid and may act as an inhibitor of citrate metabolism. In benign prostatic hyperplasia zinc levels are increased, while in cancer zinc levels are substantially lower. Low levels of zinc permit increased metabolism of citrate presumably as an energy source for the malignant cells, and citrate levels are consequently lower in prostate cancer tissue(Costello, Franklin, & Narayan

1999).

21 1.4 Natural history of prostate cancer and active surveillance

The natural history of prostate cancer is the clinical and pathological manifestations of the tumour from its inception until death of the untreated host. This depends on the malignant nature of the cancer and also the life expectancy of the individual.

The development and subsequent progression of prostate cancer is reliant on the

acquisition of genetic changes that convert a benign epithelial cell through a malignant

cell with invasive potential, to a cell capable of metastasising and in some cases

becoming hormone independent. The adaptation by the tumour requires specific genetic

changes at each step to allow progression - the multi-hit hypothesis (Figure 1.2).

Figure 1.2. Genetic changes underlying the initiation and progression of prostate cancer

Normal prostate Gains Germmline mutation epithelium + polymorphisms

HGPIN lq 24-25 Xq27-28 (HPCX) lq42-43 (PCAB) Localised lp36 prostate cancer 16q23

Xqll-12(AR) MXI1

7 Metastatic 8q PSCA prostate cancer EIF3S3

Hormone refractory Xq prostate cancer

22 This hypothesis underlies how, by being able to identify specific changes at each step, we will be able to predict disease outcomes and target new therapeutic interventions. In the meantime treatments are tailored according to our understanding of prostate cancer gained from large, but unrandomised case series.

Gaining information regarding the natural history of a life-threatening condition is almost impossible. Attempting to randomise patients with anything other than low stage and low-intermediate grade prostate cancer to receive ‘no treatment’ is unethical.

Therapeutic intervention, which may alter the natural history of the disease, is frequently required as the untreated cancer progresses. As a result good quality data for

‘conservative’ management is lacking. We rely on unrandomised or pooled data gathered from patients who have not received radical therapy for a variety of reasons.

Comparing these cohorts with the large series of patients selected for radical treatment is questionable. What is more, staging information in these ‘watch and wait’ series is based on clinical findings, which are inherently crude.

Chodak et al. published a pooled analysis of 6 unrandomised studies containing 828 patients managed by either observation alone or observation and delayed hormonal therapy, in men with clinically localised disease diagnosed with cytology or biopsy

(Chodak et al. 1994). Disease-specific survival at 10 years (survival among only those patients who did not die of causes other than prostate cancer) was 87% for those with well or moderately differentiated tumours (Gleason grade 2-4 and 5-7 respectively) compared with 34% for those with poorly differentiated disease (Gleason 8-10). An argument for radical therapy is that overall survival after treatment approximates to overall survival in the general population. However, the authors of this study concluded

23 that the same outcome could be achieved by observation alone in patients with a

Gleason score of 7 or less.

Table 1.1. Outcome of patients undergoing active surveillance (Chodak et al. 1994)

Percent of patients (95% Cl)

5 years 10 years

Disease-specific survival

Grade 1 98(96-99) 87(81-91)

Grade 2 97(93-98) 87(80-92)

Grade 3 67(51-79) 34(19-50)

Metastasis-free survival

Grade 1 93(90-95) 81(75-86)

Grade 2 84(79-89) 58(49-66)

Grade 3 51(36-64) 26(13-41)

The metastases free rate at 10 years in the Chodak study was 81% for well differentiated, 58% for moderately differentiated and 26% for poorly differentiated prostate cancers. Although 81% of patients with moderately differentiated tumours were not dying of prostate cancer during the study period, nearly 40% progressed to potentially painful and symptomatic distant metastases. In this study, the mean patient age was 70 years. The mean age for men considered for radical treatment is 60, so in these men, who have a 20-25 year life-expectancy longer follow-up for conservative treatment is required.

24 A longitudinal study in a single catchment area in Sweden examined the role of conservative treatment of prostate cancer (Johansson et al. 1997). However, for the reasons outlined, it was not randomised and was therefore biased as men with high- grade tumours received radiotherapy and were excluded. The study cohort included men with clinically localised disease, locally advanced disease and metastatic disease; the

15-year disease-specific survival was 89%, 57% and 6%, respectively. Of interest, there was no survival benefit between those who received initial androgen deprivation therapy and those in whom it was delayed; and importantly they found grade, more than stage, was the most powerful predictor of survival. Indeed, 6% of men with well- differentiated tumours died within 15 years from prostate cancer, compared with 17% with moderately differentiated and 57% with poorly differentiated disease, regardless of

stage.

In a study from Connecticut, survival was modelled on a proportional hazards

regression, Gleason score and co-morbid status were the only two variables of

significance (Albertsen et al. 1995). They concluded the age-adjusted survival for men

with Gleason 2-4 tumours was similar to that of the general population and so

questioned the use of aggressive therapy in this group. The maximum life lost for men

with Gleason 5-7 tumours was 4 to 5 years and 6 to 8 years for men with Gleason 8-10

tumours. The Connecticut group drew up computer formulae, that include Gleason

grade and co-morbid features, to estimate age-adjusted life year gains (Albertsen et al.

1996), and more recently updated this to include men aged between 55-74 (Albertsen et

al. 1998).

Whilst all these studies can be criticised for their lack of randomisation, their relatively

short follow-up and the inherent inaccuracy of clinical staging they do all show a

25 consistent pattern; the Gleason score is currently the most useful indicator of the natural history of prostate cancer. Furthermore, the Connecticut nomograms incorporating the

Gleason score and patient co-morbidity can be used in the clinic to counsel patients on treatment choices armed with an accurate prediction of their survival.

The prostate specific antigen (PSA) has played a role in altering the natural history of this disease. PSA can affect the lead-time to diagnosis, as demonstrated by the longitudinal ’s Health Study. This showed that a single PSA measurement can detect 80% of prostate cancers diagnosed at least 5 years before they become clinically apparent (Gann, Hennekens, & Stampfer 1995). What is more, PSA detected cancers are more aggressive, but generally of lower stage, than cancers detected ‘incidentally’ at transurethral resection of the prostate (TURP). Soh et al. reported a single surgeon series from 1983-1995 (Soh et al. 1997). Prior to 1989 most radical were performed in response to cancer diagnosed on TURP (Tla and Tib). Since 1989, most tumours were diagnosed as a result of a raised PSA, tumours diagnosed more recently are higher grade, but lower stage.

The role of serum PSA as a predictor of disease progression in men under active surveillance was assessed in another Swedish study, which selected 658 well men bom in 1913 who were 67 years old at randomisation (Hugosson et al. 2000) A blood sample was extracted and stored frozen and the group followed for 15 years. Patients’ management was according to local protocols and no patient had PSA screening or curative treatment for prostatic cancer. Fifty-two (7.9%) men developed clinical prostate cancer and in these men the risk of dying of the disease was related to the serum PSA at age 67. If at the age of 67 the PSA was <3ng/ml, 5% developed clinical cancer, if the

PSA was 3-10ng/ml, 30% developed the disease and if the PSA was >10ng/ml 50% men developed clinical prostate cancer.

26 In recent years there has been an increasing vogue for active surveillance. While the proponents accept there is no long-term outcome data for active surveillance they claim the number of men over treated by radical therapies warrants further investigation

(Parker 2004). Men are observed by measuring their PSA monthly and by repeating the biopsy Gleason score every 18 months. Patients with a PSA doubling time less than 1 year or men who are up graded to Gleason 8-10 on re-biopsy are referred for radical therapy. Up to 80% of men may be prevented from undergoing radical treatment, it remains to be seen how many patients ‘spared’ radical treatment eventually die of prostate cancer compared with a treated group. Furthermore, optimal surveillance programmes have not been established so the correct PSADT to trigger treatment is currently unknown, and there is a reliance on the Gleason score from prostate biopsies that prone to sampling and interpretative errors.

27 1.5 Staging prostate cancer

Staging is the systematic classification of cancer extent and as with the vast majority of , the extent of prostatic cancer both in terms of the primary site and the secondary or metastatic sites has a profound effect on "curability’ and on prognosis. The clinical staging of prostate cancer may include a digital rectal examination (DRE), prostatic biopsy. Computer Tomography (CT) or Magnetic Resonance Imaging (MRI) to ascertain the local stage and radio-isotope -scan, Positron Emission Tomography

(PET), CT or MRI to assess metastatic spread.

The adoption of an agreed classification of tumour extent enables comparisons to be made between different stages of disease and allows data from different centres to be compared in a meaningful way. Two staging classifications are used to describe the extent of prostate cancer the TNM (Tumour, Node and Metastases) and the Whitmore-

Jewet staging system. These systems incorporate clinical and pathological tumour extent running in parallel

In 1956 Whitmore subdivided prostate cancer into four clinical stages: Stage I represents clinically latent disease, Stage II early but palpable disease, Stage III locally advanced disease without metastases and Stage IV distant metastases irrespective of the local stage. These categories were subsequently revised by Prout and Jewett. This revised system was adopted in the US initially but has now been largely superseded by the TNM classification.

The TNM classification (Table 1.2) has been used to categorise all solid cancers and has been promoted by the Union Internationale Contre le Cancer (UICC), the American

Joint Committee for Cancer (AJC) and other international bodies since the 1970’s. The

28 TNM classification recognises the limitations of clinical staging by differentiating post- surgical or pathological staging by the pre-fix ‘p \ It is now the most frequently used staging system in prostate cancer and will be the system of choice in this thesis.

Table 1.2. TNM classification of prostate cancer (UICC 6th edition, 2002)

T1 Impalpable and not visible on imaging

T la Tumour incidental finding in <5% tissue resected at TURP

T ib Tumour incidental finding in >5% tissue resected at TURP

T ic Tumour identified by needle biopsy

T2 Tumour confined within the prostate

T2a Tumour involves one half of one lobe, or less

T2b Tumour involves more than one half of one lobe, but not both lobes

T2c Tumour involves both lobes

T3 Tumour extends through the prostatic capsule

T3a Unilateral or bilateral extracapsular extension

T3b Tumour invades seminal vesicle(s)

T4 Tumour is fixed or invades adjacent structures other than the seminal

vesicles’, bladder neck, external sphincter, rectum, levator muscles or

pelvic wall

N - X, 0,1 Lymph nodes - Not assessed, no regional metastases, regional node

metastases

M l - a, b, c Distant metastases - non-regional metastases, bone

metastases, other site

29 1.6(a) Grading prostate cancer

The diagnosis of prostate cancer requires the histological examination of cancer tissue.

Once established the glandular architecture and the cellular differentiation are assessed to grade the disease. Several grading systems have been proposed, though the Gleason system, described in 1974 (Gleason & Mellinger 1974), remains the standard (Figure

1.3).

Gleason, in conjunction with the Veterans Administration Co-operative Urological

Research Group (VACURG), described the overall pattern of growth of the tumour as examined at low magnification (40-100x). The cytological characteristics seen at higher

magnification were not part of Gleason’s description. The patterns were correlated with

clinical outcome by an independent observer and 5 patterns were found to be associated

with prognosis. Where more than one pattern was present in a cancer, patient mortality

fell between that seen in the two patterns in pure form. Hence the two predominant

patterns were recorded in order of area and summated as the Gleason sum score e.g

Gleason 3+4. In tumours of pure pattern the number was doubled e.g 3+3. Since

Gleason’s original description, variation has existed as to whether the second pattern

reported in biopsy specimens should be the second most frequent area or the highest

pattern grade. Generally, the two dominant areas are still reported and summated to give

the overall Gleason score (Bostwick et al. 2000). The highest grade may also be

reported if different from the predominant patterns as this area, although small, maybe

of prognostic importance, particularly at biopsy where sampling error may result in the

dominant tumour being under-represented.

30 Prostatic AocnogarginOma _ (HlfttolOgic G rades) Qy © ! © a

Figure 1.3 The Gleason grading system of prostate cancer

The problem with the Gleason scoring system is that it is a rigid scoring system based on an architectural pattern that is a continuum. The score categorises patterns 1-5, these groupings do not demonstrate malignant potential in the same stepwise fashion. Gleason sum 2-5 tumours behave in a similar manner and some authors have categorised these as ‘insignificant’ tumours, unlikely to affect the quality or quantity of the patient’s life

(Epstein, Walsh, & Brendler 1994; Jack et al. 2002). However, within the Gleason sum category 7 two distinct groups exist, 3+4 and 4+3, which are associated with differing outcomes. In a radical prostatectomy series reported from the University of Texas, 2.7% of 73 patients with 3+4 disease developed biochemical relapse compared with 12.8% of

86 men with 4+3 disease (p=0.02) after a minimum 48 months follow-up (Babaian et al.

2001). Johns Hopkins have also reported their experience of Gleason 7 disease in 564 men (Han et al. 2000). A significant recurrence-free survival advantage was found in men who underwent radical prostatectomy for Gleason 3+4 compared with those with

Gleason 4+3 disease (p<0.0001). An artificial neural network, logistic regression, and proportion hazard models all demonstrated the importance of the Gleason 7 constituent parts in predicting patient outcome (Han et al. 2000; Partin et al. 2001).

31 There is significant inter and intra-observer variability in the Gleason scoring system.

Inter-observer variability may be as high as 23%, usually as a result of under-grading

Gleason pattern 3 and 4 (Gleason 1992; Steinberg et al. 1997). Gleason himself reported a 50% intraobserver variation (Gleason 1992). These differences may be reduced by the reporting pathologist re-examining the slides and by review of all cancer diagnoses by specialist uropathologists (Steinberg et al. 1997). In practice, in the UK, this is not usually feasible. Never the less, the Gleason score currently offers one of the best prognostic indicators in prostate cancer.

1.6(b) Tumour multifocality

Prostatic cancer is both a heterogeneous and multifocal disease. It is usually made up of a dominant or index tumour consisting of several different sub-clones at different stages of tumour development - hence the Gleason scoring system includes two grading patterns. Frequently, there are also other smaller foci of adenocarcinoma within the prostate that are genetically distinct or non-clonal (Bostwick et al. 1998). The Stanford

group, using 3mm step sections, reported 83% of 486 men treated by radical prostatectomy had multiple tumour foci (Wise et al. 2002). The median number of additional foci was 2.0 (range 0-12), however the total volume of these additional tumours was <0.5cm in 234 (58%) cases. Interestingly, an increasing number of

smaller cancers had a beneficial effect on PSA failure rates. The index tumour volume was a powerful determinant of PSA failure, however there was a decline in the size of the index tumour with each increase in the number of smaller cancers. Rashid et al. also found that the index cancer and not the multifocality predicted PSA free survival

(Rashid, Wojno, & Marcovich 1999) and Haggman et al. concluded the most significant malignant features were always represented in the index tumour and less commonly in

32 the remaining foci (Haggman et al. 1997a). A more recent study has disputed these findings, and has identified genetic changes in metastases, which were only detectable in the smaller foci (Isaacs, De Marzo, & Nelson 2002). However, it is usually the largest focus that contains the highest-grade tumour and it is usually this tumour that dictates clinical outcome.

The fact that several genetically different tumours co-exist within one prostate has hampered the search for molecular markers in prostatic cancer as homogeneous samples are hard to obtain. With the advent of laser capture of individual cells and cell-sorting techniques, which enable the isolation of relatively pure populations of malignant cells, these difficulties have been partially circumvented. Now the difficulty is in selecting the

correct cell or population for investigation.

33 1.7 Localised prostate cancer

Localised prostatic cancer is cancer that is organ confined, that is it has not spread beyond the fibromuscular ‘capsule’ and at this stage the cancer is curable.

1.7(a) Presentation

In England and Wales 52% of new cases of prostate cancer are thought to be localised to the prostate (T1/2N0M0) (National Institute for Clinical Excellence 2002). Organ confined disease is unlikely to cause lower urinary tract symptoms, as tumours that are large enough to encroach on the urethra are usually locally advanced. Patients with localised disease are usually diagnosed following transrectal ultrasound guided biopsies performed as a result of an abnormal feeling prostate on digital rectal examination

(DRE) and/or an elevated serum prostate specific antigen (PSA). A minority are detected incidentally after transurethral resection of the prostate for benign prostatic hyperplasia. The proportion of men diagnosed with cancer after TURP has fallen since ^ o the introduction of PSA testing. From the Utah hospital discharge database, cancer was ^ --'i'1' «■

— y i c detected in 39% of men undergoing TURP between 1980-1984 compared with 7.4% ^ — .. ^ ,, , between 1995-1999 (Merrill & Wiggins 2002). In a recent UK series of 1001 men /. / undergoing radical prostatectomy between 1988 and 2001 cancer diagnosed exclusively ^ > // on TURP tissue accounted for 5% of cases (Bott et al. 2005c), which compares with 6- >

K / 9% in other series (Quinn et al. 2001; Renshaw et al. 1999; van den Ouden et al. 1997).

A /

The mode of tissue diagnosis is relevant to findings in the radical specimen. Tumours diagnosed on TURP are more likely to represent transition zone cancer which is commonly of lower Gleason score than tumours arising in the peripheral zone (Bott et al. 2005c; Erbersdobler et al. 2002a; McNeal et al. 1988).

34 1.7(b) Patient selection for radical prostatectomy

Patient selection involves correctly identifying those patients most likely to benefit from radical treatment. This is of paramount importance in men considered for radical prostatectomy where the desire to offer a potentially curative treatment has to be weighed up against the risk of finding disease outside the prostate where cure is less likely, or an insignificant tumour where cure is unnecessary.

Although there is no absolute means of identifying those individuals with pre-existing microscopic invasion of extraprostatic tissue, methods are now available to establish those at highest risk. Preoperative risk factors have been combined to produce a statistical model which will predict pathological or biochemical outcome (Bott et al.

2003c; Kattan et al. 1998; Partin et al. 2001).

V i f 1.7(c) Preoperative predictors of outcome

1.7(c)!. Clinical Stage

The clinical stage is derived from the digital rectal examination (DRE) and imaging investigations. Assessment of clinical stage preoperatively is important as understaging prostate cancer may mean a patient receiving radical treatment with its inherent risk of complications without the benefit of cancer cure, whereas to overstage denies a patient the chance for a potentially curative treatment.

Clinical stage is reported by some to predict pathological stage and biochemical outcome after radical prostatectomy (Hull et al. 2002; Partin et al. 2001), but not others

(Babaian et al. 2001; Kattan et al. 2003; Stamey et al. 2000). In an update of Partin’s tables published in 2001 involving over 5000 men undergoing radical prostatectomy 63% had impalpable (Tic) prostate cancer and 46% had palpable but clinically localised disease (T2a-c) (Partin et al. 2001). At examination of the prostatectomy specimen 64% had organ confined disease, 30% extracapsular extension (pT3a), 4% seminal vesicle involvement (pT3b) and 2% lymph node positive disease. While the DRE incorrectly predicted organ-confined disease in 36% of men, it was a significant predictor of pathological stage in multinomial log-linear progression analysis (p<0.001) and was included in the tables. In van den Ouden’s series of 273 men 51% were clinically understaged and 8% overstaged (van den Ouden et al. 1997). Of note, 18% of men with cT3 disease were over-staged. This highlights the problem of current preoperative predictors of outcome. Men with clinical T3 disease may be denied radical therapy as cure is considered unlikely, but nearly one in five of these men will potentially be excluded from curative treatment because of staging inaccuracies.

Clinical stage may predict biochemical outcome after radical prostatectomy. In

Catalona’s series of 1778 men, non-progression was reported in 79% (95% Cl 73-86) with cTl disease, 66% (62-69) for cT2 tumours and 44 (22-67) in men with cT3 disease

(Catalona & Smith 1998). The log rank test comparing clinical stage and cancer recurrence was significant (p<0.001). However, using multivariate analysis in this series, and in other series (Babaian et al. 2001; Kattan et al. 2003; Stamey et al. 2000) clinical stage was not an independent predictor of disease-specific survival.

In summary clinical stage does have some prognostic importance, particularly when used in conjunction with other predictors, it is a simple and ‘non-invasive’ procedure to perform. But, it lacks sensitivity and even when performed by specialist prostate surgeons DRE fails to be an independent predictor of pathological stage or biochemical outcome in most series.

36 1.7(c)2. PSA

The preoperative PSA has been shown to be a useful predictor of pathological stage and biochemical outcome after radical prostatectomy in most series (Catalona & Smith

1998; Hull et al. 2002; Partin et al. 2001; Quinn et al. 2001), but not all (Babaian et al.

2001). In the Washington series, PSA was significant in both univariate and

multivariate analysis at predicting cancer outcome at 7 years, despite some ‘high-risk’

patients receiving adjuvant radiotherapy in this series (Catalona & Smith 1998). Similar

findings were reported by others in both univariate and multivariate analysis (Kattan et

al. 2003; Quinn et al. 2001). Hull et al. reported the preoperative variables PSA, biopsy

Gleason score and clinical stage were all independent predictors of extracapsular

extension, metastatic-free survival and disease progression (Hull et al. 2002). When the

pathological features from the radical prostatectomy specimen were included in the

multivariate analysis only the pathological stage and surgical margin status were

significant independent predictors of progression, not the preoperative features. In other

series PSA has failed to predict biochemical outcome (Babaian et al. 2001).

The most frequently sited inflection points for preoperative PSA are at 4, 10 and

20ng/ml (D'Amico et al. 2000b; Hull et al. 2002; Pound et al. 1997; Quinn et al. 2001).

In the most recent version of Partin’s tables the PSA groups were subdivided in to five

categories: 0-2.5, 2.6-4, 4.1-6.0, 6.1-10 and >10 (Partin et al. 2001). In such a large

cohort these subcategories contributed significantly to the prediction of pathological

stage, in smaller series it is unlikely such narrow intervals would be predictive. In other

predictive nomograms PSA is considered a continuous variable and such artificial

grouping is avoided (D'Amico et al. 2000b; Kattan et al. 1998).

Serum PSA values may be altered by other factors including: benign prostatic

hyperplasia, prostatitis, ejaculation, tumour volume and tumour differentiation, as well

37 as differences in assay technique. Whilst attempts have been made to increase the specificity of PSA by measuring the free:total ratio, the PSA density etc these have been found to add little to the diagnostic power of PSA. Furthermore, in patients with a PSA within the range 2.5-6.0ng/ml, increasing age is more strongly associated with incurable prostate cancer than the PSA level (Carter, Epstein, & Partin 1999). On the other hand in men with PSA below the standard cut-off of 4.0ng/nl the cancer detection rate was between 15% - 25%, in the ERSPC and SWOG series (Raaijmakers et al. 2004;

Thompson et al. 2004). In men with a ‘normal’ age-related PSA the chance of finding an insignificant tumour at radical prostatectomy was greater than in men with a higher

PSA, yet a significant number of men with a ‘normal’ PSA have adverse pathological findings (Freedland et al. 2004). These studies demonstrate PSA alone cannot differentiate patients with early organ confined cancer from those with extraprostatic disease. Recently the PSA velocity has been used to predict outcome, before and after radical treatment. D’Amico reported a pretreatment PSA velocity of >2ng/ml/year was a significant risk factor for recurrence after radical prostatectomy and was associated with a significantly shorter time to death from prostate cancer (D'Amico et al. 2004a). In a separate study the same authors showed that a PSA doubling time after RP or EBRT of

<3months was a significant predictor of prostate cancer mortality (D'Amico et al.

2004b).

1.7(c)3. Biopsy Gleason score

The biopsy Gleason score independently predicts tumour stage (Partin et al. 1997;

Partin et al. 2001) and biochemical outcome (Kattan et al. 2003). In a multicentre series of 4,133 cases, 83% of men who had biopsy Gleason score 8-10 had extraprostatic extension, 24% seminal vesicle involvement and 20% lymph node metastases. Whereas,

38 38% of men with biopsy Gleason score 6 or less had extracapsular extension, 4% had seminal vesicle invasion and 3% had lymph node involvement (Partin et al. 1997).

Gleason biopsy score was a strong predictor of pathological stage (Partin et al. 1997).

A report from the same centre categorised men into biopsy Gleason 2-4, 5, 6, 7, 8-10.

Using Kaplan-Meier actuarial recurrence-free analysis; a significant difference in biochemical outcome was seen between each group, except between Gleason 7 and 8-10

(p=0.05) (Partin et al. 1993).

Studies comparing Gleason score at biopsy and the Gleason score of the whole tumour at radical prostatectomy show an approximate 50% correlation, even though therapeutic decisions are often based heavily on the biopsy Gleason score (Bott et al. 2005c; Bott et al. 2003a; Carlson et al. 1998). Correlation is improved when the primary pathologist re-examines the specimen and also reviews the slides in conjunction with 2-5 specialist uropathologists (Carlson et al. 1998). In the setting of the National Health Service this is not practical and so patients and doctors must be aware of the inherent inaccuracy of a biopsy Gleason. In our own series the biopsy and radical scores were identical in

252/536 (47%), the biopsy score being lower and higher than the radical score in

170/536 (32%) and 113/536 (21%), respectively (Bott et al. 2005c). For the purposes of the database, the biopsy Gleason score was the summation of the highest two grades in - a biopsy series rather than the overall Gleason sco based on area. The radical ^ prostatectomy Gleason score was composed of the two Gleason grades occupying the largest areas. The biopsy Gleason score equalled the radical score +/-1 in 487/536

(91%) of patients. Percentage agreement between biopsy and radical Gleason score increased with the number of cores submitted (37% agreement with < 5 cores, 48% with

6 cores, 56% with > 7 cores, p=0.006). There was no association between level of

39 agreement and number of cores containing cancer, longest length invaded with cancer in a single core and the highest percentage of a single core invaded with cancer.

Lack of correlation occurs because of inadequate sampling and interpretative error.

Sampling error may occur as prostate cancer is heterogeneous, the biopsied tumour may not be the same as the two largest areas of tumour in the radical prostatectomy specimen, which constitute the radical Gleason score. Interpretative error occurs as a result of mechanical artefact, for example Gleason 3 acini maybe compressed during paraffin embedding mimicking sheets of malignant cells typical of the Gleason 5 pattern. Alternatively interobserver variation occurs due to differences in interpretation of architectural patterns as well as level of expertise.

1.7(c)4. Biopsy features

While TRUS alone is an insensitive tool for diagnosing and staging prostate cancer

TRUS guided prostatic biopsies may be used to predict extracapsular disease and biochemical outcome. The number of positive cores (Badalament et al. 1996; D'Amico et al. 2000b; Grossfeld et al. 2002), tumour location (base vrs mid vrs apex)

(Badalament et al. 1996), percent tumour involvement (Badalament et al. 1996), highest

Gleason score and presence of Gleason 4 or 5 have all been used as preoperative staging tools (Orozco et al. 1998). The best predictor of outcome is the proportion of cores positive for cancer (Bostwick et al. 2000). In a validated series of 960 men, D’Amico reported of men with <34, 34-50 or >50% positive cores 79%, 59% and 43% had organ- confined disease, and the actuarial biochemical outcome at 4 years after surgery was

86%, 48%, and 20%, respectively (D'Amico et al. 2000b). The percentage number of positive biopsy cores was an independent predictor of time to PSA progression. The use of the number of positive cores requires careful histological processing as the submitted specimens may not be whole or may fragment during fixation and embedding. Separate

40 specimen pots and separate paraffin blocks for each core are usually required if the number of positive cores are to be used for prognostic information.

1.7(c)5. Preoperative imaging

The ability of imaging modalities to correctly stage prostate cancer remains limited.

Extracapsular extension is often microscopic and current techniques are not of sufficient resolution to detect small breaches in the capsule. The apex of the prostate is the most frequent site for extracapsular disease but as the capsule is ill-defined in this region

(31), it is difficult to distinguish extracapsular from organ-confined tumour. Where macroscopic tumour extension is present magnetic resonance imaging (MRI) is the most sensitive and specific technique available; nevertheless there is still a high rate of false positives as well as considerable interobserver variability (Table 1.3).

Table 1.3. Comparison of body coil and endorectal coil magnetic resonance imaging at detecting extracapsular extension of prostatic cancer.

Author Number Sensitivity Specificity T esla M ode Bott 2003 46 42% 78% 1 Bodycoii Comudl002 30 66% 94% 1.5 Body coil Deasy 1007 53 55% 91% 0.2 Bodycoii Rifkin 1990 230 77% 57% 1.5 Bodycoii Tuzel 109$ 61 38% 87% 1 Endorectal coil Bates 1997 20 38% 100% 0.5 Endorectal coil Tempany 1004 201 61% 62% 1.5 Body coil Temparty 1994 183 60% 42% 1.5 Endorectal coil

(Bott et al. 2003b), (Comud et al. 1992), (Deasy et al. 1997), (Rifkin et al. 1990), (Tuzel et al. 1998), (Bates et al. 1997), (Tempany et al. 1994).

41 D’Amico et al. suggested a role for endorectal MRI (eMRI) in predicting biochemical outcome after radical prostatectomy. While eMRI failed to add any useful clinical information in 81 % of over a thousand cases, it was clinically and statistically relevant in the 5-year PSA outcomes in the remaining ‘intermediate risk’ group (D'Amico et al.

2000a). Others have highlighted the accuracy of MRI varies according to the expertise of the individual radiologist (Tempany et al. 1994), so the results achieved by D’Amico et al. need to be replicated in other centres. In a meta-analysis of MRI in prostate it was noted there was significant heterogeneity of MRI results, in expert hands

MRI improves staging performance (Engelbrecht et al. 2002)

Other modalities including transrectal ultrasound (TRUS) and pelvic CT are not sufficiently accurate to detect extracapsular extension. TRUS was no better than DRE in predicting extracapsular extension in a prospective multicentre study (Yu & Hricak

2000). CT has a reported sensitivity of between 55-75% at detecting extracapsular extension (Huncharek & Muscat 1996; Manyak & Javitt 1998). In a review of 25 studies incorporating 4264 men with newly diagnosed prostate cancer who underwent staging CT, 0 and 1.1% of patients with PSA <20ng/ml and >20ng/ml, respectively had pelvic lymphadenopathy on CT crtiteria (Abuzallouf, Dayes, & Lukka 2004). CT is therefore not indicated for the vast majority of men being considered for radical prostatectomy, though may be of use in men with Gleason 8-10 or locally advanced cancer (American Urological Association (AUA) 2000).

Radio-isotope bone-scanning is a non-invasive tool for diagnosing bone metastases. In newly diagnosed clinically localised prostate cancer, where the incidence of metastases is rare and where present usually of low volume, bone-scans are unwarranted. In a review of 8644 men in who 16.8% had positive scans, metastases were identified in 5% of men with a PSA <20ng/ml and 2% in men with PSA <10ng/ml (Abuzallouf, Dayes,

& Lukka 2004). Men with clinically organ-confined disease had positive scans in 6% of

42 cases and men with Gleason score <7, 5% had a positive scan. This and other studies therefore recommend staging bone scans in men with PSA exceeding 20ng/ml or with

Gleason score 8 or above or in men with locally advanced disease (T3/4) (Abuzallouf,

Dayes, & Lukka 2004; American Urological Association (AUA) 2000).

Positron emission tomography (PET), a non-invasive imaging modality, has been investigated for its use in staging prostate cancer. It relies on the increased uptake in tumours of a radiolabelled choline, flurine or glucose analogue and prostate cancer epithelial cells do not metabolise these compounds at a sufficent rate for detection. This technique is still under evaluation in prostate cancer, however it is likely only to be of benefit in the staging of prostate cancer that has metastasised as the false positive rate for extracapsular disease is unacceptably high (DeGrado et al. 2001; Hoh et al. 1998). In practice, radiological evidence of extra-capsular disease is usually sought only in those men with a high risk - those with a PSA >20ng/ml or with a Gleason score of 8 or more

O*

1.7(c)6. Circulating prostate cells

Several groups have tried to use preoperative PSA reverse transcriptase polymerase chain reaction (RT-PCR) positivity of blood to predict the likelihood of recurrence after radical prostatectomy. Some have found a positive correlation using this technique

(Olsson et al. 1996) others have not (Ellis et al. 1998), further studies are awaited.

1.7(c)7. PSA density, PSA free:total ratio, PSMA

The PSA density (total PSA divided by the prostate volume), and the PSA free:total ratio (free PSA divided by the free PSA and the PSA bound to al-chymotrypsin) are

43 assays used to try and enhance the sensitivity and specificity of PSA prostate cancer detection. To date however studies have not been conclusive in demonstrating prognostic potential over and above total PSA measurement (Epstein et al. 1994;

Pannek et al. 1998; Raaijmakers et al. 2004; Ravery et al. 2000).

Newer techniques are evolving in an attempt to augment or improve on the ‘classical three’ preoperative predictors (biopsy Gleason grade, clinical stage and preoperative

PSA score). Kattan et al. has included molecular markers, including Interleukin-6 soluble receptor and Transforming Growth Factor betal, into nomograms used to predict biochemical failure after radical treatment and has demonstrated their independent prognostic importance (Kattan et al. 2003). Human kallikrein 2 (Hk2) has also been shown to predict organ confined disease after radical prostatectomy (Haese et al. 2003). DNA ploidy from prostatic biopsies has been reported as an independent predictor of outcome after radical prostatectomy (Ross et al. 1999), though this investigation is not available in most centres. To date the tables devised by Partin remain the most useful in the clinic as they incorporate easily obtainable variables combined to make simple-to-use tables. In the future, as in breast and other cancers, molecular markers are likely to replace the ‘classical’ triad by providing greater predictive capacity.

44 1.8 Post-operative predictors of outcome

Currently available prognostic factors have been ranked by a multidisciplinary group of clinicians, pathologists and statisticians and subsequently endorsed by The World

Health Organisation's Second International Consultation on Prostate Cancer (Bostwick et al. 2000). Category I are factors of proven prognostic importance and are currently used in patient management. These include the preoperative PSA, the TNM stage, the

Gleason grade and the surgical margin status. Category II factors are those that have been extensively studied but whose importance remains to be validated in statistically robust studies e.g. tumour volume, histological type and DNA ploidy. Category III include all other factors not studied sufficiently to demonstrate their prognostic value - oncogenes, tumour suppressor genes, apoptosis genes as well as perineural invasion, neuroendocrine differentiation, microvessel density, nuclear roundness and chromatin texture (Bostwick et al. 2000).

45 Table 1.4. Risk factors for recurrence after radical prostatectomy

(Bostwick et al. 2000)

Pre-operative factors

Higher clinical stage

Biopsy Gleason containing 4 or 5

Pre-operative PSA >10ng/ml

Greater number of positive biopsy cores

Post-operative features

Category I

Higher stage

Higher Gleason score

Positive surgical margins

Category II

DNA ploidy

Tumour volume

Histological type (signet ring, sarcomatoid, mucinous)

Category III

Tumour suppresser genes

Oncogenes

Perineural invasion

Neuroendocrine differentiation

Microvessel density

Nuclear roundness and chromatin texture

46 1.8(a) The Radical Prostatectomy Gleason Score

Gleason demonstrated a strong correlation between his grading system and patient mortality (Gleason & Mellinger 1974). Nowadays, the Gleason score is accepted as one of the most important predictors of tumour stage, progression and survival (Babaian et al. 2001; Catalona & Smith 1998; Epstein, Pizov, & Walsh 1993; Gerber et al. 1996;

Gleason 1992; Stamey et al. 2000; Steinberg et al. 1997). Epstein et al. categorised

Gleason patterns after radical prostatectomy into four groups, Gleason sum 2-4, 4-6, 7 and 8-9, in their series of 507 men (Epstein, Pizov, & Walsh 1993). There was no difference in terms of local progression or overall survival between men with Gleason scores 2-4 or 4-6. However, there was a highly significant difference between men with

Gleason 2-6 compared with Gleason sum 7 or Gleason sum 8-9 (p<0.0001). There was also a highly significant difference between Gleason 7 and Gleason 8-10 (p=0.001) in local and overall progression. In this series, like several others (Gerber et al. 1996;

Lemer et al. 1996; Zincke et al. 1994), the Gleason sum was the prognostic factor most strongly correlated with overall and local progression (p=0.0001).

Catalona reported on a series of 1778 men who had undergone radical prostatectomy, using the log rank test the radical Gleason sum was a significant predictor of survival

(Catalona & Smith 1998). The 7-year probability of nonprogression was 84% (79-89) for well differentiated (Gleason 2-4), 68% (64-73) for moderately differentiated

(Gleason 5-7) and 48 (39-57) for poorly differentiated (Gleason 8-10) tumours. This study also highlighted the impact of high tumour grade on survival. Of the 103 men who had died of any cause, including those who died of prostate cancer, high tumour grade versus moderate or well differentiated was a significant predictor of death

(p<0.0001). Other authors have also found only poorly differentiated tumours to be independent predictors of outcome in multivariate analysis (Stamey et al. 2000; van den

O udenetal. 1997).

1.8(b) Pathological stage

The pathological stage of the radical prostatectomy specimen describes the extent of the tumour and when lymphadenectomy is performed the presence or absence of obturator lymph node metastases. The presence of capsular penetration - stage pT3a disease, is associated with poorer outcome than organ confined disease (Babaian et al. 2001;

Paulson, Moul, & Walther 1990; Stamey et al. 1999; van den Ouden et al. 1997).

Babaian et al. reported 5.4% of men with organ-confined disease, 17.8% with extraprostatic cancer extension and 61% with seminal vesicle invasion developed disease progression using Kaplan Meier outcome curves (Babaian et al. 2001). Quinn et al. reported their multivariate analysis of 732 Australian men who had undergone radical prostatectomy. Independent predictors of outcome in descending order were lymph node metastases, pT3 vrs pT2, seminal vesicle invasion, Gleason score and preoperative PSA (Quinn et al. 2001). Van den Ouden et al. reported capsular penetration (pT2 versus pT3) was a significant predictor of biochemical progression and clinical progression, but not cause specific survival (RR 2.5, 2.6, 2.4, respectively) (van den Ouden et al. 1997). Despite seminal vesicle involvement in 29% of cases in this study and lymph node metastases in 10% neither was an independent predictor in multivariate analysis, probably as there were only a few deaths from prostate cancer during the follow-up period.

Not all researchers however have found local pathological tumour stage to be an independent prognostic marker. Lemer et al. reported their series of nearly one thousand patients (Lemer et al. 1996). They found preoperative PSA, grade, DNA ploidy and

48 clinical stage were highly significant indicators of progression in univariate analysis but, pathological stage and patient age did not significantly impact on progression free survival.

Stage pT3a can be further sub-classified into focal capsular penetration - where a few acini breach the ‘capsule’, or extensive capsular penetration where disease spread is more widespread, and these differences have prognostic implications. The results of the actuarial 5- and 10-year likelihood of freedom from PSA relapse in 894 men after radical prostatectomy for clinically localised prostate cancer according to pathological stage from the Johns Hopkins are shown in Table 1.5 (Partin et al. 1993).

Table 1.5. Actuarial biochemical free survival and prostate tumour stage

(Partin et al. 1993)

Pathological stage Actuarial percentage (95%C1)

5 years 10 years

Organ confined 97 (94-99) 85 (67-94)

Focal capsular penetration 90 (84-95) 82 (82-90)

Established capsular penetration, low grade 88 (81-93) 54 (20-80)*

Established capsular penetration, high grade 68 (53-80) 42 (20-64)

Seminal vesicle involvement 47 (31-63) 43 (26-60)

Lymph node involvement 15 (6-29) -

* Eight year statistics

The Kaplan-Meier actuarial likelihood of PSA relapse was significantly different between stages. There was a significant difference between men with focal capsular penetration of any Gleason score and men with high-grade (Gleason >7) extensive

49 capsular penetration, but not men with low-grade (Gleason <7) and extensive capsular penetration (p=0.06). Interestingly there was no significant difference in actuarial

likelihood of PSA relapse between men with extensive high-grade capsular penetration

and those with seminal vesicle involvement (p=0.02). Not all studies have found the

extent of extraprostatic pT3a disease to be significant (Babaian et al. 2001) and an

agreed definition of focal and extensive capsular penetration is lacking (Bostwick et al.

2000)

The commonest site of capsular penetration is ir

bundles, posterolateral to the prostate. The neurovascular bundles contain the

parasympathetic nerves to the corpora and lie outside the fibromuscular condensation

or ‘capsule’ of the prostate. Catalona and Bigg reported that all men who had

extracapsular extension in the region of the neurovascular bundles also had positive

surgical margins (Catalona & Bigg 1990). In men with induration in or around the

lateral pedicle, wide excision of the neurovascular bundle on that side did improve the

chances of negative surgical margins and decrease the risk of disease recurrence;

although using this criteria 30% of patients would have their neurovascular bundles

excised unnecessarily as the cancer is organ confined (Graefen et al. 1998). Pound et al.

compared the actuarial recurrence-free probabilities according to tumour stage and

margin status in men potent and impotent postoperatively (Pound et al. 1997). The

groups were similar for age, Gleason grade and stage. With all stages there was no

difference in actuarial recurrence-free probability between the potent and impotent

men. In appropriately selected patients nerve sparing can be performed without

compromising cancer outcomes (Pound et al. 1997; Ward et al. 2004).

The pathological definition of seminal vesicle invasion is not standardised. Some report

tumour in the peri-seminal vesicle fat as seminal vesicle invasion whilst others equate

50 invasion only once tumour involves the muscle wall, most reports do not stipulate which definition was used. This has led to variation in the prognostic importance of

seminal vesicle invasion. Most studies demonstrate only lymph node metastases exceeds the adverse prognosis associated with seminal vesicle invasion (D'Amico et al.

1995; Pound et al. 1997; Quinn et al. 2001). Seminal vesicle invasion was found in 6-

29% of men in reported radical prostatectomy series (D'Amico et al. 1995; Linzer et al.

1996; Pound et al. 1997; Sofer et al. 2003; van den Ouden et al. 1997) and in 10% of

men in the UCL database (Bott et al. 2005c). The biochemical progression free rate is 5-

60% at 5 years (D'Amico et al. 1995; Pound et al. 1997; Sofer et al. 2003), the wide

range reflecting the differences in definition of invasion.

Lymph node metastases found at radical prostatectomy are associated with an

unfavourable prognosis (Catalona & Smith 1998; Epstein, Pizov, & Walsh 1993;

Gervasi et al. 1989; Hull et al. 2002; Quinn et al. 2001; Stamey et al. 1999),

independent of the volume of metastatic malignant tissue (Gervasi et al. 1989). In many

series using multivariate analysis lymph node metastases are the most significant

predictors of progression (Hull et al. 2002; Quinn et al. 2001), and almost all men found

to have nodal metastastese ultimately develop disease progression (Epstein, Pizov, &

Walsh 1993).

The TNM system classifies invasion into a neighbouring structure as stage T4 disease.

Invasion into the bladder is a rare event at radical prostatectomy and is associated with a

poor outcome, bladder neck involvement will be discussed as a positive surgical

margin.

51 1.8(c) Positive surgical margins

A positive surgical margin is defined as the presence of tumour at the inked surface of the resected specimen (Catalona & Smith 1998; Freedland et al. 2003; Ohori et al. 1995;

Stamey et al. 1990) and implies incomplete excision of malignant tissue. An average of

f those undergoing radical prostatectomy are found to have positive surgical margins, although later series have a lower margin positive rate (Bott & Kirby 2002).

Stamey et a l subdivided positive surgical margins into two groups (Stamey et al. 1990).

In the first group the cancer is cut through when it is outside the prostate boundaries

into the fat, having penetrated the capsule - a positive extraprostatic margin (Figure

1.4c). In the second group the cancer is cut through inside the glandular area of the prostate due to inadvertent surgical incision of the periprostatic fascia and prostate

capsule, which are missing from histological specimen - a positive intraprostatic or

iatrogenic margin (Figure 1.4b). The first group may be further subdivided into focal -

where limited tumour reaches the inked limit in one or two sites, or extensive - where

multiple positive margins were present at different sites in the prostate (Blute et al.

1998; Gleave et al. 1999). Epstein et al. demonstrated significantly different progression

rates between those with negative (Figure 1.4a) compared with equivocal

(intraprostatic positive margin), focal and extensive extraprostatic surgical margins

(Epstein, Pizov, & Walsh 1993). However, the sub-classification of positive margins

has not been standardised (Bostwick et al. 2000).

Several sites are designated with margin status, including the apex - the urethral limit, the base - which includes the bladder neck margin, vasal and circumferential - anterior,

lateral, rectal or posterior surface. The overall incidence of positive surgical margins is

very similar for each of the three approaches to radical prostatectomy - the retropubic, perineal and laparoscopic (transperitoneal and retroperitoneal). However, the location of the positive margin varies. The apex is the most common site for positive margins in the retropubic and both laparoscopic approaches, whilst the anterior prostate is the most common site for the perineal approach (Abbou et al. 2000; Barocas et al. 2001; Blute et al. 1997; Bott et al. 2005c; Bott & Kirby 2002; Cathelineau et al. 2004; Epstein, Walsh,

& Brendler 1994; Guillonneau & Vallancien 2000; Van Poppel et al. 1995; Watson,

Civantos, & Soloway 1996; Weldon et al. 1995). Table 1.6 shows the incidence and site of positive margins.

53 Table 1.6. The site of non-iatrogenic (extraprostatic) solitary positive surgical margins after radical prostatectomy(Abbou et al. 2000; Bott & Kirby 2002; Epstein, Walsh, & Brendler 1994; Guillonneau & Vallancien 2000; Van Poppel et al. 1995; Watson, Civantos, & Soloway 1996; Weldon et al. 1995)

No. of % site of solitary positive surgical margin Author Stage margins Apical Anterior Posterior Lateral Postero­ Bladder Other lateral neck

Radical retropubic prostatectomy

Epstein Tic 29 31 24 10 7 24 3 1994 shave

Watson T ic 17 59 6 18 18 1996 shave

Van T3 19 37 53 10 Poppel cone 1995 Bott Tlc-T3 131 79 16 2 2 2002 shave

Radical perineal prostatectomy

Weldon Tl-2 88 7 25 -- 16 7 45 1995 cone

Laparoscopic radical prostatectomy

Abbou T2-3 4 50 50 2000

(The cone and shave are two methods for deriving apical margin status. The cone technique involves amputating the distal 1.5cm of the prostate and the tissue is sectioned in a sagittal plane, with the urethra at the centre. Using the shave technique a section is taken for the most inferior part of the apicalparaffm block. The cone technique allows examination of the prostate at its most distal limit, although only 5pm for every 3-5mm section of the distal margin are examined histologically. The shave technique on the other hand is simpler and less time-consuming to perform and the entire surface area of the distal margin is examined).

54 Figure 1.4 a) Low power of radical prostatectomy wholemount section demonstrating organ confined prostate cancer (EPT = extra- prostatic tissue, E = edge of prostate, CAP = prostate cancer)

CAP

Figure 1.4 b) Low pow er of radical prostatectomy w holemount section demonstrating an intraprostatic positive surgical margin (CAP = nrostate cancer. IM = inked marpin'l

CAP

IM

55 Figure 1.4 c) Low power of radical prostatectomy w holemount section demonstrating extraprostatic carcinoma w ith a positive surgical margin (EPT = extra-prostatic tissue, E = edge of prostate, CAP = prostate cancer)

56 Patients with positive surgical margins are at a significant risk of biochemical relapse -

50-60% at 5 years, and subsequent clinical relapse, although by no means every patient will suffer eventual disease recurrence (Cheng et al. 1999; Epstein et al. 1996; Ohori et al. 1995; van den Ouden et al. 1993). Epstein et a l , reported 79% of men with negative margins were progression free over a 10-year period compared with 55% of those with positive margins (p<0.00001) (Epstein et al. 1996). The Johns Hopkins group subsequently demonstrated the effect of Gleason grade on outcome in men with positive margins. They reported that positive surgical margins had no impact on 10-year probability of biochemical recurrence in men with Gleason score less than 7. However, men who had a Gleason score of 7 and positive surgical margins did significantly worse than those with extracapsular extension and negative margins (Epstein et al. 1998).

Grossfeld et al. calculated that, after adjusting for PSA, pathological tumour stage and

Gleason grade, patients with positive margins were 2.6 times more likely to have disease recurrence than those with negative margins (Grossfeld et al. 2000a).

Furthermore patients with positive margins were significantly more likely to receive adjuvant or non-adjuvant secondary treatment (p=0.0001). Stamey et al. however contested these findings, stating that margin status is not an independent predictor of failure after radical prostatectomy when adjusting for the percent of Gleason 4 and 5 cancers, tumour volume and lymph node status (Stamey et al. 1999).

Cheng and co-workers examined the correlation between margin status and PSA relapse in a series of 377 patients from the Mayo Clinic (Cheng et al. 1999). Their overall margin positivity rate was 29%; 19% of patients had positive margins without extracapsular disease (intraprostatic), 14% had extracapsular extension with negative margins and 10% had both extracapsular extension and positive margins. Men with organ confined disease and negative margins had a 90% 5-year progression-free survival. In the cohort with positive margins the 5-year progression-free survival was 57 78% for those without extracapsular extension and 55% for those with extracapsular extension.

Other authors have shown intraprostatic positive margins have little or no effect on outcome, at least in the medium term (Blute et al. 1998; Ohori et al. 1995). Ohori et al. reported 100% of men with intraprostatic positive margins were free from tumour recurrence after 5 years compared with 42% with positive margins and extracapsular extension (Blute et al. 1998). Barocas et al. compared 91 cases who had isolated intraprostatic margins with three groups matched for age, Gleason grade, preoperative

PSA and clinical stage, at the Johns Hopkins (Barocas et al. 2001). Their follow-up was on average only 31 months, but they were able to show a 4.3% recurrence rate for men with organ-confined prostate cancer, 9.8% for specimen confined, margin negative cases, 10.9% for men with a positive intraprostatic surgical margin, but otherwise organ confined disease, and a 25% relapse rate for men with extracapsular disease and positive surgical margins. They concluded, from this small series, that inadvertent capsular incision, resulting in a positive intraprostatic limit, gave similar PSA outcomes to specimen-confined disease.

Features including the location, extent and number of positive margins may have an impact on disease recurrence (Table 1.7) (Blute et al. 1997; Obek et al. 1999). From

The UCL database the positive margins of 347 patients operated on by a single surgeon were examined (Bott et al. 2003d). Apical margins were associated with a lower incidence of PSA recurrence than other sites after a median 24 months (range 1 -89).

Solitary margins had lower recurrence rates (22%) than multiple positive margins

(53%). This is in agreement with the literature. Blute et al. reported that the site of positive margins was a significant predictor of progression (Blute et al. 1997). Patients with pT2N0 disease who had a single margin at either the apex/urethra or anterior/posterior prostate or multiple positive margins in these sites had only slightly 58 decreased clinical recurrence or PSA failure rates compared with patients with negative margins at 5 years (79%, 78%, 68% compared with 86% with negative margins).

However, patients in whom the prostate base limit was positive had significantly lower clinical or PSA failure free rates - 56% at 5 years. The risk of PSA progression was 1.68 times higher in men with positive margins after matching for Gleason score, pre­ operative PSA and DNA ploidy in this series.

Table 1.7. Biochemical recurrence and positive surgical margins

Margin status PSA recurrence Bott 2003 Blute 1997 Obek 1999 AU margins- 14% 14% 7% Single apical + 22% 21% *29% Any single* 24% 23% 24% Multiple+ 53% 32% (2) 43% Number n=347 n=2334 n=495 Follow-up (snaths) 24 60 25

(Blute et al. 1997; Bott et al. 2003d; Obek et al. 1999)

* apical +/- other positive margins

Obek et al. reported similar findings from their series from Miami of 495 men with clinical T1-T3 disease who underwent radical prostatectomy (Obek et al. 1999). After a mean follow-up of 25.3 months their recurrence rate was significantly higher in men with positive compared with negative surgical margins, 27.8% versus 6.9% respectively, despite a quarter of margin positive cases receiving adjuvant androgen deprivation. These were not isolated margins and patients with other adverse pathological features, such as seminal vesicle and lymph node involvement were

59 included. However, they concluded that time to recurrence was shorter in patients older than 70, with a Gleason score of 7 or more, seminal vesicle invasion, multiple positive margins and a positive margin at the bladder neck or the posterolateral surface of the prostate. Of men with multiple positive margins 43% developed disease recurrence, compared with 24% of men with a single positive margin.

Solitary bladder neck margins occur in less than 5% of radical prostatectomy specimens

(Soloway & Neulander 2000; Watson, Civantos, & Soloway 1996) and in the UCL radical prostatectomy database a bladder neck margin was the solitary positive margin in 10 (1%) of 968 cases (Bott et al. 2005c). Bladder neck involvement is almost invariably associated with adverse prognostic factors including large tumour volume, high preoperative PSA, high Gleason score, seminal vesicle involvement and lymph node metastases (Wieder & Soloway 1998).

Positive vas deferens margins are also associated with a poor prognosis. From the UCL

Radical Prostatectomy series of 1001 men 16 (1.5%) had positive vasal margins, all these men also had positive seminal vesicles. In a series of 105 men undergoing radical prostatectomy 7 (6.7%) cases had vasal involvement; in these patients there was a significant correlation with seminal vesicle involvement, extracapsular extension, extensive carcinoma, Gleason score 7-10 and positive bladder neck margins. In this small series all five men with follow-up developed disease recurrence (Billis A, Freitas

LLL, & Magna LA 2002).

Several groups have highlighted the prognostic difference between a focal compared with an extensive positive margin (Babaian et al. 2001; Epstein et al. 1996; Wieder &

Soloway 1998). They reported focal margins were associated with recurrence in 40% and extensive margins had a 65% risk of progression after 5-years follow-up (Epstein et

60 al. 1996). In summary the site, number and extent of positive surgical margins all provide valuable prognostic information.

1.8(d) Tumour volume

Of the 998 cases with tumour volume measurements in the UCL database the median tumour volume was 2 ml, ranging from atypical acini (too small to measure) to 31 ml

(Bott et al. 2005c). Some authors report tumour volume to be a significant independent predictor of disease relapse (Humphrey & Vollmer 1997; Renshaw et al. 1999; Stamey et al. 2000), though others contest these findings (Epstein et al. 1993; Salomon et al.

2003). This may reflect the different zone of origin of the tumour. Tumours arising in the transition zone tend to be larger (Stamey, Dietrick, & Issa 1993) but of lower grade and have lower risk of progression than peripheral zone tumours (Grignon & Sakr 1994;

Noguchi et al. 2000). In studies where cancers of both peripheral and transition zone tumours are included, the tumour volumes may not be significant. In the one study where only the prognostic features of peripheral zone tumours were examined the tumour volume was an independent predictor of disease relapse (Stamey et al. 2000).

PSA concentrations correlate significantly with volume of prostate cancer in radical prostatectomy specimens (Stamey et al. 1989), it may be that in multivariate analysis involving PSA and tumour volume, tumour volume is not an independent variable.

Tumour volume remains a category 2 prognostic factor.

1.8(e) Lymphovascular invasion

Overall lymphovascular invasion is reported in 5.2-13.2% of cases (Quinn et al. 2001;

Stamey et al. 2000; van den Ouden et al. 1997) and 9.1% in the UCL series. It is

strongly associated with Gleason score, pT stage, tumour volume and lymph node status

(p<0.001) (McNeal & Yemoto 1996). Vascular invasion has been found to be both the

61 most significant risk factor (van den Ouden et al. 1997) and a significant but not an independent variable (Babaian et al. 2001; Stamey et al. 2000) for disease progression following radical prostatectomy. Many large series do not report invasion into the lymphatic of vascular spaces, further reports are required before this becomes a recognised independent prognostic feature.

1.8(f) DNA ploidy

Benign prostate cells and tumours of low stage and grade are diploid, as assessed by static or flow cytometry. As prostate cancer progresses several copies of the DNA within a single cell may be found - aneuploid. Patients with aneuploid tumours have reduced survival and several authors report that DNA ploidy is an independent predictor of progression (Adolfsson 1994; Shankey et al. 1993). Furthermore, DNA ploidy status at biopsy can be used to predict ploidy status at radical prostatectomy (Ross et al. 1999).

Despite these findings DNA ploidy remains a Category 2 prognostic factor, probably as the technology to assess ploidy status is not widespread and so multi-centre confirmation of its prognostic value is lacking.

62 1.9 Neoadjuvant androgen deprivation

Neoadjuvant androgen deprivation has been shown to reduce the incidence of positive surgical margins and extracapsular disease in a number of studies (Goldenberg et al.

1996; Lee, Warde, & Jewett 1999; Montironi et al. 1999; Soloway et al. 1995). This may be because the whole prostate shrinks allowing wider resection margins, down- staging of the tumour itself or because of difficulties with accurate pathological assessment after hormonal treatment (Wieder & Soloway 1998). However, there is no evidence from long-term prospective randomised trials that neoadjuvant androgen deprivation has any effect on biochemical relapse or survival (Goldenberg & et al 1997;

Lee, Warde, & Jewett 1999; Soloway & et al 1997). Furthermore difficulties may arise giving the prostate specimen a reliable Gleason grade as androgen withdrawal affects the glandular architecture, so important prognostic information may be lost. Given the current data the routine use of neoadjuvant androgen deprivation before radical prostatectomy is unjustified (Lee, Warde, & Jewett 1999).

63 1.10 Management of recurrent disease after radical prostatectomy

The incidence of prostatic cancer has remained fairly static over the last 8-10 years while the proportion of men with recurrent disease after radical treatment has fallen.

The advent of PSA testing in the late 1980s has meant more men are diagnosed and treated with lower stage disease, both in the US (Han et al. 2001) and in the UK (Bott et al. 2005c). The proportion of men from the radical prostatectomy databases held at

University College London with organ confined disease rose significantly, from 30% between 1988 and 1991 to over 50% in 2001(p<0.001) (Bott et al. 2005c). Despite this

‘stage migration’, 27-53% of men will develop prostate specific antigen (PSA) recurrence up to 10 years after radical prostatectomy (Hull et al. 2002; Pound et al.

1997; Quinn et al. 2001; Trapasso et al. 1994; Zincke et al. 1994) and 16-35% of patients receive second-line treatment within 5 years of surgery (Fowler, Jr. et al. 1993;

Lu-Yao et al. 1996). After radical prostatectomy half the men treated for recurrent disease receive external beam radiotherapy (EBRT) to the prostatic bed, with curative intent. In contrast, over 90% of men with relapse following radical radiotherapy receive palliative hormonal manipulation. This reflects the limited therapeutic options ^ patients who have had radical radiotherapy or brachytherapy, though this may with the increasing role of salvage cryotherapy, salvage brachytherapy and in some cases salvage prostatectomy. n ^ . AW?*

64 1.10(a) Definition and timing of relapse

Following surgery the PSA should fall to undetectable levels within 4 weeks, given the half-life of PSA is 3.15 days (Oesterling et al. 1988). A PSA nadir that is not

‘undetectable' represents either residual benign tissue or local/metastatic malignant prostatic tissue. A rising PSA following a detectable nadir is likely to represent metastatic disease.

There is no universally agreed PSA cut-off to define ‘recurrent disease’ following an undetectable PSA nadir (Bott 2004). This is because different assays have differing sensitivities and the contribution of residual benign glands to a postoperative PSA remains unresolved. The conventional assay, the Hybritech (Tandom-R), has a detection threshold of O.lng/ml. Using this assay, Epstein et al. proposed a cut-off as >0.2ng/ml, given that 53% of men with this PSA value eventually develop clinical progression

(Epstein, Pizov, & Walsh 1993). Lange et al. proposed a higher cut-off of 0.4ng/ml as all men with a postoperative PSA at this level had detectable disease progression within

49 months of surgery (Lange et al. 1989). The introduction of ultra-sensitive assays which are able to detect PSA to O.Olng/ml has meant men being diagnosed and treated for recurrence before their PSA reaches O.lng/ml (Doherty et al. 2000). While at this level potentially curative treatment is expedited there is at least a theoretical risk of treating residual benign glands only.

The seminal paper by Pound et al. demonstrated the extended period between biochemical failure after RP and the development of clinical progression (Pound et al.

1999). The median time for the 315 of 1,997 men to develop PSA relapse, defined as a serum value of at least 0.2ng/ml, was 2.3 years. Of the 34% of these men who subsequently developed metastases, the median actuarial time to do so was 5 years from

65 the time of PSA progression. Once men developed metastases the median time to death was just under 5 years (Pound et al. 1999). The probability and time to develop metastases was dependent on the Gleason score (5-7 or 8-10), the time to biochemical progression (>2 or <2 years) and the PSA doubling time (>10 or <10months). Using these variables ‘Pound tables’ have been constructed to predict the likelihood of clinical disease progression in men with biochemical recurrence after radical prostatectomy.

1.10(b) Site of relapse

The decision of what and when to treat depends on whether disease progression is confined locally to the prostatic fossa or whether distant spread has occurred. Local relapse alone occurs in 50% of cases and metastases with or without local recurrence in

50% (Pound et al. 1997). Clinical and imaging modalities currently available are not of , sufficient sensitivity to detect low volume locally recurrent disease, th e digital rectal examination is not useful if men have an undetectable PSA (Pound et al. 1999) and is only sensitive once the PSA has been elevated for a median 2.3 years after surgery (Bott

2004). Less than 5% of bone-scans are positive until the PSA value is above 40-45ng/ml

(Cher et al. 1998a). CT scans are not sufficiently sensitive at detecting local recurrence until the PSA is increasing by a rate of more than 20ng/ml per year (Johnstone et al.

1997). The sensitivity and specificity of MRI is improving, though remains most sensitive at detecting nodal and bony metastases (Harisinghani et al. 2003; Hricak et al.

2003). The role of evolving technologies such as 111 Indium capromab and PET scanning in the diagnosis of recurrent disease has yet to be established. ^

The use of transrectal ultrasound (TRUS) guided anastomotic biopsies remains controversial. In a series from the University of California 54% of 114 men with persistently elevated PSA levels undergoing TRUS and biopsy had confirmed local recurrence, though a third of men required more than one biopsy session (Connolly et al. 1996). In this study the median PSA level of men with local recurrence was 5.7ng/ml

(range 0.2-35). Saleem et a l reported none of 11 patients with normal DRE and a PSA

<0.5ng/ml had a positive biopsy (Saleem et al. 1998). Shekarritz et al found only 25% of patients with a PSA of <1 .Ong/ml had a positive biopsy compared with 71% of those with a PSA >1 .Ong/ml (Shekarriz et al. 1999). These studies highlight the lack of

sensitivity of TRUS and anastomotic biopsies at low PSA values, salvage radiotherapy

is most effective while the PSA is still low. Scattoni et al reported men undergoing 6-8

anastomotic biopsies following a PSA recurrence or abnormality on DRE (Scattoni et

al. 2003). No patient had evidence of metastatic disease on bone-scan or abdominal CT.

In 68 men with a PSA 1 .Ong/ml or below, 45% had biopsies containing tumour. From

our own series of anastomotic biopsies 33% of 12 cases were positive for cancer

recurrence, in a further 5 patients with negative biopsies the PSA stabilised after

prostate bed radiotherapy, suggesting a high false negative rate (Foley et al. 2002). In

addition, in 5 patients salvage radiotherapy for local recurrence was delayed by a benign

anastomotic biopsy. Anastomotic biopsies are usually unwarranted, a positive result

does not exclude metastatic disease and a negative result does not exclude local

recurrence. Furthermore, a positive anastomotic biopsy is not associated with improved

outcome after salvage radiotherapy (Koppie et al. 2001).

The best predictors of local versus metastatic recurrence are the pathological findings of

the radical prostatectomy specimen and the post-operative PSA parameters. Patients

with Gleason grade >8 disease, positive seminal vesicles or lymph node metastases have

a 95%, 86% and 100% likelihood of developing metastatic disease, respectively (Han et

al. 2003). Conversely, biochemical relapse is more likely due to local failure alone if

there are multiple positive surgical margins, extensive extracapsular disease, Gleason

67 score <7 and an absence of nodal or seminal vesicle involvement (Anscher & Prosnitz

1991; Connolly et al. 1996; Kupelian et al. 1996; Nudell et al. 1999).

The PSADT is a useful marker of local versus distant recurrence after radical prostatectomy (Partin et al. 1994; Patel et al. 1997; Pound et al. 1999). Patel et al. reported a PSADT of <6months was indicative of metastatic disease in 77 men (Patel et al. 1997). Pound used a PSADT cut-off of <10months to indicate distant spread (Pound et al. 1999). Another study found the median PSA doubling time was 4.3 months for men who were ultimately found to have metastatic disease compared with 11.7 months for men with local failure alone (Trapasso et al. 1994). The problem lies in the men at an intermediate risk of metastatic disease. For these men nomograms, artificial neural networks or biostatistical equations incorporating the pathological and PSA features, and in some cases race, may be used to predict disease failure (Diblasio & Kattan 2003;

Moul et al. 2001; Partin et al. 1995). While these are cumbersome to use in the clinic, some are now available on the internet and may facilitate the management of intermediate risk patients.

Lack of a PSA nadir below undetectable levels and PSA recurrence within a year of surgery all reduce the likelihood of remaining PSA free after salvage radiotherapy

(Cadeddu et al. 1998). In general, men who develop PSA recurrence within 2 years of surgery, have tumour grade greater than 7, higher pathological stage, positive seminal vesicles or lymph node involvement are more likely to harbour metastatic disease and are better suited to systemic treatment (Han et al. 2003; Pound et al. 1999). Table 1.8 shows the Johns Hopkins data comparing the pathological and PSA parameters in men who developed local versus metastatic disease relapse.

68 Table 1.8. Comparison of Gleason score, pathological stage and timing of PSA recurrence by site of recurrence after anatomic radical prostatectomy (Pound et al. 1997) ______Variable Local recurrence Distant metastases +/- local recurrence Number of patients 41(34% 88(66%)

Gleason score 2-4 0% 0% 5-6 55% 45% 7 39% 61% 8-10 11% 89%

Pathological stage Organ confined 40% 60% Capsular penetration, negative 54% 46% surgical margins Capsular penetration, positive 48% 52% surgical margins Seminal vesicle involvement 16% 84% Lymph node metastases 1% 93%

Timing of PSA recurrence In 1 year 1% 93% Within 1-2 years 10% 90% After year 2 61% 39% After year 3 74% 26% Reproduced with kind permission from A W Partin

69 1.10(c) Treatment of relapse

The main management options available in patients who have biochemical failure after radical prostatectomy include active surveillance, prostate bed EBRT and hormonal manipulation. Attempts are being made to use salvage brachytherapy (Losa et al. 2003), and high intensity focused ultrasound (Kiel, Wieland, & Rossler 2000) in the management of local recurrence, though currently these techniques remain experimental. The decision on whether to treat or observe a patient with PSA only recurrence depends on the risks of clinical failure, perhaps aided by the Pound tables or biostatistical equations outlined above. In an American Urological Association questionnaire undertaken in 1998 only 13% of urologists said they would offer salvage radiotherapy following an asymptomatic PSA rise, 54% preferred observation and 31% opted for hormonal manipulation (Omstein et al. 1998). This is because the role of external beam radiotherapy in men with PSA recurrence after radical prostatectomy remains unresolved and the subject of ongoing research. To date no randomised study has been reported and small retrospective case series with limited follow-up have been conflicting. This is in part due to variation in patient selection, disparity in the definition of biochemical failure and differences in the use of adjuvant hormonal manipulation.

Studies reported highlight predictive factors that determine biochemical recurrence-free rates after salvage radiotherapy. Essentially where there is a high risk of metastases the likelihood of curative salvage treatment is reduced. The timing and dose of radiotherapy also impacts on PSA free survival.

The importance of the radical prostatectomy Gleason grade in men undergoing salvage radiotherapy was demonstrated in several studies (Cadeddu et al. 1998; Peyromaure et al. 2003; Tsien et al. 2003). In 57 patients, reported from the University of Michigan,

70 with a median follow-up of 7 years the 5 and 8-year biochemical disease-free survival was 58% and 37%, respectively (Tsien et al. 2003). The only significant risk factor for progression was Gleason score. At 8 years biochemical-free survival for Gleason 2-6, 7.

8-10 was 48%, 37% and 0%, respectively (Tsien et al. 2003). Cadeddu et al. reported no patients with Gleason score 9 or more, seminal vesicle invasion or lymph node metastases were PSA free 2 years after salvage radiation (Cadeddu et al. 1998).

The number of adverse risk factors also impacts on PSA free survival. In 115 men receiving salvage radiotherapy for PSA recurrence or clinical local recurrence 70 patients had radiotherapy and 45 men had neoadjuvant hormonal treatment before radiotherapy (Katz et al. 2003). The 4-year PSA relapse-free survival was 77% if men had positive surgical margins and no poor prognostic features, compared with 20% if they had one or more poor prognostic feature including seminal vesicle involvement, a pre-radiotherapy PSA >0.6ng/pl or extracapsular disease. In patients receiving radiotherapy alone, 39% were recurrence-free compared with 59% for those on combined radiotherapy and neoadjuvant treatment. Other studies have failed to show an improvement in PSA outcome with neoadjuvant treatment (Katz et al. 2003). The role of adjuvant hormonal manipulation in salvage radiotherapy after radical prostatectomy remains unresolved. This issue has been addressed in a randomised study: the Radiation

Therapy Group (RTOG) 9601. This study is awaiting further follow-up before reporting.

Patients who receive radiotherapy after surgery fare better if the PSA falls to an undetectable level than those with a persistently detectable PSA (66% at 40 months compared with 20% at 12 months, respectively)(Coetzee, Hars, & Paulson 1996). What is more, men in whom the PSA rises more than one year after surgery have a better

71 response to salvage EBRT (Cadeddu et al. 1998; Coetzee, Hars, & Paulson 1996;

Grossfeld et al. 2000b).

The PSA level at which salvage treatment is given is critical in terms of radiotherapy efficacy. The American Society of Therapeutic Radiology and Oncology (ASTRO)

Consensus Panel recommended salvage EBRT be given when the PSA level is less than

1.5ng/ml (Cox et al. 1999). Nudell et al. reported a cohort of 105 men treated with either adjuvant or salvage EBRT (Nudell et al. 1999). He defined disease free as the achievement and maintenance of a PSA of <0.2ng/ml; giving a 5-year overall disease- free survival of 43%. Outcomes were equivalent for the adjuvant and salvage EBRT groups where therapeutic irradiation was administered when the serum PSA was

<1.Ong/ml (Figure 1.5). In 27 men receiving 60-67 Gy, Schild et al reported 3-year progression-free survival of 76% with pre-radiotherapy PSA

26% for pre-treatment PSA levels >l.lng/m l (Schild et al. 1994). Peschel et al. reported a series of 52 men who had either adjuvant or salvage EBRT. He found the pre­ operative PSA level, the pre-radiotherapy PSA and seminal vesicle involvement were significant risk factors for actuarial biochemical recurrence following post-operative radiotherapy. However, the risk of recurrence was most significant for a pre­ radiotherapy PSA of <0.3ng/ml (Peschel et al. 2000). The PSA cut-off to trigger salvage

EBRT is variable; initiating salvage radiotherapy with a PSA <1.Ong/ml is likely to confer the best chance of biochemical survival. More data will be available on timing of salvage radiotherapy when two prospective randomised studies, European Organisation for Research and Treatment of Cancer (EORTC) 30943 and South West Oncology

Group (SWOG) 8974 report.

72 Q) 0 8

00 0 .6

Mos.: 0 12 24 36 48 60 L No at Risk: — Adjuvant RT: 36 26 19 12 10 6 — Non-Adj. RT PSA < 1: 38 29 23 10 4 2 ------Non-Adj. RT PSA > 1: 30 13 10 6 6 4

12 24 36 48 60 Months From End of RT

Figure 1.5 Time to develop PSA recurrence (PSA >2) after adjuvant radiotherapy

(red) and salvage radiotherapy with a pre-treatment PSA <1.0ng/ml (green) and

>1.Ong/ml (blue).

The dose of radiation is important in terms of biochemical disease-free survival after salvage therapy. Several studies have shown a dose of >64.8Gy is a significant prognostic factor for biochemical recurrence-free survival (Anscher, Clough, & Dodge

2000; Valicenti et al. 1999). Valicenti et a l reported a 3-year biochemical disease-free survival of 52% for salvage EBRT in high-risk patients, if men received a dose of

>64.8Gy compared with 18% for those receiving <64.8Gy. In a study from Duke

University 89 patients received salvage EBRT for persistently elevated or delayed PSA

73 recurrence after RP, 72% had an undetectable PSA response, with a 4-year disease-free survival of 53%. The only significant prognostic factor for biochemical-free survival in multivariate analysis was radiation dose >65Gy. Following these and other studies the

1999 ASTRO consensus statement recommended a salvage radiotherapy dose of at least

64.8Gy (Cox et al. 1999).

Salvage radiotherapy is usually well tolerated and toxicity has been markedly reduced by 3-D conformal techniques. Although there are no prospective quality of life studies reported, several studies demonstrate significant late genitourinary (GU) or rectal toxicity to be less than 10%. Using 3-D conformal EBRT the groups from Michigan and

Memorial Sloane-Kettering found grade >2 late GU toxicity of 6.3-10% and grade >2 rectal toxicity of 8.9-12%, respectively (Katz et al. 2003; Tsien et al. 2003). The figures for the Michigan series were based on physician reporting and may underestimate the actual complication rate.

Hormonal manipulation is used frequently in men with presumed or confirmed metastatic disease after radical prostatectomy, though there is scant data to support this therapy in men with PSA only recurrence. No randomised study has been reported on the use of hormonal manipulation in lymph node negative men with biochemical failure after radical prostatectomy. The PSA level at which to institute therapy is unknown.

Extrapolating from the major studies on hormonal therapy in patients with advanced disease, hormonal therapy either alone or as maximum androgen blockade, appears to have a small, but significant survival benefit, but more importantly reduces the risk of serious complications from disease progression (Bennett et al. 1999; Medical Research

Council Prostate Cancer Working Party Investigators Group 1997; Messing et al. 1999).

The Medical Research Council (MRC) study demonstrated modest improvements in

74 cancer-specific survival in the immediate versus delayed hormonal therapy men with

MO. MX or Ml disease (Medical Research Council Prostate Cancer Working Party

Investigators Group 1997). The greatest benefit was seen in men with MO disease - which may most closely reflect the post radical prostatectomy patients with biochemical recurrence alone.

The Casodex Early Prostate Cancer Program (EPC) is addressing the question of whether hormonal manipulation reduces the risk of recurrent disease. This has reported a relative reduction of 42% in clinical disease progression in all men with localised or locally advanced prostate cancer (Tl-4, NO/l/x, MO), treated with bicalutamide 150mg

(Casodex, AstraZeneca) in conjunction with their standard treatment compared with standard treatment alone (9% versus 13.8%, respectively) (See et al. 2002). This was irrespective of stage or primary treatment, although the North American arm (study 23), which consisted of a greater proportion of radical prostatectomy patients, failed to show such significant differences in progression rates. This data is still immature, with a median treatment time of 2 years (range 1.8-2.5 years), unsurprisingly most men have responded to hormonal manipulation, whether this translates to an increase in survival as men inevitably become hormone refractory remains to be seen. A recent study comparing flutamide, another non-steroidal antiandrogen, versus no adjuvant treatment also found a reduction in rates of disease progression after radical prostatectomy.

However, following a median follow-up of 6.1 years there were no differences in terms of overall survival and considerable toxicity was reported in the flutamide arm.

Bicalutamide is well tolerated in advanced prostate cancer (Anderson 2003; Kolvenbag,

Iversen, & Newling 2001). Whether the favourable toxicity profile results in fewer withdrawals and enhanced efficacy remains to be seen.

75 The role of maximum androgen blockade or intermittent hormonal therapy remains contentious. The results of further studies, including the SWOG, NCI Canada and the

EORTC GU-Group (NCI-INT 9346/EORTC 30958), are required before these potentially beneficial developments become part of routine clinical practice.

The management of recurrent disease after radical prostatectomy will be greatly assisted by the publication of ongoing randomised studies. Still a significant proportion of men undergoing surgery develop PSA relapse and many of these men may be cured by second line therapies. Defining the population most likely to harbour local recurrence only is critical in deciding management strategies. Men with moderate grade disease

(Gleason <7), who do not have nodal or seminal vesicle involvement, and develop a

PSA relapse more than 2 years after surgery with a doubling time of >lyear and whose pre-treatment PSA is <1.5ng/ml are most likely to benefit from local treatment with salvage radiotherapy. Men with one or more poor prognostic feature may be offered hormonal manipulation, the PSA level to initiate this treatment remains unknown. In all men a balance must be made between exposing men to a treatment with little proven survival benefit, but with associated unwanted effects and the patient’s quality of life.

76 1.11 Molecular changes in prostate cancer

Prostate cancer, like all other cancers, develops and progresses as a consequence of an accumulation of genetic changes. While several putative genes have been isolated for the development of breast, ovarian ( BRCA1. BRCA2 ) and colon cancer (hMLHl,

hMSH2 ), the aetiology and pathogenesis of prostate cancer remains poorly understood.

However with advances in molecular genetics it will be the analysis of the genetic code

and its products that will, in the future, provide both prognostic information and targets

for therapeutic intervention.

1.11(a) Background

Under normal conditions every cell in the body (except red blood corpuscles) has the

same DNA, but perform a range of functions. Abnormal changes in this DNA affect the

expression and function of critical genes. Alterations in genes can result in altered

mRNA expression and often, altered proteins. Although many changes may occur, 5-10

specific alterations are required to convert a normal prostate epithelial cell into a

malignant cell capable of invasion (Elo & Visakorpi 2001). In a small minority of cases

these alterations are inherited, in the majority however they are acquired or somatic

mutations (Fearon & Vogelstein 1990; Kallioniemi & Visakorpi 1996). The problem for

researchers is to identify: which genetic changes are important in cancer progression,

which can be used to predict prognosis and which can act as targets for therapeutic

intervention.

l.ll(a)l. Oncogenes

A proto- is a normal regulatory gene. If its activity is increased as a

consequence of genetic alteration it becomes an oncogene. Oncogenes are described as

77 dominant as only one allele needs to be changed for a biological effect to occur.

Modification of the proto-oncogene to form an oncogene leads to a gain in function; this may be due to a qualitative or quantitative change in protein product. Alteration in the regulatory sequence, for example by amplification, brings about quantitative change in a normal product. Qualitative change generates abnormal protein product as a result of mutation in the coding region or by rearrangement (translocation) of two genes resulting in production of a fusion protein made up of parts of each participating gene.

Oncogenes implicated in prostate cancer include the RAS oncogenes, MYC, ERBB2,

PSCA, EIF3S3 and the AR gene. The RAS oncogene is activated by point mutation, the

MYC gene is activated by amplification and ERBB2 is either amplified or modified through transcriptional or post-transcriptional deregulation. Amplification occurs relatively late in the development of prostate cancer; it is found in metastatic {MYC,

PSCA, EIF3S3) and hormone refractory disease (AR). Altered regulation may be seen in the early as well as the later stages of cancer development.

l.ll(a)2. Tumour suppressor genes

A tumour suppressor gene (TSG) encodes for an inhibitory protein whose function is lost in cancer. TSGs are inactivated by deletion of one or both genes or by a point mutation in one gene and deletion in the other. Three types of TSG have been described.

Firstly, recessive TSGs, these are inactivated only if both gene copies of a diploid cell are lost e.g. RB. Secondly, the dominant negative TSG where a mutation in one allele generates an abnormal protein that inactivates the normal product from the unaffected allele. p53 may be inactivated in this manner. Finally, haplo-insufficiency is where inactivation of one allele copy is sufficient to change the phenotype, for example PTEN.

78 The majority of TSGs code for proteins that complex with other effector proteins and inhibit their actions. Where an individual has a hereditary mutation of one TSG they have an increased susceptibility to develop cancer and for developing cancer at a younger age.

TSGs may be inactivated by mutation, deletion or by abnormal methylation.

Methylation changes including hypermethylation, demethylation and redistribution of methyl groups have all been reported in prostate cancer (Ruijter et al. 1999). The significance of demethylation in prostate cancer is unknown (Bedford & van Helden

1987), however hypermethylation is of importance. The promotor and exons at the 5’ end of some recognised TSG contain a high density of C-G (cytosine-guanine) repeats, called CpG islands. Hypermethylation of these CpG islands is believed to result in transcriptional inactivation of the associated gene and is found in malignant tissues.

79 1.11(b) Hereditary prostate cancer

In 1895, Dr Alfred Warthin’s seamstress told him she was convinced she would die from cancer as so many of her family had. Indeed she died at a young age from endometrial cancer; this prompted Warthin to publish a description of her family, which included relatives dying prematurely from gastric, endometrial and colorectal cancers

(Wartin 1913). Progress was made in understanding familial cancers in the 1960’s by

Lynch (Lynch et al. 1966), Li (Li & Fraumeni, Jr. 1969) and Knudson (Knudson, Jr.

1971) who systematically examined at risk families. Knudson went on to predict the

‘two-hit’ hypothesis based on his study of a rare eye tumour in children - retinoblastoma. He postulated the presence of regulatory TSGs that are normally involved in cell-cycle, apoptosis and proliferation. Humans normally have two copies of these TSG, one from each parent. Where inherited mutations occur one copy is affected.

The remaining TSG maintains sufficient control of its regulatory function, however should a mutation occur in the other copy of the TSG in one cell a malignant potential is acquired. This hypothesis was confirmed 20 years later when germ-line mutations in the

RB gene were detected in hereditary retinoblastoma (Hogg et al. 1993).

The problem with studying familial prostate cancer is that as the incidence of the disease in the general population is so high there is a high rate of sporadic cases amongst the familial cluster. Furthermore, these families undergo PSA screening which will further enhance the detection rate. Despite these factors, a family history of prostate cancer still confers an increased risk of being diagnosed with prostate cancer and of developing the disease at an earlier age. A man with a first-degree relative with prostate cancer has a relative risk of 2.0 (1.2-3.3, 95%CI), with a second-degree relative risk of

1.7 (1.0-2.9) and with both a first and second degree relative an 8.8 (2.8-28.1) relative risk of developing prostate cancer (Steinberg et al. 1990).

80 Further evidence for a hereditary component and a link with another histologically similar malignancy is that prostate cancer is higher in men who have relatives with breast cancer (Thiessen 1974; Tulinius et al. 1992). Likewise, the risk of breast cancer is doubled in families with a history of prostate cancer (Anderson & Badzioch 1992;

Sellers et al. 1994).

Hereditary prostate cancer is usually defined by the pedigree as no associated genes have yet been firmly identified. This definition therefore includes nuclear families with three cases of prostate cancer, the presence of prostate cancer in each of three generations in the maternal or paternal lineage and a cluster of two relatives diagnosed with prostate cancer before the age of 55 years (Carter et al. 1993). Using this definition

3-5% of patients with prostate cancer has hereditary disease. This definition however excludes families with a hereditary susceptibility to prostate cancer (with mutations in autosomal dominant susceptibility genes), so the true proportion is likely to be 5-10%.

1.1 l(b)l. Candidate genes

Prostate cancer susceptibility loci have been reported at lq24-25, lp36, lq42, 20ql3,

Xq27-28 (Ostrander & Stanford 2000), 16q23 (Suarez et al. 2000) and 17p(Tavtigian et al. 2001) (Table 1.9). Several authors have reported a linkage (the inheritance of a recognised genetic marker with the disease) on the long arm of chromosome 1 (Cooney et al. 1997; Hsieh et al. 1997; Smith et al. 1996). The exact site of the susceptible genes on this chromosome region remains unclear; lq24-25, named HPC1 (Smith et al. 1996) and lq42, termed PCAP (Berthon et al. 1998) have been implicated. Two further studies confirmed linkage at the HPC1 locus on lq24-25 (Cooney et al. 1997; Hsieh et al.

1997), other studies did not (Berthon et al. 1998; Eeles et al. 1998; Mclndoe et al. 81 1997). In a study from the CRC/BPG UK Familial Prostate Cancer Study Collaborators the incidence of allelic imbalance at HPC1 was low in both sporadic tumours and small prostate cancer families (Dunsmuir et al. 1998). The authors concluded HPC1 is unlikely to be acting as a TSG in the development of familial or sporadic prostate cancer.

Other candidate genes include the breast cancer susceptibility genes BRCA1 (17q21) and BRCA2 (13ql 2.3) and these confer a relative risk of prostate cancer of 3.0 and 2.6-

7.0 respectively (Eeles & the UK Familial Prostate Study Group & the CRC/BPG UK

Familial Prostate Cancer Study Collaborators 1999). The UK/Canadian/Texan

Consortium found up to 30% of familial clusters may be linked to BRAC1/2, although the confidence intervals were wide and included zero (Eeles et al. 1998). Indeed the UK

Familial Prostate Cancer Study was unable to find any mutations in the BRAC1 gene but did find two germline mutations in BRCA2. They postulated the BRCA2 mutations may be a coincidental finding or may play a role in modifying the expression of another cancer gene (Eeles & the UK Familial Prostate Study Group & the CRC/BPG UK

Familial Prostate Cancer Study Collaborators 1999). While an association between breast and prostate cancer clearly exists, the molecular basis for this association is not yet fully understood.

Using segregation analysis, prostate cancer was found 1.5 times more often in brothers than in fathers of men with prostate cancer (Schaid & Rowland 1998). This may indicate that prostate cancer is inherited in some families in a recessive or an X-linked way. A study of 360 families in North America, Sweden and Finland strongly implicated a region on the long arm of the X- chromosome at Xq27-28, which the

82 authors named HPCX (Xu et al. 1998). This gene accounted for 15-16% of cases in

North America and 41% of Finnish cases of hereditary prostate cancer.

From the Utah Population database workers mapped an inherited predisposition to prostate cancer to the ELAC2 gene on chromosome 17p in 33 families (Tavtigian et al.

2001). When the number of families studied increased to 127, the linkage was no longer significant as the prevalence of the mutation fell. The authors conclude this was due to the heterogeneity of the families involved. ELAC2 codes for a protein involved in interstrand-cross link repair mechanisms, which is consistent with a possible role in prostate cancer.

Table 1.9 Chromosomal loci reported to contain hereditary prostate cancer genes

Locus Putative gene Number of Two-point lod Multipoint lod

families score score

lq24-25 HPC1 91 3.7 5.4

lq42 PCAP 47 2.7 3.1

lp36 CAPB 12 3.7 2.2

17p ELAC2 33 4.5

20ql3 HPC20 162 2.7 3.0

Xq27-28 HPCX 360 4.6 3.9

(The lod score is a statistical estimate of whether two (two-point) or several (multipoint) loci are likely to lie near each other on a chromosome, and are therefore likely to be inherited together, a score over 3 is statistically significant.)

Confirmatory studies showing weak or no linkage to these regions and a putative susceptibility gene has yet to be identified. This has led investigators to conclude that

83 hereditary prostate cancer is a heterogeneous disease and no single gene is responsible for the high incidence in certain families. Alternatively, there may be no major susceptibility genes for prostate cancer as there is for say breast or colon cancer (Elo &

Visakorpi 2001; Ostrander & Stanford 2000).

1.1 l(b)2. Low penetrance polymorphisms

Mutations in high penetrance susceptibility genes resulting in a significantly increased risk of prostate cancer are relatively uncommon in prostate cancer. On the other hand polymorphisms (the occurrence of allelic variations) in low penetrance genes increase the risk of developing the disease only modestly, but occur with greater frequency in a population. Low penetrance polymorphisms may therefore have a greater impact on the frequency of prostate cancer in the population as a whole.

The polymorphic CAG (cytosine-adenosine-guanine) repeat in the androgen receptor gene has been studied extensively. This codes for a polyglutamine tract of varying lengths (dependent on the number of CAG repeats in the gene). Workers have shown an inverse relationship between the CAG repeat length and prostate cancer risk

(Chamberlain, Driver, & Miesfeld 1994; Kazemi-Esfaijani, Trifiro, & Pinsky 1995;

Stanford et al. 1997). In normal healthy males there are 13-30 CAG repeats, however

African-American men have a higher prevalence of <22 CAG repeats in the AR gene

(Irvine et al. 1995), and Chinese men have a higher prevalence of longer CAG repeats

(Hsing et al. 2000). The prevalence of prostate cancer is higher in African-Americans and lower in Chinese man (Elo & Visakorpi 2001). Giovannucci et a l showed that if an individual had <18 CAG repeats they had an increased relative risk of 1.52 for developing prostate cancer compared with men with >26 CAG repeats (Giovannucci et al. 1997).

84 Polymorphisms in the SRD5A2, the gene that codes for the enzyme 5 a-reductase type

II, have been reported (Elo & Visakorpi 2001). 5 a-reductase is responsible for the conversion of testosterone to its more active metabolite dihydrotestosterone in the prostate. Polymorphisms in the SRD5A2 gene result in increased catalytic activity of this enzyme and are associated with an increased risk of developing prostate cancer

(Jaffe et al. 2000).

The vitamin D receptor ( VDR) gene is also a polymorphic steroid hormone receptor gene implicated in prostate cancer. Vitamin D is antiproliferative to prostate cancer cell lines (Peehl et al. 1994; Skowronski, Peehl, & Feldman 1993) and low serum vitamin D may be associated with an increased risk of prostate cancer (Ahonen et al. 2000; Corder et al. 1993). Whether the VDR polymorphisms confer an increased susceptibility to prostate cancer remains contested (Blazer, III et al. 2000; Habuchi et al. 2000b; Ingles et al. 1998; Nelson & Wilding 2001; Taylor et al. 1996). However, a number of vitamin D analogues are currently undergoing clinical trials for prostate cancer treatment and prevention (Nelson & Wilding 2001).

Other polymorphisms reported to influence the risk of developing prostate cancer include the cytochrome P450 family genes (CYP3A4 and CYP17). The ELAC2 gene may be a low as well as possibly a high penetrance gene (Habuchi et al. 2000a; Rebbeck et al. 1998; Rebbeck et al. 2000). Larger studies are required to corroborate these findings.

85 1.11(c) Somatic changes

While some individuals develop tumours as a result of inherited genetic changes all tumours contain acquired or somatic alterations. Advances in the techniques used to examine these changes resulted in the identification of chromosomal regions deleted or amplified that may play a role in tumourigenesis. Chromosomal changes have been identified in specific stages of disease progression with the average number of alterations significantly higher in distant metastases than primary tumours (Alers et al.

2000).

One of the most widely reported techniques for examining chromosomal alterations is comparative genomic hybridisation (CGH). This allows the detection of DNA sequence copy number changes throughout the genome and can therefore identify regions where deleted TSG or amplified oncogenes may be harboured. Using CGH researchers have shown that locally recurrent hormone refractory prostate cancers contain almost four times as many alterations as untreated primary tumours (Nupponen & Visakorpi 1999).

The early development of prostate cancer is associated with inactivation of TSG, whereas later progression, including the development of hormone refrac associated with activation of oncogenes.

CGH and Al (allelic imbalance) studies have demonstrated the most common chromosomal aberrations in prostate cancer are deletions in chromosome regions 3p, 6q,

7q, 8p, 9p, lOq, 13q, 16q, 17q and 18q and gains in 7p, 7q, 8q and Xq (Alers et al. 2000;

Cooney et al. 1996b; Elo & Visakorpi 2001; Latil et al. 1997; Li et al. 1999).

Furthermore, the frequency of Al at several loci increases with higher grade (Hugel &

Wemert 1999; Leube et al. 2002) and higher stage disease (Hugel & Wemert 1999;

Perinchery et al. 1999).

86 1.11(c)!. Early prostate cancer

Genetic changes have been identified at all stages of prostate cancer development and progression. In this section we will examine the changes in chromosomal regions and genes identified in early, metastatic and hormone refractory prostate cancer (Bott et al.

2005b).

8p

Two of the most common deletions found using Al and CGH are on 8p and 13q

(Visakorpi et al. 1995b). Laboratory techniques including fluorescence in situ hybridisation (FISH) and Al have found that losses in these regions are frequently found in HGPIN (Emmert-Buck et al. 1995; Saric et al. 1999). It would appear that inactivation of one or more as yet unidentified TSG in 8p and 13q is an early event in prostate cancer pathogenesis.

At least two regions of loss on 8p have been identified 8pl2-21 (Haggman et al. 1997b) and 8p22 (Cher et al. 1996; Emmert-Buck et al. 1995; Oba et al. 2001). Loss of 8 p l2-21 is an early event in prostate , occurring in HGPIN and early invasive disease, whereas loss of 8p22 is a late event, found more commonly in advanced cancers. The NKX3.1 gene (Brothman et al. 1999) at 8p 12-21, and FEZ1 (Ishii et al.

1999) at 8p22 have been implicated. In mice with a disrupted NKX3.1 gene, defects in prostate duct morphology and secretory fuction occur (Bhatia-Gaur et al. 1999).

Furthermore NKX3.1 mutant mice develop lesions that closely resemble HGPIN

(Bhatia-Gaur et al. 1999). Yet mutations in the coding region of NKX3.1 gene have not been identified (Voeller et al. 1997) nor is NKX3.1 mRNA lost in prostate cancer (Xu et al. 2000). Knock-out of a single NKX3.1 allele may be sufficient to lead to the development of HGPIN, perhaps haploinsufficiency at 8p 12-21 accounts for the lack of

87 mutations found in this region. NKX3.1 expression is found only in adult prostate

(Bhatia-Gaur et al. 1999), whereas deletions of 8p are common in other malignancies including lung and colorectal tumours (Chang et al. 1994) suggesting the existence of other TSG in the 8p22 region.

FEZ1, at 8p22, codes for a leucine zipper protein and alterations in this gene have been reported in oesophageal and breast as well as a prostate cancer cell line (Ishii et al.

1999). There are no reports of alteration of this gene in radical prostatectomy specimens.

13q

Loss of chromosome 13q occurs in at least 50% of prostate tumours (Cooney et al.

1996b; Li et al. 1998). TSGs on the long arm of chromosome 13 include the retinoblastoma gene RBI (13ql4), DBM (13ql4) and the breast cancer susceptibility gene BRCA2 (13ql2-13). Mutations in the RB gene and loss of Rb protein has been reported in localised and advanced prostatic carcinomas (Ittmann & Wieczorek 1996).

RB is thought to regulate apoptosis in prostatic cells, particularly in response to androgens (Bowen, Spiegel, & Gelmann 1998; Yeh et al.). Several studies have however failed to demonstrate mutations in the RB gene (Cooney et al. 1996b; Li et al.

1998; Sarkar et al. 1992), so an alternative gene on 13q may be a significant gene in prostatic carcinogenesis.

GSTP1

Hypermethylation of the GSTP1 gene is seen in 70% of cases of high-grade prostatic intraepithelial-neoplasia (HGPIN) and over 90% of prostate cancers, but is a rare event

88 in benign prostate tissue (Brooks et al. 1998; Lee et al. 1997). The GSTP1 gene on

llql3, encodes for glutathione S-transferase which conjugates electrophilic and

hydrophobic environmental with glutathione and protects the cell

(Mannervik et al. 1985). Hypermethylation of the CpG islands of the promoter region

prevents the transcription of the gene, thus removing the cell’s in-built protection

mechanism against potential environmental carcinogens, including dietary factors

(Nelson & Wilding 2001).

GSTP1 inactivation occurs in the vast majority of cases of HGPIN and prostate cancer

early in the disease course, so may be used as a molecular marker (Lee et al. 1997).

Indeed studies have already reported measurement of GSTP1 methylation in cells in the

urine of men with prostate cancer (Cairns et al. 2001; Goessl et al. 2001). Cairns et a l detected GSTP1 methylation in the urine of 27% of patients with GSTP1 methylation in their primary tumour (Cairns et al. 2001). Goessl et a l improved the sensitivity of this technique to 73% by looking at GSTP1 methylation in the urine sediment after prostatic massage. They found GSTPJ methylation in 1 (2%) of 45 patients with BPH, 2 (29%) of 7 patients with HGPIN, 15 (68%) of 22 patients with localised prostate cancer and 14

(78%) of 18 with locally advanced or disseminated disease (Goessl et al. 2001). In a

study examining the methylation status of GSTP1 in prostate biopsies (Harden et al.

2003) the sensitivity of cancer detection rate was 11% greater using GSPT1 methylation

status compared with histology alone. Included in this study were men undergoing cystoprostatectomy for bladder cancer or radical prostatectomy for prostate cancer.

After prostate resection sextant needle biopsies were taken and examined by an expert uropathologist. DNA was extracted from the needle cores and GSTP1 quantitative methylation specific PCR (QMSP) was undertaken. The whole prostate specimen underwent standard serial sectioning. Histological examination of the biopsy tissue

89 revealed tumour with a 64% sensitivity (95% Cl = 51 -76%), whereas the combination histology and QMSP with an assay threshold >10 detected cancer with 75% sensitivity

(95% Cl = 63-86%). In men with GSTP1 methylation in biopsy material, but negative histology, early repeat extensive biopsy may be recommended.

Furthermore, GSTP1 methylation status in serum is related to biochemical outcome

(Bastian et al. 2005). GSTP1 methylation was detected in 12% of men with clinically localised disease, 28% of men with metastatic cancer and in 15% of men who developed PSA recurrence after surgery. While both these studies demonstrated high specificity for GSTP1 methylation in prostate cancer, the overall sensitivity of the assays was low. Augmentation of GST activity, using pharmaceutical GST inducers, may in the future have a role in prostate (Wilkinson & Clapper 1997).

Cell cycle regulatory genes

In the normal prostatic epithelium the relatively low rate of cell proliferation is matched by the low rate of apoptosis. In HGPIN and early invasive disease the cell proliferation rate increases by 7-10 fold. In advanced and metastatic prostate cancer the apoptotic rate is reduced by 60% (Abate-Shen & Shen 2000). Changes in the genes that code for the cell cycle regulatory proteins may be critical early events in the development and progression of prostate cancer.

The Gl/S transition checkpoint in the cell cycle is an important point of control as DNA damage can be repaired before replication occurs (Sherr 1996). Progression through this checkpoint is regulated by a family of cyclin dependent kinases (CDKs), whose activity is in turn controlled by CDK inhibitors (CDKIs). CDKIs have been sub-classified into two classes on the basis of their structure and their CDK target (Sherr & Roberts 1999).

90 The first class, the INK4 proteins, inhibit CDK4 and CDK6 and include pl6INK4a and pl^iNK4b (serrano> Hannon, & Beach 1993). The second class of CDKIs are the Cip/Kip family, which includes p27WAF1/CIPI and p27K,pl (Harper et al. 1993) and they regulate the cyclin D-, E-, and A-dependent kinases (Sherr & Roberts 1999).

p 2i WAFI cipi was identified simultaneously as a mediator of p53 induced growth arrest

(Wild type p53-Activated Fragment 1) (el Deiry et al. 1993) and as a regulator of CDK

(CDK-Interacting Protein) (Harper et al. 1993). p27WAF1/CIP1 has two binding sites for p53 and its transcription is activated by wild type p53, but not mutated p53 (el Deiry et al. 1993). Therefore, cancer cells expressing mutated p53 protein may be unable to induce expression of p21WAFI/CIPI. Factors including TGF-p (Datto et al. 1995), IFN-a,

IFN-p, IFN-y (Chin et al. 1997; Gartel & Tyner 1999; Xu et al.), IL-6 (Bellido et al.

1998) and EGF (Chin et al. 1997) have been shown to up-regulate />27WAF1/CIP1, via the

JAK-STAT and other pathways, in a p53 independent manner (Gartel & Tyner 1999). p 2i WAFI/CIPI aiSo inhibits DNA replication directly, by binding to the proliferating nuclear antigen (PCNA), a subunit of DNA polymerase 8 (Li et al. 1994). In addition to its associations with CDKs and PCNA, p 2 ] WAF,/CIP1 has been found to participate in specific protein-to-protein interactions, acting both on other cell cycle regulators and as a modulator of apoptosis (Dotto 2000).

Malignancy is characterised by the inability of cells to respond correctly to growth arrest signals, such as cell-to-cell contact. The various roles of p 2 iWAF,/CIPI and other

CDKIs as inhibitors of the cell cycle and as a modulators of apoptosis and differentiation implies a tumour suppressor function (Toyoshima & Hunter 1994).

Decreased levels of CDKIs like p21WAFI/CIP/, or increased cyclin- CDK levels, may lead

91 to inappropriate cell division. Alternatively, decreased levels of CDKIs may limit the cellular response to DNA damage.

Lack of p21WAhl/CIPI expression has been related to poor prognosis in several solid cancers including pancreatic (Jarrard et al. 1997a), gastric (Noda et al. 2001), bladder

(Shariat et al. 2004) and non-small cell lung cancers (Shoji et al. 2002). A single nucleotide polymorphic region in p 2 lWAFIC,PI and in p27k,pI has been identified as a risk factor for developing advanced prostate cancer (Kibel et al. 2003), and in men with this polymorphism close screening is recommended. Immunohistochemical studies have shown reduced p21WAF1/CIPI protein expression in prostate cancer can be used to predict poor survival outcome (Cheng et al. 2000; Matsushima et al. 1998). Although decreased p21 WAFI CIPI expression is often detected in cancer cells, the mechanism of inactivation is not known, mutations and allelic loss of p21WAFt C,PI are rarely found (Gartel & Tyner

1999; Shiohara et al. 1994). In acute lymphoblastic leukaemia poor prognosis is associated with transcriptional silencing of the p 2 JWAFI C,PI gene by methylation of the

CpG islands in the promoter (Roman-Gomez et al. 2002).

The CDK4 inhibitor p27kipI is a cell cycle regulatory gene whose function is frequently lost in advanced prostate cancer (Kibel et al. 1998). Loss o f p27kipl expression, which is mapped to 12pl2-13.1, is associated with increasing tumour grade and poor clinical outcome (Cote et al. 1998; Guo et al. 1997). This gene is rarely mutated, with lack of expression usually due to aberrant phosphorylation and/or ubiquitinylation (Loda et al.

1997; Singh et al.; Tan et al.). In a recent study using a murine model, the p27kipI gene along with the PTEN and NKX33.1 gene were reduced or eliminated (Gao et al. 2004).

Mice that were heterozygous (p27+/-) displayed enhanced prostate carcinogenesis, whereas mice that were homozygous (p27-/-) showed inhibition of cancer progression.

92 Moreover, expression profiling revealed that Cyclin D1 was up-regulated in

heterozygotes, but was down-regulated in the homozygous mutants. This suggests that p27kipl possesses dosage-sensitive positive as well as negative modulatory roles in

prostate cancer progression.

Another cell cycle regulatory gene is the CDKN2 {pi6) gene located on 9p21. CDKN2

codes a cyclin-dependent kinase inhibitory protein, which controls passage through the

G1 phase of the cell cycle. Inactivation of the CDKN2 gene may facilitate progression

through the cell cycle. Increased expression has been reported in proliferative

inflammatory atrophy (PIA) a common histological finding that maybe associated with

prostate cancer (Faith et al. 2005). Furthermore, increased expression has been reported

in 83% of primary prostate tumours and was associated with higher grade and worse

clinical outcome, whereas loss was seen more frequently in metastatic lesions (Jarrard et

al. 2002).

Mutations in the p i 6 gene have been identified in advanced and metastatic prostate

cancer but not in primary tumours. In primary prostate cancer it has been suggested that

inactivation may occur by point deletion of one pi 6 allele and promoter methylation of

the remaining allele (Jarrard et al. 1997b). Jarrard et al. reported 3 (13%) of 24 primary

and 1 (8%) of 12 metastatic tumour samples had hypermethylation in the promoter

region. Deletions near the CDKN2 gene were detected in 12 (20%) of 60 primary

tumours and in 13 (46%) of metastatic lesions. Gu et al looked at CDKN2 in early

prostate cancers and could find no evidence of methylation, although deletions close to

the gene were identified (Gu et al. 1998). Surprisingly, up-regulation of the pi 6 protein

has been reported as a prognostic marker of disease recurrence in prostate cancer

(Halvorsen et al. 2000; Lee, Warde, & Jewett 1999). Similarly, up-regulation of p i 6 has

93 been reported in ovarian cancer and is associated with disease progression and unfavourable prognosis in this tumour (Dong et al. 1997). The reason for this anomaly is not clear, it may be that the up-regulated p i6 protein is abnormal and this confers a growth advantage, but there is no evidence for this theory.

Other cell cycle regulatory genes include cyclin D1, this gene is over-expressed in metastatic prostate cancer (Drobnjak et al. 2000) but not primary tumours (Gumbiner et al. 1999).

lOq

Up to 50% of prostate cancers have been found to harbour deletions on lOq detected by

FISH, CGH, Al and cytogenetics (Carter et al. 1990; Macoska et al. 1993;

Madersbacher et al. 1998; Sakr et al. 1994; Saric et al. 1999; Visakorpi et al. 1995b).

Two commonly deleted regions have been identified, 10q23.1 and 10q24-5. Putative

TSG in these regions include PTEN/MMAC1 (Phosphatase and Tensin homologue/

Mutated in Multiple Advanced Cancers) at 10q23.23 (Cairns et al. 1997) and MXI1 on

10q25 (Di Cristofano & Pandolfi 2000; Schreiber-Agus et al. 1998), though their respective roles in prostate carcinogenesis is not fully established.

Germ line mutations are found in the autosomal dominant PTEN in cancer syndromes, such as Cowden’s disease (breast and thyroid cancer) (Liaw et al. 1997) and Bannayan-

Zonana syndrome (multiple lipomas and haemangiomas) (Marsh et al. 1997). PTEN codes a lipid phosphatase, which induces p27k,pl expression. Loss of PTEN activates

PKB/AKT kinase resulting in enhanced cell proliferation, decreased apoptosis, and increased angiogenesis (Zhong et al.).

94 PTEN mutations have also been found in 5-27% of early and 30-58% of metastatic prostate cancers (Alers et al. 2000; Cairns et al. 1997; Dong et al. 1998; Feilotter et al.

1998; Halvorsen, Haukaas, & Akslen 2003; Suzuki et al. 1998). In 104 radical prostatectomy specimens the expression of PTEN and p27k,pl was examined in tissue microarrays using immunohistochemistry (Halvorsen, Haukaas, & Akslen 2003). PTEN was negative in 27%, and combined loss of PTEN and p27kipl expression was seen in

18% of tumours; combined loss was associated with adverse pathological features, as well as being an independent predictor of time to biochemical failure and clinical recurrence.

Mice deficient in one PTEN allele and both Cdknlb alleles (which codes for p27kipl) all develop prostate cancer within 3 months of birth (Di Cristofano et al. 2001).

Inactivation of one of the two PTEN genes may be sufficient to allow the progression of tumour - haploinsufficiency. To demonstrate this PTEN knock-out mice were bred with

TRAMP mice (Transgenic Adenocarcinoma of Mouse Prostate) (Kwabi-Addo et al.

2001). All TRAMP mice develop prostate cancer, mice with one PTEN gene knocked out developed larger more aggressive tumours and had a significantly reduced survival compared with mice with both PTEN genes activated. Mice with both PTEN genes knocked out had the same survival rate and had similar sized tumours to those with one

PTEN gene knocked-out. Rapamycin, an immunosuppressive agent that acts on

FRAP/mTOR protein kinase, a downstream effector protein of PTEN, reduces the volume of prostate cancer xenografts in PTEN knock-out mice. Clinical trials are currently underway to see whether rapamycin has a similar affect in patients with prostate cancer that do not express PTEN (Neshat et al.).

95 MXI1, located at 10q24-q25, is implicated as a TSG as it codes for a MYC-binding protein (Eagle et al. 1995; Schreiber-Agus et al. 1998). The MYC gene has been implicated in a number of cancers, it is mapped on 8q, a region frequently amplified in prostate cancer (Bubendorf et al. 1999). Some studies have found a high mutation rate in MXI1 (Eagle et al. 1995) (Prochownik et al. 1998) others have failed to detect mutations in primary prostatic tumours (Kawamata et al. 1996).

96 1.1 l(c)2. Metastatic prostate cancer

Metastasis requires a complex interrelated series of events including vascularisation, invasion, survival in blood or lymph, adhesion, extravasation and proliferation at a distant site (Fidler 1990). This requires several critical genetic changes and these changes may act as molecular markers and targets for therapeutic intervention.

16q

Several CGH and Al studies have shown losses of 16q in late stage prostate disease

(Cher et al. 1998b; Latil et al. 1997; Li et al. 1999), though up to 68% of primary tumours may also harbour losses of 16q (Harkonen et al. 2005). Loss of 16q23 and q24 are associated with metastatic potential (Elo et al. 1999; Suzuki et al. 1996) and loss of

16q24.3 with risk of disease recurrence and progression (Harkonen et al. 2005).

A candidate gene, CDH1 coding for the E-cadherin protein, is located on 16q22.1.

Decreased expression of this gene is associated with high-grade prostate cancer (Dong et al. 1998; Elo et al. 1999; Feilotter et al. 1998; Latil et al. 1997; Pesche et al. 1998;

Richmond et al. 1997; Suzuki et al. 1996), in addition decreased expression is associated with metastatic potential in the primary tumour (Umbas et al. 1994).

Although no mutations have been described, hypermethylation of the CpG island has been reported in cell lines. E-cadherin is calcium dependent and is responsible for cell- to-cell recognition and adhesion (Takeichi 1991). a and P-catenins form part of the same adhesion mechanism and decreased expression of a-catenin has been demonstrated in prostate cancers which have normal E-cadherin expression (Richmond et al. 1997). About 5% of prostate cancers contain mutations in P-catenin (Chesire et al.

2000; Voeller, Truica, & Gelmann 1998). P-catenin has two known roles, firstly in cell adhesion with a-catenin and secondly as part of the Wnt signalling pathway. Another

97 potential TSG, located at 16q24.2-q24.3, is H-cadherin, which is associated with breast lung and colorectal tumours, though this gene has not been directly implicated in prostate cancer tumourigenesis.

17p

Mutations in the TSG p53 have been described in the majority of cancers including breast, colon and lung (Hollstein et al. 1991). Located at 17pl3.1, p53 codes for a nuclear phosphoprotein that has a negative effect on cell growth. In the event of DNA damage, the cell cycle is arrested or apoptosis is induced. If p53 is inactivated and loses its function, damaged DNA may be transcribed (Ruijter et al. 1999). The frequency of p53 mutations in prostatic cancer ranges from 1-42% and a consistent finding is its association with high grade and high stage disease coupled with disease recurrence and poor prognosis (Aprikian et al. 1994; Brewster et al. 1999; Navone et al. 1993; Shi et al.

1996; Stackhouse et al. 1999). The frequency of p53 mutations is relatively low in prostate compared with other cancers (Abate-Shen & Shen 2000). Prostate cancer occurs only rarely in the Li-Fraumeni cancer syndrome where patients carry a germline mutation in p53 (Kleihues et al. 1997), although this may be because patients die from other malignancies before they acquire prostate cancer.

The low mutation rate, particularly in early stage disease, precludes p53 from being an effective target for curative intervention in prostate cancer (Brooks et al. 1996; Stattin

1997). p53 may however play a role as a molecular marker in radiotherapy resistance along with BCL, MYC and RAS. Radiotherapy induces apoptosis and may be less effective in patients who have defective apoptotic pathways as a result of mutations in these genes (Prendergast et al. 1996). The p53 gene mutation predicts poorer overall survival for patients treated with antiandrogen therapy after radiotherapy (Grignon et al.

98 1997). The presence of mutations in the p53 gene may therefore be used before radiotherapy and antiandrogen therapy to predict treatment failure (Huang et al. 1998;

Prendergast et al. 1996).

The breast cancer susceptibility gene BRCA1 is also mapped to 17p. This gene is not deleted however in prostate cancer (Brooks et al. 1996).

18q

Loses of 18q have been reported to be associated with metastatic prostate cancer (Saric et al. 1999). Two TSGs implicated in cancer progression are located in this region, DCC

(Deleted in Colon Cancer) and DPC4 (Deleted in Pancreatic Cancer), however no mutation in these genes have been found in prostate cancer (Gao, Porter, & Honn 1997;

U edaetal. 1997).

99 EZH2

The EZH2 gene, situated on 7q35-36, encodes the EZH2 protein, of which levels are significantly increased in prostate cancer compared with BPH tissue. EZH2 increases histone deacetylation, which is a repressor of transcription in prostate cancer

(Varambally et al. 2002). Elevated levels of EZH2 mRNA and protein in clinically localised prostate cancer predicts poor prognosis (Varambally et al. 2002). In a separate immunohistochemistry study moderate or strong expression of EZH2 coupled with expression of ECAD was strongly associated with the recurrence after radical prostatectomy in 183 men even after adjusting tumour stage, Gleason score, and PSA level (Rhodes et al. 2003). This gene represents an exciting development both as a prognostic marker and also as a therapeutic target.

ERBB2

The ERBB2 gene is located on 17q21.1 and mutations have been widely reported in a number of tumours including breast and ovarian. ERBB2 or HER-2/neu, codes for a transmembrane tyrosine kinase growth factor, which is involved in interleukin 6 (IL6) signalling through the MAP kinase pathway. IL6 induces phosphorylation in the ErbB2 and ErbB3 receptors in prostate cancer cell lines (Qiu, Ravi, & Kung 1998).

Phosphorylation inactivates the receptors resulting in inhibition of the MAP kinase pathway (Qiu, Ravi, & Kung 1998). The recent development of a monoclonal antibody to the epidermal growth factor receptor, anti-ERBB2 (Herceptin, Genetech Inc, South

San Francisco, CA) for use in advanced breast cancer (Pegram & Slamon 1999) has led researchers to re-focus on the role of this oncogene in prostate cancer. Trials on the use of Herceptin in prostate cancer are ongoing (Scher 2000), but it is clear that high level

ERBB2 amplification does not occur in this disease (Kallioniemi & Visakorpi 1996).

100 However, Herceptin does inhibit growth in prostate cancer cell lines, including LNCaP

(Agus et al. 1999) and ERBB2 enhances androgen receptor signalling in the presence of low androgen levels (Craft et al. 1999).

KAI1

The KAIl gene (Kang Ai, Chinese for anticancer), found on 11 pi 1.2-13, encodes a

2.4kb mRNA for a membrane glycoprotein belonging to the TM4 superfamily, suggesting a role in cell-to-cell interaction or cell-extracellular matrix binding (Yu et al.

1997). KAI-1 has been shown to suppress metastatic potential in highly metastatic

Dunning rat tumour models (Behrens et al. 1989). In addition expression of this gene is reduced in metastatic prostate cancer cell lines (Gao et al. 1997), but no mutation in this gene has yet been found, suggesting epigenetic silencing may be the mechanism of down regulation.

The KAIl gene is in close proximity to CD44 (11 p i3), this latter gene codes for a membrane glycoprotein, which is also present on lymphocytes and participates in cell- to-cell interaction. Dong et al. transfected the CD44 gene into metastatic rat prostate cells and demonstrated suppression of metastatic potential, without suppression in vivo of growth rate or tumorigenicity (Dong et al. 1995). Moreover down regulation of CD44 at a protein and mRNA level correlated with metastatic potential in the Dunning rat tumour model (Dong et al. 1995). Studies in patients have found an inverse correlation between prostate cancer histological grade and stage with CD44 expression (Noordzij et al. 1997). Immunohistochemical studies have shown CD44 expression is either negligible or non-existent in prostate cancer lymph node metastases.

101 l.ll(c)3. Hormone refractory prostate cancer

The androgen receptor

During embryological development and puberty androgens play a key role in prostate development. Testosterone is converted to its more potent metabolite dihydrotestosterone by the enzyme 5 a-reductase. There are no reported cases of prostate cancer in men who have a congenital deficiency of 5 a-reductase. This enzyme can be antagonised pharmacologically by finasteride (Proscar®) and this drug has recently undergoing a phase III trial in over 18.000 men to evaluate a possible role in prostate cancer chemoprevention (Feigl et al. 1995). Early reports suggest a reduction in the incidence of prostate cancer in the treated arm, however there appears an increased risk of high-grade tumours in men receiving finasteride. Whether this is a genuine difference or a result of the difficulties in interpreting the glandular architecture in men receiving finasteride remains to be seen (Thompson et al. 2003).

Dihydrotestosterone and testosterone bind to the androgen receptor (AR), which

initiates phosphorylation and dimerisation of the receptor. The AR binds to a specific

DNA sequence, the androgen responsive element, located in the promoter region of androgen responsive genes. The AR complex, in collaboration with cofactor proteins, up or down regulates the transcriptional activity of the androgen response genes (Figure

1.6 ).

102 Low androgen level: Testes Adrenal glands IGF-1, KGF, EGF

prostate cell cytoplasm

Androgen 5a -reductase

prostate cell nucleus

DNA

Androgen responsive element

Figure 1.6 The androgen receptor ligands and androgen responsive element. (AR = androgen receptor, DHT = dihydrotestosterone, IGF-1 = insulin-like growth factor-1,

KGF = keratinocyte growth factor, EGF = epidermal growth factor)

The removal of androgens by surgical castration or therapeutic blockade initiates apoptosis in prostate cancer cells resulting in a clinical response in up to 95% of men for a median 18 months (Petricoin et al. 2002). However, some prostate cancer cells have the ability to grow even in low or absent androgen levels. These cells have a selective growth advantage over the androgen dependent cells following androgen withdrawal. Androgen independent cells are able to proliferate and typically have a highly aggressive phenotype.

103 The androgen receptor is amplified in 20% of hormone refractory prostate cancers, but in only 2% of the same untreated individuals (Edwards et al. 2003). This suggests AR gene amplification is selected following androgen withdrawal though is not solely responsible for androgen independence. Amplification is associated with over­ expression of the AR gene (Koivisto et al. 1997; Visakorpi et al. 1995a) and some hormone refractory tumours express AR protein highly even in the absence of AR gene amplification (Edwards et al. 2003; Linja et al. 2001). Over-expression allows even very low levels of androgen (such as adrenal androgen levels after surgical or medical castration) to promote androgen dependent growth and therefore, patients with AR amplification may respond better to therapeutic maximum androgen blockade

(Visakorpi et al. 1995a).

There is also evidence that growth factors including insulin-like growth factor 1 (IGF-

1), keratinocyte growth factor (KGF) and epidermal growth factor (EGF) may be able to activate the AR either through a ligand dependent or independent pathway. For example, the cAMP, protein kinase A and protein kinase C pathway may be able to directly activate the androgen receptor (Ikonen et al. 1994).

As well as amplification of the AR gene several hundred other AR gene mutations have been reported (www.mcgill.ca), although not all these reported mutations are associated with prostate cancer. These mutations may alter the AR function and ligand specificity allowing other steroid hormones to bind, overcoming the need for androgens to stimulate growth. Mutations are uncommon in untreated prostate cancer and in cases treated with surgical castration, though they are found in patients treated with the testosterone antagonist flutamide (Taplin et al. 1995; Wallen et al. 1999).

Approximately one third of patients treated with flutamide have specific mutations,

104 which result in AR activation (Eeles et al. 1998). This mutation, T887A, found only in patients treated with flutamide, changes the AR response from an antagonist to an agonist. This may explain why a minority of patients paradoxically, show a temporary clinical improvement when flutamide therapy is withdrawn.

Up to 10-15% of hormone refractory metastatic lesions express only low AR levels as a result of methylation in the promoter regions (Kinoshita et al. 2000). Clearly, a number of genetic and epigenetic changes occur in relation to the androgen receptor in late stage prostate cancer, and AR amplification, overexpression and mutation are not the exclusive cause the development of androgen insensitivity.

The androgen receptor complex exerts its effect on the target gene in conjunction with various cofactor proteins. Several of these proteins have been cloned including SNURF,

ARIP3, ARA54, 55, 70, 160 and ANPK. Their role in cancer progression in vivo remains unknown at present.

BCL2

Bcl2 over expression is a characteristic of advanced and hormone refractory prostate cancer and it may account for the dramatic reduction in apoptosis that occurs at this late stage (McDonnell et al. 1997). Like p53, Bcl2 expression maybe used as a molecular marker that correlates with disease outcome (Mackey et al. 1998). Over-expression in prostate cell lines confers a resistance to chemotherapeutic agents (Tu et al. 1995) and in a clinical setting several studies are testing whether Bcl2 inactivation can be used to prevent tumour recurrence (Gleave et al. 1999; Miyake et al. 1999).

105 8q

Several studies using CGH have shown that gain in 8q is a characteristic of hormone refractory disease (Cher et al. 1996; Visakorpi et al. 1995b). Almost 90% of samples of hormone refractory prostate cancer and distant metastases compared with 5% of untreated primary tumours contained gains of the whole of 8q. Alers et al. found the number of individuals with 8q gains increases with advancing tumour stage (Alers et al.

2000). Furthermore, gain in 8q, as well as loss of 6q, is significantly less frequently encountered in regional lymph node metastases than in distant metastases (Alers et al.

2000). This implies differential genetic changes for haematogenous versus lymphatic spread (Alers et al. 2000).

Several sites within 8q have been suggested as containing critical oncogenes, 8q21

(Cher et al. 1996) and 8q 23-q24 (Nupponen et al. 1998). MYC is located at 8q24, and codes for a nuclear phosphoprotein whose function is in the promotion of DNA replication, regulation of the cell cycle and control over cell differentiation (Yeh et al.).

Amplification of this oncogene has been found in 8% of primary and up to 72% of hormone refractory or advanced prostate cancers (Bubendorf et al. 1999; Nupponen et al. 1998; Sato et al. 1999). Furthermore amplification of this gene is associated with poorer prognosis in patients with locally advanced disease (Sato et al. 1999). In androgen dependent LNCaP cell lines, cells treated with the androgen antagonist bicalutamide underwent growth arrest, whereas cells over-expressing c-myc were able to grow in an androgen independent environment (Bernard et al. 2003). This did not occur as a result of a change in AR activity, rather c-myc is a downstream target of AR and overexpression of c-myc may be an alternative mechanism of androgen independence.

106 Other sites of amplification on 8q include 8q23 and 8q24.2, which harbours the putative oncogenes EIF3S3 and PSCA (Reiter et al. 2000; Visakorpi et al. 1995a). Using FISH,

Nupponen et a l found EIF3S3 in 30% of cases of hormone refractory prostate cancers.

Moreover EIFS3S was frequently, but not exclusively, co-amplified with MYC

(Visakorpi et al. 1995a). PSCA is often co-amplified with MYC, and over-expression of

PSCA has been reported in high grade and metastatic prostate cancer (Reiter et al. 2000)

l.ll(c)4. Telomerase activity

Ageing is one of the major risk factors for the development of prostate cancer. HGPIN has been found in prostate samples from men in their twenties and occurs with increasing frequency with advancing age. Invasive cancer generally manifests itself in the sixth and seventh decade. This is because the prostate epithelium has a low proliferation rate and so takes longer for transforming events to occur. In addition, it may be related to the ability of telomerase over time to overcome replicative senescence in cancer cells. Telomeres are the non-coding repeat base sequences found at the end of eukaryotic chromosomes. With each cell division the telomeric sequence is shortened as

DNA polymerase starts at an RNA primer that is subsequently removed. The telomere thus protects the coding regions until it has been shortened to a critical length. Then, at a critical age, it is thought to act as a signal to initiate apoptosis. Telomerase is an enzyme that adds telomeric sequences to newly formed DNA, therefore compensating for loss of sequences as a result of replication. This prevents the telomere reaching its critical shortened length and allows unlimited cell division.

Telomerase activity has been examined in exfoliated epithelial cells in the urine

(Botchkina et al. 2005). All 56 men with prostate cancer had high levels of telomerase activity, though it was also detected in 12% of men who were thought to have BPH, this

107 represents a potentially exciting new non-invasive test. Telomerase activity has not been found in normal or benign prostate tissue, it has however been found in 172 (89%) of

194 cases of prostate cancer where it is associated with a reduced telomere length

(Sommerfeld et al. 1996; Zhang et al. 1998). Increased levels of telomerase activity correspond more frequently to poor tumour differentiation (Lin et al. 1997).

Furthermore telomerase activity was found in HGPIN in 14 (74%) of 19 samples and 5

(12%) of 42 samples of normal prostate tissue adjacent to prostate cancers (Zhang et al.

1998). This suggests ‘normal’ adjacent tissue may have genotype changes before malignant phenotype changes can be identified.

Attempts have been made to use telomerase activity in prostate fossa biopsies at radical prostatectomy to predict local recurrence. Increased telomerase activity was found in the prostatic fossa samples of 5 (10%) of 48 men with pT2 and 7 (15%) of 47 men with

>pT2(Straub et al. 2001). Recurrence rates in these trials are awaited, although telomerase activity may enable more accurate molecular staging of prostate cancer.

108 1.12 Molecular markers in cancer

Gold and Freedman were the first to observe the potential of molecular markers when they discovered carcinoembryonic antigen (CEA), a protein usually only found during fetal development, is expressed in some colorectal cancers (Gold & Freedman 1975).

This work has been expanded upon, augmented by the discovery of the human genome sequence, such that now genetic and epigentic changes are being used as molecular markers. Identification of specific genetic changes already allows the subclassification of several malignancies according to prognosis, including B cell lymphoma (Alizadeh et al. 2000) malignant melanoma (Bittner et al. 2000), breast cancer (Iwao et al. 2002), lung adenocarcinoma (Beer et al. 2002) and Wilms’ tumours (Lu et al. 2002).

Prognostic markers currently used in patients with prostate cancer are limited as they measure tumour differentiation and the extent of malignant tissue rather than the underlying biological properties that drive tumour behaviour (Singh et al. 2002).

Attempts to identify genetic changes associated with prostate cancer progression have yielded alterations in a number of candidate genes associated with cancer progression, including loss of p53, amplification of MYC, loss of p27 and loss of PTEN (Sellers &

Sawyers 2001). However, presently no single genetic change has been identified in a sufficiently significant proportion of prostate cancer specimens to have prognostic value which will warrant its use in everyday clinical practice (Singh et al. 2002).

Novel markers have several potential roles:

• Detect early disease before symptoms develop when cure is still possible.

• Allow the detection of metastases before they are clinically evident thus

preventing unnecessary radical intervention when cure is not possible.

• Assess the possible effect of a therapy on a tumour to facilitate treatment

planning. For example, breast cancer cells that express high levels of tyrosine

109 receptor kinase (ERBB2 or HER2/neu) are more likely to respond to Herceptin

treatment. Assays are used to measure tyrosine receptor kinase expression before

treatment is initiated.

• Monitor the effects of treatment. PSA plays an important role in the follow-up of

men after radical treatments for localised prostatic cancer, detecting disease

relapse several years before symptoms develop. Newer therapies, for example

agents that disrupt angiogenesis or tumour invasion, may produce more subtle

changes in the tumour so markers may be used to monitor therapeutic response.

1.12(a) DNA markers and microarray technology

Several genetic changes have been identified in early cancer and include: oncogenic mutation, microsatellite instability and methylation of the promoter region. These genetic modifications were originally detected in biopsy specimens but have subsequently been identified in cells from bodily fluids including: serum - TP53 in hepatocellular carcinoma, urine - telomerase and GSTP1 in prostate cancer and telomerase and TP53 mutations in bladder cancer, saliva - CDKN2A in , sputum - RAS mutations in men with lung cancer and faeces- RAS mutations in colorectal cancer (Sidransky 2002).

Gene expression profiling using microarrays allows a large-scale analysis of gene expression and cancer characterisation. DNA microarrays consist of thousands of spots of DNA oligonucleotides on a ‘chip’ each coding for part of a known or unidentified gene. Radio or fluorescently labelled DNA or RNA is applied to the chip and hybridises where there is a complementary oligonucleotide sequence. The site on the chip and intensity of the radioactive or fluorescent signal is proportional to the amount of a

110 particular DNA/RNA. Rather than studying individual genes this technique allows examination of as many as 25,000 genes at one time.

Microarray chips have been used in young women with node negative breast cancer to identify a gene expression signature that predicts poor prognosis (Veer et al. 2002). In this study 5,000 genes were either significantly up- or down-regulated, a group of 70 genes could be used to accurately predict disease outcome in 83% of 78 patients. This had important clinical implications as patients with the good prognosis genetic signature could safely avoid adjuvant chemotherapy, which they would otherwise have received, saving unnecessary side effects and cost.

The signature derived included genes involved in regulating the cell cycle, invasion, metastases, cell signalling and angiogenesis (eg cyclin E2, MCM6, VEGF receptor

FL1). In breast cancer, like prostate cancer, numerous studies have shown associations

between individual gene expression and patient outcome (cyclin D2, ER-, HER2/neu and MYC) (Bieche & Lidereau 1995). In the microarray analysis none of these genes

was included in the 70 marker genes. The authors concluded cancer occurs as a result of

a coordinated action of several genes, so genes looked at in isolation have only limited

predictive power.

Dhanasekaran et al. were one of the first groups to report gene expression profiling in

prostate cancer. Using a selection of normal tissue prostate, benign prostatic

hyperplasia, primary cancer, metastatic cancer and cell lines they identified a number of

genes with altered expression patterns depending on the stage of disease. A hierarchical

clustering algorithm was used to group genes based on the similarities of genes

expressed. They found benign prostatic conditions clustered separately from malignant

111 tissue or cell lines. Within the cancer cluster, metastatic and clinically localised cancers clustered into two succinct groups. Several genes already implicated in prostate cancer carcinogenesis had altered expression levels. PTEN and CDH1 (E-cadherin) were down-regulated and MYC and the fatty acid synthase ( FSAN) gene were up-regulated. In addition numerous genes not previously implicated in prostate cancer were identified.

Notably, hepsin, a cell-surface serine protease, was up-regulated by 4.3 fold on microarray analysis and this was confirmed on Northern Blot studies.

Immunohistochemistry showed hepsin was preferentially expressed by neoplastic and

HGPIN samples over benign prostate. In 78 men with clinically localised disease, absent or low hepsin expression was significantly associated with PSA recurrence after radical prostatectomy (p=0.03). In multivariate analysis the association of weak (HR 2.9

(p=0.0004)) or absent (HR1.65 (p= 0.037)) hepsin protein expression was similar to having high-grade disease in terms of PSA outcome.

Microarrays have also been used to predict tissue identity and pathological outcome after radical prostatectomy (Singh et al. 2002). This study also examined gene expression in relation to clinical outcome after radical prostatectomy, albeit involving small patient numbers. Using RNA extracted from tumours and normal tissue of fresh radical prostatectomy specimens the authors were able to identify 317 genes highly expressed in the tumours and 139 genes more highly expressed in normal prostatic tissue. Using a 4- or 16-gene model they were able to predict tumour from normal tissue with an accuracy of 77% and 86%, respectively. When the gene expression patterns were compared with the clinico-pathological prognostic indicators only the Gleason score correlated with gene expression, not pathological stage, positive surgical margins or perineural invasion. Fifteen genes had expression positively correlated with Gleason score with 14 genes had expression negatively correlated with Gleason score. Most

112 high-grade cancers expressed the positively correlated genes. Interestingly, a subset of intermediate grade tumours also expressed these genes, possibly identifying a subset of phenotypically intermediate grade tumours with a more aggressive genotype.

Singh et al also examined the relationship between gene expression and tumour recurrence in 21 unmatched patients, 8 having relapsed and 13 having remained PSA free for at least 4 years. No single gene was statistically associated with recurrence, when a 5-gene model was used (including chromogranin A, platelet-derived growth factor receptor p, HOXC6, inositol triphosphate receptor 3 (ITPR3) and siayltransferase-1), 90% accuracy could be achieved in predicting recurrence. The

Gleason score of tissue adjacent to the sample tissue was >8 in 4/8 relapse patients compared with 0/13 in the non-relapse group. In this small sample of unmatched patients it is not possible to draw meaningful conclusion about the gene expression patterns and clinical outcome. This study does however demonstrate that gene expression patterns can be used to distinguish benign from malignant prostatic tissue with 86-92% accuracy.

Microarray analysis was performed on the cRNA of 24 patients with clinically localised prostate cancer (Welsh et al. 2001). A large number of genes with expression specific to malignant cells including, hepsin and fatty acid synthase ( FASN), were identified. The gene expression of malignant cells could be divided into two groups by the expression patterns of a group of ribosomal genes and patients with higher Gleason score disease had significantly lower ribosomal gene expression (p<0.01). The authors emphasised the potential for therapeutic targeting of differentially expressed genes including FASN, hepsin and MIC-1, though no new evidence was presented in this study to support this

(Welsh et al. 2001).

113 The results of gene expression profiling across the cancer spectrum are promising, and classification of cancers according to their genetic expression are now possible

(Alizadeh et al. 2000; Bittner et al. 2000; Blaveri et al. 2005; Eschrich et al. 2005; Perou et al. 2000; Yang et al. 2005). Tailoring cancer treatments according to gene expression patterns is already underway and therapeutic targets are being tested. Prostate cancer poses several difficulties. The heterogeneous nature of the disease means that while the expression pattern of one area of tumour may accurately predict prognosis for that tumour the other foci with different genetic make-up may behave differently to the index lesion. For the same reason therapeutic modulation of genes with altered expression patterns may select out those tumours which do not have altered expression of the particular gene, in the same way that antiandrogen treatment selects androgen independent prostate cancer cells. In prostate cancer, the results of prospective studies to examine gene expression and patient outcome take many years to mature. The studies to date are retrospective and have involved small patient numbers, larger scale trials are required to demonstrate the validity and reproducibility of these results before these experimental techniques are used in the prostate cancer clinic.

1.12(b) Allelic imbalance

The protein coding regions account for just 3% of the whole human genome. The remaining 97% is located either within genes, the introns, or between genes. Since these regions do not code proteins variation is tolerated resulting in genetic diversity. Much of this non-coding DNA consists of repeated nucleotide sequences or Variable Number

Tandem Repeats (VNTR). These may be single tandom repeats (STRs), often a dinucleotide consisting of CA (cytosine and adenosine) on the sense strand and GT

(guanine and thymidine) on the antisense strand. These 2-5 nucleotide repeat segments

114 are called microsatellite regions. Microsatellite regions also occur within or close to tumour suppressor genes. If a cell develops a mutation of one allele of a TSG and loss of the other allele, loss of heterozygosity (LOH), then tumourigenesis may occur.

Detecting deletion of a microsatellite region may therefore implicate the coding sequence near to the polymorphic region in the initiation or progression of the disease.

In a normal cell the microsatellites of the paternal and maternal alleles are usually of different length. In a PCR reaction a single pair of oligonucleotide primers

(microsatellite markers) that surround such sequences produce variably sized DNA fragments depending on the number of repeats. If either the maternal or the paternal microsatellite region is lost during tumorigenesis (LOH) the product of only one of the alleles will be amplified - this is allelic imbalance (Al) (Figure 1.7).

The role of Al as a molecular marker was first described in 1996 from urine samples of

25 patients with bladder cancer (Mao et al. 1996). Al alterations that matched those in the tumour were detected in 95% of those with bladder cancer whereas urine cytology detected cancer in only 50% of samples. This study demonstrated the role of microsatellite markers in the screening of cancer, subsequent studies have used Al analysis in monitoring the response of bladder cancer to therapy.

The problem with this technique is that it is not very sensitive and so large quantities of cancer DNA is required. Furthermore a panel of microsatellite markers is required, perhaps 15-20, to reliably detect cancer.

115 Figure 1.7. Demonstrates locus ‘a’ in a normal cell and a malignant cell containing microsatellite instability and allelic imbalance (LOH).

Normal cell Canoer cell

XV ♦ 22 x C homologous pairs ot chromosomes Miaosatellite instability (MIN) Loss oi heterozygosity I.UOH)

I Paternal- Maternal- (CA)n repeat lengths Deletion by partial or total derived derived are altered at several chromosomal b ss chromosome chro moeo me sites i n genome

Allele Longer Shorter JL repeat X repeat lost at Two homologous \ at locus a / ‘ at locus a \ locus a chromosomes N! n y (e.g. visualised Locus a ^ M Locus a ■ Locus a by FISH analysis) V \

Number of (CA)n (TCA), (CA>3

® (CA),, repeals detected -CA-CA-CA-CA- - CArCA-CArCA-CA- -CA-CA-CA-CA- by PCR - CA-CA-CA- - CArCA- (Allele lost) at locus a Normal Microsatellite instability Allelic imbalance

1.12(c) Microsatellite instability

Errors generated during DNA replication are usually corrected by DNA repair enzymes,

such as those involved in nucleotide excision repair and mismatch repair. Short

stretches of mismatched base pairs occur in microsatellite regions during normal DNA replication, however these errors are not corrected if mismatch repair genes have been

116 inactivated (Figure 1.7). Inactivating mutations of mismatch repair genes have been described in cancer. Errors in mismatch repair lead to instability at microsatellites and increased mutation rates causing differing lengths of microsatellite regions between the tumour and normal tissue of the same individual. This is termed microsatellite instability (MSI) and is responsible for the development of hereditary nonpolyposis colon cancer (HNPCC).

Several mismatch repair genes have been identified, including hMLHl, hPMSl, hPMS2, hMSH2. Using immunohistochemistry hMSH2 staining was minimal to low in normal glands (Velasco et al. 2002). In 32% of benign prostatic hyperplasia cases hMSH2 staining was increased, in 71% of cancer specimens moderate to high staining was reported. Microsatellite instability was detected in 60% of absent to low staining and 26% of moderate to high staining prostate carcinoma specimens. Independent of tumour grade and stage, decreased risk for PSA recurrence after radical prostatectomy correlated with absent to low hMSH2 staining. Some studies have shown extensive MSI is associated with high grade and more advanced tumours (Uchida et al. 1995) though others have failed to confirm this (Dahiya et al. 1997a). Mutations or epigenetic silencing of the mismatch repair genes may augment cancer progression by failing to correct potentially pro-carcinogenic DNA defects, identifying these changes may potentially provide prognostic and therapeutic targets.

1.12(d) Hypermethylation

Epigenetic changes are alterations in gene expression without change in DNA sequence.

The most common somatic alteration in prostate cancer is hypermethylation in the regulatory region of certain genes, most commonly in GSTP1. Areas rich in the cytosine-guanine motif (CpG islands) occur in the promoter region of some genes,

117 including tumour suppressor genes. Methylation of these cytosine residues, which occurs in cancer as well as normal fetal development, inactivates the promoter and blocks transcription.

Inactivation of TSGs by methylation occurs in a number of cancers. To assess DNA methylation DNA is treated with sodium bisulphite. Sodium bisulphite converts unmethylated cytosines to uracil while the methylated cytosines are protected and remain as cytosine. Methylation specific PCR assays can be performed with great sensitivity, amplifying only methylated DNA and enabling the detection of 1 cancer cell amongst 1000 normal cells in a body fluid. This technique is currently being used to assess the urine of men with prostate cancer, saliva in patients with , bronchiolar lavage fluid in patients with lung cancer and the serum of people with lung, head and neck and colorectal cancers.

Efforts to quantify the degree of methylation are being developed. Real-time-PCR

involves the monitoring of fluorescent signals during DNA amplification. Cancer cells that have a high degree of methylation will amplify at a faster rate than benign cells

with low or no methylation. Real-time-PCR has been used to assess the degree of

methylation of cells obtained by prostatic massage (Goessl et al. 2001). GSTP1

(glutathione-S-transferase) encodes for an enzyme that detoxifies potential carcinogens.

This is inactivated, by promoter methylation, in 70% of patients with HGPIN and 90%

of men with invasive prostate cancer (Brooks et al. 1998; Lee et al. 1997).

1.12(e) Mitochondrial DNA mutation

Mitochondrial DNA is more susceptible to damage by exogenous mutagens than

nuclear DNA and has less efficient repair mechanisms (Sidransky 2002). Several

118 tumour specific mutations have been described in mitochondrial DNA initially in men with colorectal carcinoma and subsequently in bladder, breast, lung and head and neck cancers.

Mitochondrial DNA is significantly shorter than genomic DNA at 16000 base pairs.

Furthermore mutations are most frequently found in a region probably involved in mitochondrial replication called the mitochondrial D-loop. While the detection of mutations in mitochondrial DNA may prove useful, methods are not yet available to carry out high volume analysis as required in the clinical setting.

1.12(f) Viral DNA

Viral DNA can be detected in a number of tumours, including human virus

(HPV) in cervical and head and neck squamous-cell-carcinomas and Epstein-Barr virus

(EBV) in . These viruses are strongly implicated in the aetiology of these diseases. While viruses are probably not aetiological factors in prostatic neoplasms, viral DNA may become a useful molecular marker in other carcinomas.

1.12(g) RNA markers

Genetic alterations lead to fundamental changes in the level of mRNA and this can be detected by the reverse transcriptase polymerase chain reaction (RT-PCR). This involves the conversion of RNA into a copy or cDNA, catalysed by the enzyme reverse transcriptase. This is then amplified by PCR to allow detection. This has been used to detect mRNA for tyrosinase to distinguish benign skin moles from melanoma. It has also been used to detect transcripts of PSA and CEA in blood, lymph node and bone metastases to confirm detection.

119 Quantitative RNA analysis may be required as altered gene expression may be present in normal as well as malignant cells. Quatitative RT-PCR is currently used to monitor treatment response in, for example mRNAs that encode thymidine phosphorylase in colon tumour samples in patients treated with 5-flourouracil (Salonga et al. 2000).

1.12(h) Protein markers

Several protein-based assays have been developed to detect malignant cells. Most use antibody based assays, although other approaches are being developed. These techniques either detect levels of over-expressed protein or structurally altered protein in cancer cells. The measurement of serum PSA involves a protein assay, other protein markers used to detect cancer or to diagnoses recurrent disease include human chorionic gonadotrophin (HCG) and a-fetoprotein (AFP) in testis and hepatocellular cancers,

CEA in colorectal cancer and CA125 in ovarian carcinoma. Some newer markers, like

CA15-3 in breast cancer or CA19-9 in pancreatic cancer are only found in a fraction of affected individuals (Canizares et al. 2001; Slesak et al. 2000). In addition all these markers lack specificity - they may be elevated by other disease processes, which limits their role in cancer screening. PSA, CEA and CA125 are used to diagnose recurrent disease before radiological evidence or symptoms are apparent, enabling further treatment while the tumour burden is low.

As discussed, telomerase is required for cellular immortalisation and is therefore expressed in almost every cancer type. Telomerase has been detected in the urine of patients with bladder and prostate cancer and the saliva of people with oral malignancies (Califano et al. 1996; Yoshida et al. 1997). Several assays are available, including the telomerase repeat amplification protocol (TRAP). Lymphocytes also

120 express low levels of telomerase, leading to false positive telomerase assays (Califano et al. 1996; Yoshida et al. 1997). A more quantitative telomerase assay may be more useful to determine the presence of cancer cells in a clinical sample.

Antibodies to aberrant proteins expressed by malignant cells may be detected in the serum of patients with certain cancers. For example, antibodies to p53 have been detected in the serum of patients with small cell lung carcinoma (Jassem et al. 2001).

Since the mapping of the human genome, attention has now focused to proteomics as a source of cancer markers. This is because post-translational protein modification may be cancer specific and cannot be determined at the genomic level. Proteomics is the analysis of large number of proteins. It has already been used to compare normal and malignant prostate cells and this information was used to correctly identify unique proteins in the sera of patients with prostate cancer (Paweletz, Liotta, & Petricoin, III

2001). High throughput mass spectroscopy can be used to detect cancer in patient samples including blood and urine. Techniques including matrix-assisted-laser- desorption-ionisation time-of-flight (MALDI-TOF) mass spectroscopy or liquid- chromotography-ion-spray tandem mass spectroscopy (LC-MS) have revolutionised protein analysis. In these examples a patient sample, such as urine or serum, is fractionated and digested with proteolytic enzymes and separated using chromatography

(Wall et al. 2000). The partially separated sample is then put into a mass spectrophotometer. While individual proteins are not identified a proteomic profile containing several thousand points is produced. The protein profile of normal and malignant cells differs and these techniques have been used in diagnosing ovarian and bladder cancers. A diagnostic sensitivity of 100% and a specificity of 95% was reported in one series examining ovarian cancer, including patients with early stage disease

121 (Petricoin et al. 2002). These techniques have the potential to diagnose and detect disease relapse at an earlier stage without the need for a biopsy. Their role as prognostic markers has not yet been examined, though it is likely it may be extended to allow specific proteomic profiles to correlate with patient outcome in the future.

122 1.13 Summary

The current clinico-pathological markers used in the management of men with clinically localised prostate cancer undergoing radical prostatectomy and their limitations have been reviewed. Inaccuracies in these markers results in a significant proportion of men developing recurrent disease after surgery, despite stage migration and modest improvements in staging techniques. The management options for men with recurrent disease are limited to radiotherapy for local control and androgen ablation as systemic therapy. Prospective data on these treatments in the setting of PSA only recurrence after radical prostatectomy is currently lacking.

The fact that specific molecular changes are associated with each step in the pathogenesis of prostate cancer has been emphasised. By identifying these molecular changes it may be possible to use genetic biomarkers to predict the likelihood of relapse after surgery. In the future these changes may be identified in biopsy material to enable better patient selection for surgery, as well as directing treatment in the adjuvant and salvage setting. Ultimately genetic or epigenetic changes may act as targets for therapy.

This thesis aims to identify molecular changes in radical prostatectomy specimens in an attempt to identify genetic alterations that predict patient outcome after surgery. The patient selection, methods and results of this experiment will be described in chapters

2,3 and 4 respectively. The experiments examining an epigenetic change in the promoter of the p 2 iWAhI/CIPI gene are described in Chapter 5.

123 u; CHAPTER 2 r

Patient Selection

Aim

The aim of this study was to address the question: In two matched groups of men who have undergone radical prostatectomy, can molecular markers be used to predict biochemical outcome? This chapter outlines the rationale for patient selection and how this process was undertaken.

2.1 Introduction

The increasing use of serum PSA as a biomarker has led to more men being diagnosed with prostate cancer when it is still clinically organ-confined (Bott et al. 2005c; Han et al. 2001) (Figure 2.1). Early detection allows radical treatment with curative intent.

However, 25-53% of men will develop prostate specific antigen (PSA) recurrence up to

10 years after radical prostatectomy (Hull et al. 2002; Pound et al. 1997; Quinn et al.

2001; Trapasso et al. 1994; Zincke et al. 1994), often as a result of micrometastatic disease present, but undetected, at the time of surgery (Roberts et al. 2001). Identifying markers that predict metastases would preclude some men from undergoing the procedure with its inherent risk of short and long-term side effects but without any evidence of a survival benefit.

124 % of patien:s assays up to 90% of potential disease recurrences can be recognised within 3 years of of years 3 within recognised be PSA can sensitive recurrences highly disease of potential advent of 90% the to With up 1997). assays al. et (Pound surgery of undetectable time but the present at presumably disease, metastatic have surgery of years 2 PSA within rising a with men of 90% over States, United the in series Hopkins Johns the Prostatectomy From Radical UCL the in time 2005c). al. et over (Bott stage database Pathological 2.1 Figure radical prostatectomy, two groups of patients were selected. One group had rapid rapid had group One selected. were patients of groups two prostatectomy, radical years, 3 after relapse who men of minority the In 2000). al. et (Stamey surgery radical te gop eand ies fe fr t es 3 er. ains ee ace fr the for matched were Patients years. 3 least at for free disease the and remained surgery of group time other the at metastases occult represent to likely progression disease after outcome poor predict to To used 1997). be al. could et markers (Pound molecular metastases whether distant investigate without recurrence, local develop 74% 1X% 90% 50% 70% 30% 40% 30% 10 20 30% U% % % yaiy iyy iy«a-iyyi 2 iy 'y iy iy iy auu iyyy iyyo iyy/ 'yyb -iyyt Year of surgery of Year 2UU1 btal 125 □ pT3a □ 1pT2o pro pT2a most important clinical predictors of outcome after surgery: radical Gleason sum score, preoperative PSA and pathological tumour stage.

126 2.2. Methods

2.2(a) The UCL Radical Prostatectomy Database

The radical prostatectomy database held at University College London contained the pathological details of 1001 men who had undergone radical prostatectomy, between

1988-2001. All specimens were examined by a single histopathologist, MCP, and there were 16 surgical contributors, operating at nine hospitals. The database was used to identify patients according to their pathological stage and once the study population was recruited the clinicopathologic-1 ------

In this series, the UICC 1997 TNM classification was used. Prostate cancer was reported as organ confined if the tumour did not extend outside the smooth muscle limits of the gland and into the periprostatic fibroadipose tissue. If tumour was close to the prostatic limit but separated by a few fibroblasts or smooth muscle cells it was still designated as gland-confined (Epstein 1990; Epstein & Sauvageot 1997).

Apical/urethral, base/bladder neck and vasal margins were assessed using the shave technique (Figure 2.2). Thios involved taking a section from the inferior and superior aspects of the apical and basal blocks, respectively. Where cancer was present in the shave limit the margin was considered positive. If the bladder neck shave margin contained tumour the specimen was classified as margin positive and stage pT4. If the apical or vasal shave contained tumour these were reported as a positive margins but this did not impact the stage.

Positive circumferential margins (anterior, lateral and posterior) were classified as either extraprostatic where tumour was in contact with the inked specimen limits having spread beyond the gland or intraprostatic when the cancer was cut through inside the glandular area of the prostate due to inadvertent surgical incision of the periprostatic

127 fibroadipose tissue and prostate capsule. The term pTx was applied to tumours that could not be assessed for stage. This included specimens with an intraprostatic positive limit and no evidence of extraprostatic spread at any other site. Eighty-one patients in whom stage was indeterminate (pTx) were excluded.

Vasal

Bladder neck

Urethral

Figure 2.2 Assessment of margin status using the shave technique.

The Gleason score at radical prostatectomy consisted of the sum of the two Gleason grades, each 1-5, from the two largest areas of cancer. At biopsy the two highest

Gleason grades were summated to give the overall biopsy Gleason score.

Tumour volume was assessed by identifying areas of tumour microscopically and outlining them in ink. A microscopic measurement grid was used to calculate the areas of all tumours, which were summated and multiplied by 0.5cm (the depth of each block).

Biochemical recurrence in this study was defined as a serum PSA level at least 0.2ng/ml and rising after radical prostatectomy (Pound et al. 1999; Salomon et al. 2003) or

0.5ng/ml (Partin et al. 1993) if a less sensitive assay had been used.

128 2.2(b) Consent

Consent for the radical prostatectomy database and for this project was obtained from the Joint UCL/UCLH Committees on the Ethics of Human Research (Study number

01/0093), in compliance with the ICH GCP (International Committee on Harmonisation

Good Clinical Practice - www.corec.org.uk). The committee stated retrospective patient consent was not required. From January 2000 individual patients were asked to consent for their clinicopathological details to be stored on the database and for tissue from their surgical specimen to be used in genetic research. Patients undergoing radical prostatectomy were provided with an information sheet and consent form to sign and return. The vast majority of men gave their consent, where consent was refused patients were excluded from the database and laboratory based research. The database was registered under the Data Protection Act 1998.

2.2(c) Identifying two groups of patients

2.2(c)l. Tumour stage

The findings from the whole radical prostatectomy database were collated. The most common pathological stage in the UCL database was pT2 i.e. pathologically organ- confined prostate cancer; of 1001 cases 435 had pT2 disease. However, in our series and from the literature an insufficient number of these patients develop biochemical recurrence for optimal matching (Babaian et al. 2001; Pound et al. 1997). The reported

PSA free rate for men with pathologically organ confined disease is 81-85% at 7-10 years (Babaian et al. 2001; Catalona, Ramos, & Carvalhal 1999; Pound et al. 1997). In the UCL series, one surgeon (RSK), from who PSA follow-up data was available, had operated on 196 men with a pT2 disease. After a median follow up of 25 months only

129 14 (7%) of these men developed a PSA rise >0.2ng/ml (Bott SRJ et al. 2002). Despite having 435 men with pT2 disease overall, there were an insufficient number of men who have relapsed to enable matching.

A total of 415 men had pT3 disease, the second most common pathological stage in the

UCL database. The literature reports between 24-82% of men with stage pT3 prostatic cancer develop biochemical relapse after 7-10 years follow-up (Catalona, Ramos, &

Carvalhal 1999; Hull et al. 2002; Pound et al. 1997). Data were available on the biochemical outcome of 143 men with pT3a from RSK, of who 49 (34%) developed biochemical recurrence after a median 25 months (range 2-89 months). It was estimated

140 men with pT3 from the whole UCL radical prostatectromy database were predicted to relapse so this cohort was selected for further investigation.

2.2(c)2. Exclusion criteria

Lymph node metastases are one of the most significant independent predictors of biochemical failure after radical prostatectomy (Epstein, Pizov, & Walsh 1993; Gervasi et al. 1989; Quinn et al. 2001; Stamey et al. 2000) and almost all patients with lymph node involvement develop biochemical relapse within 10 years of surgery (Epstein,

Pizov, & Walsh 1993; Gervasi et al. 1989). Molecular markers are unlikely to add to the prognostic data from routine histological examination in cases with lymph node metastases, as their outcome is known. Patients with lymph node disease were excluded, leaving 401 men with pT3N0 disease.

Men who received antiandrogen therapy either preoperatively as neoadjuvant therapy, or as adjuvant treatment immediately postoperatively as a result of the pathological findings, were excluded. Antiandrogen therapy causes apoptosis in the hormone

130 sensitive prostatic epithelial cells and may result in down staging of the tumour, inaccurate Gleason scoring and a reduction in serum PSA concentration - all potentially leading to inappropriate matching. Men receiving finasteride, a 5a-reductase inhibitor used in the treatment of benign prostatic hyperplasia, were also excluded as finasteride reduces PSA levels and may alter the glandular structure affecting Gleason grading.

Men receiving antiandrogen therapy as salvage treatment, after a rising post-operative

PSA, were included in the ‘relapse’ group if they fulfilled the other matching criteria.

Patients who received adjuvant radiotherapy based on adverse pathological features identified on the radical prostatectomy specimen were excluded from the non-relapse group. Men who received adjuvant radiotherapy and subsequently relapsed within 2 years were included in the relapse group as these men were still likely to be harbouring micrometastases at the time of surgery.

Patients who receive primary radiotherapy and subsequently underwent salvage prostatectomy for cancer recurrence are a selected group of men with poor prognosis disease. Furthermore, radiotherapy leads to DNA damage, which may have adversely affected the study findings; consequently these men were excluded from this study. In total 83 men were excluded as they had received either radiotherapy or hormonal manipulation.

Tumours diagnosed on TURP are more likely to represent transition zone cancer which is commonly of lower Gleason score, higher tumour volume and has different gene expression patterns compared with tumours arising in the peripheral zone (Bott et al.

2005c; Erbersdobler et al. 2002a; McNeal et al. 1988; Pavelic, Zeljko, & Bosnar 2003).

131 Tumours diagnosed after TURP were excluded so all cancers included in this study were diagnosed by prostatic biopsy.

2.2(c)3. Biochemical outcome

The UCL Radical Prostatectomy database has limited PSA follow-up uaia. num. patients who met the inclusion criteria could be identified, their biochemical outcome was not known. Furthermore, data on adjuvant and salvage treatments was required.

The notes of all 49 NHS patients were retrieved and their PSA follow-up and adjuvant therapy noted. NHS patients who were referred for surgery from another hospital were often followed up and had their PSA measured at their local hospital. In order to obtain their follow-up details, notes were retrieved and reviewed in the local hospital. If insufficient information was derived from these, the patient or their General Practitioner was contacted directly.

Private patient records were also examined in the case of three consultants (RSK, NOD,

JB). These cases were often referred from across the UK and overseas so these data were incomplete. A questionnaire was therefore devised and sent to all 440 private patients operated on by the most prolific private surgical contributor, RSK. This asked if patients had achieved a PSA nadir of <0.2ng/ml within six weeks of surgery and whether the PSA had subsequently risen above 0.2ng/ml. Patients were asked whether they had received radiotherapy, hormonal manipulation or if they were taking the herbal remedy PC-SPES. As part of a separate study addressing the impact of radical prostatectomy on erectile dysfunction, continence and quality of life further questions derived from the Short Form 36, International Index of Erectile Function and

International Continence Society-male questionnaires were included. If a response had not been received within 6 weeks, a further questionnaire was sent. A final response rate

132 of 81% (360) was achieved including 145 men had pT3N0 disease. Patients with pT3N0 disease who did not respond to the questionnaire were contacted directly. The questionnaire results have been presented (Bott SRJ et al. 2002; Thomas K et al. 2002).

Once biochemical outcome and adjuvant therapy status was established patients were assigned to either a ‘relapse’ group or a ‘non-relapse’ group. The relapse group consisted of patients who developed biochemical relapse, defined as a PSA of

>0.2ng/ml (Aus et al. 2003; Pound et al. 1999; Salomon et al. 2003) or >0.5ng/ml if a less sensitive assay was used (Partin et al. 1993), within two years of surgery. The non­ relapse group were defined as patients whose serum PSA level remained undetectable for at least 3 years after radical prostateetemy. Men who relapsed in the third postoperative year were excluded.

2.2(c)4. Matching Matching2.2(c)4.

The relapse and non-relapse groups were matched for pathological tumour stage, radical

Gleason grade and pre-operative PSA. Pathological stage was split into pT3a and pT3b and men matched accordingly. The Gleason grade was divided into Gleason sum 5, 6, 7 or 8. The two components of the sum score were matched where possible. The preoperative PSA was grouped 4-9.9, 10-19.9 and 20ng/ml or above. This is in line with recognised inflection points for prognosis in pre-treatment PSA levels (Quinn et al.

2 0 0 1 ).

To avoid bias once patients had been identified from the database they were given a randomisation number and the study investigators were blind which numbers were allocated to a particular patient until the final analysis.

133 Other possible prognostic factors examined but not included in the matching process included the surgical margin status, tumour volume and the presence of lymphovascular invasion.

A positive surgical margin implies incomplete tumour resection at the time of radical prostatectomy and most authors consider positive margins an independent predictor of disease outcome (Cheng et al. 1999; Epstein et al. 1996; Ohori et al. 1995). Positive margin status maybe sub-classified into intra or extra prostatic, focal or extensive - though a universally agreed definition of each of these has not been established

(Bostwick et al. 2000). Furthermore, the site as well as the extent and multifocality of positive margins all affect PSA outcomes (Blute et al. 1997; Obek et al. 1999). Due to the complexity surrounding positive margins and the lack of agreed definitions patients were not matched for surgical margin status. Statistical analysis was undertaken to examine whether differences existed with respect to margin status between the two groups.

Tumour volume as a prognostic indicator has been extensively studied but its importance remains to be validated in statistically robust trials (Bostwick et al. 2000;

Epstein et al. 1993; Stamey et al. 1993). Tumour volume was not included in the matching criteria although statistical analysis was undertaken to examine any association with biochemical outcome in this cohort.

Some authors have found the presence of lymphovascular invasion is an independent prognostic predictor after radical prostatectomy (Stamey et al. 1999), though this is not universally accepted (Bostwick et al. 2000). Differences between the two groups with

134 respect to lymphovascular invasion were examined statistically but were not included in the matching criteria.

2.2(d) Statistical Analysis

Matching two groups of patients for the main prognostic features limits the impact of confounding variables and increases statistical power. For this reason, in a matched study the sample size required to show statically significant differences is smaller than a comparable randomised study. The disadvantages of using matched groups for this study include:

• between participant difference still exists (e.g. PSA levels not identical in

matched pairs)

• uninformative locus in one individual results in the loss of a pair of subjects

• additional potential prognostic variables are not included in matching (e.g.

tumour volume or lymphovascular invasion).

The outcome variable for the allelic imbalance experiments was qualitative binary - either the patient has or does not have allelic imbalance at an informative locus. Where an individual was homozygous or uninformative they were not included in the statistical analysis for that locus. The McNemar test is used for paired or matched samples in which the outcome variable is binary. A significance level of 5% (p=0.05) was set and a power calculation of at least 80% required. The power is the probability of finding a difference if one truly exists (Appendix 1).

To establish the minimum sample size a realistic expectation of what difference in allelic imbalance between case and control was set at 40%. The statistical power increases with greater subject numbers. For example, if 20 paired cases are matched a power value of >80% cannot be achieved. If 30-paired cases are matched and 70% of

135 test cases and 30% of controls (a difference of 40%) have allelic imbalance at a locus

the power value is 86%. If the same 30-paired cases have allelic imbalance at a single

locus in 70% of test cases and 40% of controls (a difference of 30%) then the power

value is 62% and a significant difference between the test and control groups is not

evident. To demonstrate a significant power difference between the two groups a

sample size of at least 30 matched pairs was required using the McNemar test. This

statistical test was selected at the outset and before patient matching was undertaken.

From 413 men with pT3a prostate cancer two groups of only 31 men could be

identified. Men with an uninformative locus or in who a result was unobtainable for / j experimental reasons could not be included such that the 30 matched pairs required for

the McNemar test could not be achieved for any locus. ^

y c O z .

Fisher's exact test is a statistical test used to determine if there are non-random

associations between two categorical variables. It has the advantage that it does not

make any approximations, and so is suitable for small sample sizes (expected frequency

in any of the cells is <5). However, it assumes both groups are independent (unmatched)

and so the increased statistical power gained by paired matching in this series was lost.

Fisher’s exact test was used to assess differences in allelic imbalance between the two

study groups. The Fisher’s exact test allowed all results to be analysed irrespective of

whether the matched pair was informative.

Chi square test is a method to assess statistical significance for bivariate tabular analysis

and provides the degree of confidence in accepting or rejecting a hypothesis from non-

parametric data (e.g. lymphovascular invasion). The paired /-test is a parametric test to

compare means and determines whether two matched groups differ from each other in a

significant way under the assumptions that the paired differences are independent and normally distributed (e.g. patient age). The Wilcoxon Signed Ranks is non-parametric

136 alternative to the paired t test and compares medians, as it is a non-parametric test ordinal data may be analysed (e.g. preoperative PSA).

137 2.3 Results

2.3(a) The Radical Prostatectomy Database

The radical prostatectomy database contains clinicopathological features of 1001 men who have undergone radical prostatectomy between 1988-2001 (Bott et al. 2005c). The surgeons contributed the following number of radical prostatectomy specimens each: 44

MB, 31 JB, 3 AC, 42 ME, 510 RK, 201 NOD, 1 CO, 1 AP, 1 JP, 38 JS, 122 PS, 7 PW.

Seven hundred and forty two radical prostatectomies were performed in the private sector and 259 in the National Health Service. The selection criteria (preoperative PSA and biopsy Gleason grade) and pathological outcome (pT stage, radical Gleason score and tumour volume) did not differ significantly between the NHS and private sectors.

Although the mean age at surgery did show a statistically significant difference, 63 years (NHS) and 61 years (private sector) (paired t test, p < 0.01) this was not of major clinical importance.

The preoperative features and pathological variables of the whole UCL database are shown in Table 2.1 and the incidence and site of positive surgical margins is summarised in Table 2.2.

138 Table 2.1 Preoperative and pathological features of the UCL Radical

Prostatectomy database.

Age (years) Mean 62 SD 6 Pre-op PSA (ng/ml) Minimum 0.10 Median 8.00 Maximum 146.00 Biopsy Gleason sum Minimum TSTG score Median 6.00 Maximum 10.00 Tumour volume (mis) Minimum TSTM Median 1.40 Maximum 30.60 Pathological stage pTO 3 (<1%) / pT2a 94 (9%) pT2b 341 (34%) V pT3a 311 (31%) pT3b 104 (10%) pT4 67 (7%) pTX 81 (8%) Number of patients =1001, TSTG = too small to grade,r'STM = too small to measure

Table 2.2 The number and site of positive surgical margins from The UCL Radical

Prostatectomy database.

All m argins -ve 465 (48%) 1 margin +ve 336 (35%) Apical 215 (22%) Circumf-CIP 71 (7%) Circumf - CEP 31 (3%) Circumf- CIP+CEP 5 (<1%) Bladder neck 10 (1%) Vasal 4 (<1%) 2 margins +ve 135(14%) CIP + Apical 69 (7%) CEP + Apical 34 (4%) Others 32 (3%) 3 margins +ve 26 (3%) 4 margins +ve 5(1%)

CIP = Circumferential intraprostatic margin CEP = Circumferential extraprostatic margin

139 2.3(b) Patient matching

Four hundred and fifteen patients had pT3 disease of which 14 had lymph node

metastases and were excluded. Eighty-three patients who received preoperative

treatment with antiandrogens, finasteride or radiotherapy were also excluded. Of the

remaining 307 patients, 17 relapsed in their third postoperative year and so were

ineligible for either the relapse or non-relapse groups. Of the remaining contactable men

there were 91 men with adequate follow-up. Forty-four men had developed PSA

recurrence within 24 months of surgery and 47 had remained PSA free for at least 36

months without receiving adjuvant therapy. These patients were matched for stage

(pT3a or pT3b), Gleason sum score (5,6,7,8) and the constituent scores where possible,

and PSA (4.1-10, 10.1-20 ng/ml). Two groups of 31 patients were matched (Table 2.3);

all had been diagnosed with prostate cancer as a result of prostatic biopsy.

140 Table 2.3 Patient matching

7 i. vJUbuljb * Reiapao ^ n J t e i a p ^ ' ...... 1 32 3+4=7 3+4=7 14.2 11.3 pT3a pT3a 9 40 6 10 4+3=7 4+3=7 16.7 18.5 pT3a pT3a 6 44 4 26 4+3=7 4+3=7 10.1 12 pT3a pT3a 18 47 2 49 4+3=7 4+3=7 12 10 pT3a .._ pT3a 13 60 3 54 4+3=7 3+4=7 18 11 pT3a pT3a 4 53 55 56 4+3=7 3+4=7 14 16.6 pT3a pT3a 7 49 42 21 4+4=8 4+4=8 15.7 16.3 pT3a pT3a 19 47 16 30 4+3=7 3+4=7 8 8.8 pT3a DT3a 17 39 25 13 4+3=7 3+4=7 8.6 5.9 PT3a pT3a 11 64 58 57 4+3=7 3+4=7 16 10.5 pT3a pT3a 21 48 59 9 3+4=7 3+4=7 8 8 pT3a pT3a 6 56 27 12 4+3=7 4+3=7 7.5 8.8 pT3a pT3a 16 66 60 5 4+3=7 3+4=7 8.7 7.9 pT3a pT3a 3 42 52 29 4+3=7 3+4=7 5.9 7.4 pT3a PT3a 5 37 11 62 3+3=6 3+3=6 4.3 9.3 pT3a pT3a 3 70 38 48 2+4=6 4+2=6 10.2 12.9 pT3a pT3a 12 62 18 33 3+2=5 3+2=5 18.6 16 pT3a PT3a 18 70 7 23 2+3=5 2+3=5 14.7 17.8 PT3a pT3a 9 60 79 77 3+5=8 4+4=8 9.7 7.5 pT3a pT3a 12 46 22 28 2+3=5 2+3=5 8.6 7 PT3a pT3a 16 87 15 17 2+4=6 3+3=6 9.1 5.1 pT3a pT3a 20 62 24 19 4+3=7 3+4=7 16 8.3 pT3a pT3a 21 71 43 44 4+3=7 3+4=7 5.2 5.5 pT3a pT3a 3 53 45 20 3+4=7 3+4=7 9.8 6.4 pT3a pT3a 8 59 64 61 5+3=8 4+4=8 14.4 15.1 pT3a pT3a 11 48 75 37 4+3=7 3+4=7 15.8 13 pT3a pT3a 5 45 74 31 3+4=7 3+4=7 9.9 9.7 pT3a pT3a 3 36 50 51 4+3=7 3+4=7 15.7 17 pT3a pT3a 15 36 35 34 3+2=5 2+3=5 5.1 9 pT3b pT3b 3 36 78 53 4+3=7 4+3=7 12.4 16.6 pT3b ...... 5______57 39 36 4+3=7 4+3*7 18 13.4 pT3b pT3b 6 55

Median 10.2 Median 10.0 Median 9 Median 53 (range 5.1-18.6) (range 5.1-17.8) (range3-21) (range 36-87) 2.3(c) Comparison between the groups

The complete clinicopathological details of the matched groups are in Appendix 3. The mean age at the time of surgery was 59.6 (SD 6.1) years in the relapse group and 64

(SD 6.2) years in the non-relapse group (paired t test, p<0.01). Patients for pathological stage and Gleason sum score. It was not always possible to match patients exactly for their constituent Gleason grades. Fifteen (48%) were matched exactly. In 12 matched pairs the constituent scores were the same but the relapse group had the higher grade in the predominant tumour (4+3 vrs 3+4 in 11, 3+2 vrs 2+3 in 1), in one matched couple the constituent scores were the same but the predominant

Gleason pattern was higher in the non-relapse group (4+2 vrs 2+4). In the remaining 3 patients, the relapse group vrs the non-relapse were 3+5 vrs 4+4, 2+4 vrs 3+3 and 5+3 vrs 4+4, respectively.

The median preoperative PSA was 10.2ng/ml (range 5.1-18.6) in the relapse group and lO.Ong/ml (range 5.1-17.8) in the non-relapse group. There was no significant difference between the two groups with respect to PSA (Wilcoxon Signed Ranks test p=0.39).

The median PSA free follow-up for the non-relapse group was 53 months (range 36-

87), most men were PSA free for well in excess of the stipulated 36months. The median time to PSA recurrence in the relapse group was 9 months (range 3-21).

Surgical margin status was not included in the matching criteria; 20 patients in the recurrence free group and 21 patients in the relapse group had at least one positive margin. Statistical analysis showed neither the site (apical, C-IP or C-EP - Table 2.4)

142 nor the number of positive surgical margins (negative, solitary or multiple margins -

Table 2.5) was significantly different between the two groups (Wilcoxon Signed Ranks test p=0.11 and p=1.0, respectively).

143 Table 2.4 The site of solitary positive surgical margins

(Wilcoxon Signed Ranks test, p=0.11)

Non-relapse Relapse Site of solitary positive margin 11 10 Margin negative Urethral/Apical only (U) 14 9

Circumferential- 1 1

intraprostatic (C-1P)

Circumferential- 1 3

extraprostatic (C-EP)

Table 2.5 Patients with 0,1 or >1 positive surgical margins

(Wilcoxon Signed Ranks test, p=1.0) ______Number of positive No recurrence Recurrence

margins

11 10 Margin negative Solitary (U or C-EP) 15 12

Multiple (U and C-EP) 1 5

C-IP (excluded) 4 4

Total 31 31 In this matched series, 2 patients had lymphovascular invasion in the non-relapse group compared with 9 in the relapse group - the relapse group were significantly more likely to have lymphovascular invasion (Chi square, p=0.02). Most large series do not report the prognostic significance of lymphovascular invasion, while two found vascular invasion to be a significant but not independent predictor of outcome (Babaian et al.

2001; Stamey et al. 2000).

The median tumour volume was 3.0mls (range 0.5-10.5) in the relapse group and 1.8mls

(range 0.8-8.4) in the non-relapse group. There was significant difference between the tumour volumes of men with disease relapse versus men without PSA recurrence

(Wilcoxon Signed Ranks test p=0.01).

While both lymphovascular invasion and tumour volume are not proven to be prognostic factors, in this small series there is a significant difference between these factors in the relapse and the non-relapse groups. In univariate analysis, tumour volume was not an independent predictor of PSA relapse within two years of surgery (p=0.063), but lymphovascular invasion was (p<0.05). The clinical significance of this finding in this series in not known.

2.3(d) The unmatched patients

The remaining unmatched patients tended to have either lower stage, grade and preoperative PSA if they had not relapsed or and a higher stage, grade and preoperative

PSA in they had relapsed (Table 6).

145 Table 2.6. The stage, grade and preoperative PSA group for the unmatched patients (PSA group A=4-10ng/ml, B=10.1-20ng/ml and C= >20ng/ml) Unmatched non-relapse group Pathological stage Gleason grade Preoperative PSA pT3a 3+4=7 A pT3a 3+4=7 A pT3a 3+2=5 A pT3a 3+2=5 A pT3a 2+3=5 B pT3a 2+3=5 B pT3a 3+3=6 A pT3a 3+3=6 A pT3a 3+3=6 B pT3a 3+3=6 A pT3a 3+2=5 A pT3a 3+3=6 A pT3a 3+3=6 B pT3a 4+3=7 A pT3a 2+4=6 A pT3a 3+3=6 A

Unmatched relapse group Pathological stage Gleason grade Preoperative PSA pT3b 4+4=8 B pT3b 3+4=7 B PT3b 4+5=9 B pT3b 4+3=7 B pT3b 4+3=7 C pT3b 4+3=7 B pT3a 4+3=7 B pT3b 4+3=7 A PT3b 4+3=7 A pT3b 4+5=9 A pT3b 4+5=9 B pT3b 3+4=7 B pT3b 3+4=7 A

146 I' v vf - , y ) , 0 ^ 2.4 Discussion ^HT yft W7- V , k- ^ ' ‘ y ' U^ A \) The Radical Prostatectomy Database at UCL identified men by stage and full pathological details with respect to the biopsy and radical prostatectomy findings were available. However, without adequate PSA follow-up or details of adjuvant therapy the

biochemical outcomes could not be ascertained. A self-completed questionnaire was

sent to all patients operated on by the highest volume surgeon, clinical notes were

reviewed and patients or their General Practitioners were contacted directly. Using this

data two groups were matched for the molecular study.

It may be surprising that from a large series of over a thousand patients more men could

not be included in the study. There are several reasons why this was not the case.

Firstly, during the first six years (1988-1993) less than 20 radical prostatectomies were

performed per year. From 1994 the number of operations performed and submitted to

the database rose dramatically. In 1995 forty-one were submitted, in 1996 ninety-six,

1999 one hundred and fifty one and in 2001 just over two hundred. While the total

number in the database represents a large series the number with 3-year follow-up was

limited.

Secondly there was a propensity for men with poor prognosis tumours, based on the

pathology report, to be treated with adjuvant therapy. Indications for adjuvant therapy

are variable (Bott & Kirby 2002), though positive surgical margins frequently triggered

adjuvant radiotherapy. From the UCL Radical Prostatectomy database 465 (52%) men

had positive surgical margins. A significant proportion of these men had adjuvant

radiotherapy and were excluded.

147 Pre-operative androgen ablation was shown in the mid 1990’s to reduce the incidence of positive margins and extracapsular disease after radical prostatectomy (Goldenberg et al. 1996; Soloway et al. 1995; Van Poppel et al. 1995; Witjes, Schulman, & Debruyne

1997). At the time these findings were reported, the proportion of men in this series receiving androgen ablation was at its highest (24% in 1996). More recently, series with longer follow-up have reported no benefit in terms of overall survival (Goldenberg

& et al 1997; Lee, Warde, & Jewett 1999; Soloway & et al 1997). The use of these agents in the UCL series mirrored these findings and declined in the latter years, however the patients treated with neoadjuvant therapy tended to be the men who had the longer follow up.

Half the patients in the series were operated on by one surgeon in the private sector.

Many of these men were referred from across the UK, Europe and the Middle East.

After their radical prostatectomy, their PSA follow-up was performed locally. While some were traced many were lost to follow-up.

The matching criteria were based on the most widely accepted features that predict progression after radical prostatectomy (Bostwick et al. 2000). Patients were grouped according to accepted inflection points for prognosis with respect to PSA (4-9.9, 10-

19.9 or 20ng/ml), Gleason sum (5, 6, 7 or 8) and pathological stage (pT3a or pT3b). As these are strong predictors of biochemical outcome, identifying men with poor prognostic features who had not relapsed and vice versa proved difficult.

148 2.5 Summary

The UCL radical prostatectomy database is the largest of its kind in the UK, with complete clinicopathological details on over 1000 men who had undergone radical prostatectomy. Despite this the cohort size was insufficient to allow matching of the required number of relapse and non-relpase pairs for this study, because of insufficient length of follow-up and exclusion of patients who had nodal metastases and treatments that might affect the matching features. Future studies may enable more patient matching as the duration of follow-up becomes longer and the strict matching criteria could be broadened to include a wider range of relapse and non-relapse men.

149 CHAPTER 3

Materials and Methods

Aim

This chapter will outline how the study DNA was extracted and amplified and the genotyping technique.

3.1 The pathological specimen and tumour dissection

The radical prostatectomy specimen was received by MCP in 10% formal saline. The specimen was weighed and laterality indicated by different coloured inks. The prostate was sliced between parallel glass rods of 5 mm diameter, producing a series of 4-6 coronal sections and embedded in paraffin (Figure 3.1a). A 5 pm ‘wholemounf section was taken from each coronal slice, stained and examined by MCP.

The existence of allelic imbalance (AI) in tumour tissue may be masked if DNA from normal cells contaminates the cancer DNA. Areas of tumour within a radical prostatectomy wholemount also contain areas of normal stromal tissue (fibroblasts and smooth muscle cells) and benign prostatic glands. The wholemount containing the most tumour was selected and the largest area containing at least 70% malignant cells was marked on the cover-slip by MCP (Figure 3.1b). Other authors have used areas containing 50-70% cancer by area (Cunningham et al. 1996; Macoska et al. 1995;

Pesche et al. 1998; Uchida et al. 1999; Wang, Parsons, & Ittmann 1998; Wu et al.

2001), 70% was used in this study to maximise the proportion of tumour DNA. The paraffin block corresponding to the slide was retrieved together with the lymph node paraffin block. In cases where a lymphadenectomy had not been performed or no lymphoid tissue was identified in the ‘lymph node’ specimen, seminal vesicle tissue was

150 used (patients 6, 27 and 37) - in these three cases the seminal vesicles were free from tumour involvement.

Once the block containing the tumour and control tissue had been identified a further four to six 5pm sections were taken using a microtome (Figure 3.1.c) and floated on water at 40°C. A microscope slide partially submerged at an angle of 45° was used to lift the section from the water and left to dry. The slide and section was placed in an incubation oven at 60°C for 20min to attach the specimen to the slide. The specimens were dewaxed in 100% xylene, rehydrated in ethanol and stained with haemotoxylin and eosin (Appendix 5.)- Slides were passed through an ethanol series and left to dry overnight.

The cover-slipped master slide was placed over the sample slide so that the wholemounts were aligned. The H&E staining facilitated this alignment in the prostate

sections (malignant areas stained strongly with the nuclear staining haematoxylin) as

well as identifying nodal tissue from adipose tissue in the lymph node sections. Both

nodal and adipose tissues contain ‘normal’ DNA. The DNA yield is significantly greater

from lymph nodes as lymphocytes have a high nuclear to cytoplasmic ratio.

The areas of cancer marked on the master slide were then transcribed by marking the

underside of each sample slide with a fibre-tip pen. Areas of cancer of >0.5cm2 were

scraped off using a number 20 scalpel blade within the region marked. The scrapings

were prone to ‘blow-away’ or their static charge adhered them to the scalpel blade.

Therefore, for ease of handling, a drop of 100% ethanol was added before samples were

scraped. The scalpel blade was changed between each patient and between tumour and

151 control slides to avoid contamination. The dissected sample was placed in a labelled

Eppendorf™ tube and placed at 37°C for 15 mins to allow the ethanol to evaporate.

Figure 3.1 a) Coronal block of prostate embedded in paraffin, b) master slide with area of prostate containing at least 70% malignant cells marked, c) 10pm section unstained, d) section stained, tumour area marked and excised with scalpel blade.

152 Where the tumour area was small (<0.5cm2) the malignant tissue was obtained by laser capture microdissection (LCM, Arcturus) (patients 12 and 28). For LCM sections were soaked in 70% xylene for 3 mins followed by 100% xylene for 3 minutes to facilitate transfer. The area of tumour was marked (Figure 3.2a) and a polymerising film attached to an Eppendorf™ cap was applied to the section. The section was viewed microscopically and the cells of interest orientated to the centre of the field of vision. A

45mW pulsed laser beam was activated and produced a 30pm diameter spot on the transfer film immediately above the cells of interest. At this precise location the laser heated and focally polymerised the film, thus fusing the film with the underlying cells

(Figure 3.2b). The transfer film, fused with the selected cells, was lifted off the tissue section. This technique allows malignant acini and even individual malignant cells to be dissected from the section.

Figure 3.2 Laser capture microdissection

a) Area of tumour marked

b) Tumour captured on polymerised film

c) Defect left on slide

153 3.2 DNA extraction

DNA extraction was performed using the Qiagen DNeasy™ kit. This involved: lysis of the cells, binding of DNA to a silicone membrane, removal of proteins and divalent cations and the elution of DNA, in fragments of 50-30kb. The protocol was as follows:

1. To each of the tumour and control samples 180pl of buffer ATL and 20pl of

proteinase K was added. Buffer ATL was an enzymatic buffer and also

contained the detergent sodium dodecyl sulphate. Together these performed an

enzymatic and chemical digestion to lyse the connective tissue between cells and

the cell membrane. Lysis occurred over 16-20 hours at 55°C and the samples

were vortexed intermittently.

2. To the samples 200pl of buffer AL was added. This solution was votexed and

incubated at 70°C in a water bath for 10 mins. Buffer AL was a chaotrophic salt,

which dehydrated the DNA and exposed the negatively charged phosphate

backbone to optimise binding to the column. It also inactivated proteins and

carbohydrates in the solution.

3. 200jnl of ethanol (96-100%) was added, to further dehydrate the DNA and

expose the phosphate backbone

4. The solution was pipetted onto the spin columns provided and centrifuged at

8000rpm for 1 minute. The spin column contained a positively charged silicone

membrane that bound to the negatively charged phosphate backbone of the

DNA.

5. Two wash steps were performed. 500pl of buffer AW1 (a dilute chaotrophic salt

solution removed residual protein and carbohydrate) was added and the column

centrifuged at 8000rpm for one minute and the eluate was discarded. 500pl of

154 buffer AW2 (dilute ethanol solution) was added and the column centrifuged at

8000rpm for 3 minutes to dry the silica membrane.

6. The column was placed in a clean labelled collecting tube and the DNA was

eluted by adding lOOpl of buffer AE and centrifuging at 8000rpm for one

minute. This step was repeated. Buffer AE is a simple alkaline buffer (pH 8-9)

which bound to the acidic DNA.

7. The sample DNA was stored at -20--30°C.

When the DNA from all 124 tumour and control samples had been extracted, aliquots of

50pl of DNA were placed in an ABgene DNA block 96 plate and sealed with the corresponding Cap mat™ and Parafilm™ (Appendix 6).

155 3.3 DNA amplification

3.3(a) Principles of the polymerase chain reaction

The polymorphic regions of DNA to be studied were amplified using the polymerase chain reaction (PCR), to increase the quantity of study DNA. This is a cyclical process of heating and cooling to denature, anneal and enzymatically amplify DNA. Specific

DNA sequences can be multiplied exponentially using a system that combines a computerized thermal cycler with a heat stable enzyme (Taq polymerase). The standard reaction uses two oligonucleotide primers that are complementary to and hybridize with opposite DNA strands flanking the region of interest in the target DNA. The primers are generally around 20-30 nucleotides in length and sufficiently long to be unique within the genome. The reaction consists of the following steps:

template denaturation at 90-94°C for 30-120 seconds - this separates the two DNA

strands

primer annealing at 50-60°C for 30-60 seconds - the primers anneal to the template

extension at 72°C for 30-120 seconds; from the sense and antisense primer on

complementary DNA strands, nucleotides in solution bind to their complementary

base, catalysed by the enzyme Taq polymerase.

These steps are repeated 20-60 times depending on the quality and length of the DNA to be amplified and the efficiency of primer annealling. The cycling means that the newly synthesized DNA strands can themselves act as templates for further rounds of DNA synthesis in the subsequent cycles of the reaction. The DNA is therefore amplified exponentially, theoretically 2" times where n is the number of cycles, although in practice the efficiency is not 100%. Typically, 30-40 cycles are performed and, although

156 it is easier to amplify small fragments of a few hundred base pairs, up to lOkb can be amplified.

In this project, DNA was extracted from formalin fixed paraffin sections. DNA from archival tissue is fragmented, consequently the primers were designed such that the product length was kept below 275 base pairs.

3.3(b) Microsatellite marker selection

PubMed and MedLine searches were performed and 184 papers were identified relating to ‘prostatic neoplasia’ and ‘loss of heterozygosity’ or ‘allelic imbalance’. These papers were reviewed and the most frequent sites of Al in prostate cancer noted. Al is the most frequently found genetic alteration in prostate cancer (Dahiya et al. 1997c) and is reported consistently on chromosomes lq, 3p, 5q, 6q, 7q, 8p, 9p, lOp, lOq, lip , llq ,

12p, 13q, 16q, 17p, 17q, 18q and 21q (Carter et al. 1990; Cooney et al. 1996a; Dahiya et al. 1997b; Dahiya et al. 1997c; Elo et al. 1999; Emmert-Buck et al. 1995; Fromont et al.

2003; Gao et al. 1995; Hugel & Wemert 1999; Kawana et al. 1997; Macoska et al.

1995; Narla et al. 2001; Saric et al. 1999; Smith et al. 1996). The sites most frequently lost in somatic cancers include 8p, lOq, 13q, 16q and 18q (Carter et al. 1990; Elo et al.

1999; Emmert-Buck et al. 1995; Fromont et al. 2003; Hugel & Wemert 1999; Macoska et al. 1995; Saric et al. 1999) and losses on lq are associated with hereditary prostate cancer (Smith et al. 1996). Some losses, such as those on 8p and 13q, are associated with early prostate cancer (Emmert-Buck et al. 1995; Fromont et al. 2003; Macoska et al. 1995; Saric et al. 1999). Others, such as those on 1 lp and 16q (Carter et al. 1990; Elo et al. 1999; Kawana et al. 1997; Saric et al. 1999; Suzuki et al. 1996), are more frequent in advanced disease.

157 Fluorescently labelled microsatellite markers for the most common polymorphic regions that are lost in all stages of prostate cancer were identified (Appendix 7) and obtained and are shown in Table 3.1. Markers were selected where a heterozygosity rate of

>70% was reported. Several markers were obtained which despite attempts over a range of conditions failed to generate a PCR product. These included D8S549 on 8p22,

D9S2136 on 9p (p i6), D16S413 on 16q, D17S855 on 17q, D17S960 on 17p, KLF6M1 and KLF6M2 on lOp.

158 Table 3.1 Details of microsatellite markers

Marker Sequence Size Het Locus D1S158 GGGCCTT CTT AT ATT GCTT C 137-163 0.89 1q25.2-31.1 GGAAAGACT GGACCAAAGAG D1S422 CAT GGGGTAT AGC AACAG AC 160 0.76 1q31.2-42.3 T GATTT CCTGCAAACATTTT D1S305 CCAGNCT CGGT AT GTTTTT ACT A 156-176 0.8 1q21.1-21.2 CT GAAACCTCTGTCCAAGCC D1S414 GCACAGTTCAACATCCATT 195 0.78 1q31.3-32.3 T CT CT GT CATTTT AGGT CT ATTT CT D2S222 ACAAAT GCAG AAAAAGC AT AT G 118-138 0.9 2q23.1 CT GT CAGGCT G AGG AAATTT D5S500 ACCTATTCGACCTAATGACTAAAGA 213 0.84 5q31.1 ATCGGT G AAAT GCAACT ACTT D5S656 GCT AAG AAAAT ACG ACAACT AAAT G 189 0.75 5q21.3 CAT AAT AAACT G ATGTT G AC AC AC D6S501 GCT GGAAACT GAT AAGGGCT 173 0.75 6q16.3-21 GCCACCCT GGCT AAGTT ACT 06S314 AAAAT GACTT CTTT GGGTGGGC 243-259 0.81 6q16.3-27 GTGGGT AGCAACACT GTGGC D7S523 CT GATT CAT AGCAGCACTT G 224-240 0.81 7q31.1-31.33 AAAACATTT CCATT ACCACT G D7S480 CTT GGGGACT G AACC AT CTT 189-206 0.87 7q31.31 AGCTACCATAGGGCTGGAGG D8S1991 GT GAAGGAGGGCAGTCAT 200-220 0.71 8p22 CAGGGTT GAAGCAAT CT G D8S136 GCCC AAAGAGG AGAAT AAA 71-89 0.88 8p12-21 CT GTTT CCACACCGAAGC NEFL GCAGT AGT GCCGCAGTTT C A 137-147 0.81 8p21.3 T GCAATT CAT CTT CCTTT CT 01OS1246 CT ACGGACT CATT G AAG ACT AGG 219 0.8 10q25.1 AGCGTTTT CT AT AGCT CT G ACG D10S587 CCCAG ATT CAT GGCTTT C 172-186 0.8 10q25.3 TT CT GCT GACACGGGC D10S211 CT CCT GGT CT CAT GCG 195-211 0.84 10p12.1 CAGGCT CCT ACT ACCGT C D11S902 CCCGGCT GT G AAT AT ACTT AAT GC 145-163 0.8 11 p15.1 CCCAACAGCAAT GGGAAGTT D11S916 CAGACTATTCTCATT GCTGC 135-153 0.73 11q14.1 GG ACTT CT AAGCCT CCAT AA D11S2000 AGT AGAGAACAAAACACTGTGGC 213 0.87 11q22.3 TTT GAAGATCTGT GAAATGTGC D12S89 ATTT GAGAGCAGCGT GTTTT 254-288 0.79 12p13.1 -13.31 CCATT ATGGGGAGTAGGGGT D12S1697 CAT CTT GGCCCAGT CAAT 218-234 0.83 12p12.3 CCTT CT GTTT AT AGC AAT GGGA D13S269 T GT CTT CCAGCAGGGC 116-134 0.71 13q21 CAAAGT GGTT CAT CTT GGT CT D13S171 CCT ACCATTGACACT CT CAG 227-241 0.71 13q12.3 T AGGGCCAT CCATT CT D13S263 CCT GGCCT GTT AGTTTTTATT GTTA 149 0.84 13q14 CCCAGT CTT GGGT AT GTTTTT A D15S1232 CC AG AG AG AT CTTT CCCCAT 289 0.84 15q13.1 TT GCT CCACT GTTTT CT CAC D16S413 ACT CCAGCCCGAGTAA 131-149 0.85 16q24.3 GGT CACAGGT GGGTT C D18S541 CT CCCAAAT AT GG AAT GG AA 275 0.75 18q21.1-21.3 T GAGCT GAGAT CAT AT CAATGC D19S223 CAAAAT CG AGGT GC AT AGAA 228-246 0.82 19q13.2 ACCAT GACT GGCT AATT GT G D21S156 GT CAACAT AGT G AG ACCCCA 77-107 0.84 21q22.3 ATCCAGCCT GT AACACATT C

159 3.3(c) PCR optimisation

The PCR reaction consists of the following reagents:

• lOx reaction buffer - a commercially available solution containing 750mM Tris-

HC1 (pH 8.8), provides optimal pH to prevent dNTP hydrolysis.

• dNTP - dinucleotide triphosphates (A,T,G and C)

• MgCb - required for efficient enzymatic activity of Taq

• sense primer (forward microsatellite marker)

• antisense primer (reverse microsatellite marker)

• Distilled water 18Q

• DNA

• Taq - A heat stable DNA polymerase enzyme derived from Thermus aquaticus

bacterium.

The concentration of lOx buffer, dNTP and Taq is standard, however the concentration of the other components and the PCR conditions was optimised for 3 sample markers -

D10S211 (200bp), D16S413 (140bp), D21S156 (lOObp) of varying size.

3.3(c)l. DNA concentration

The total quantity of patient DNA available was limited. The minimum volume of DNA required for PCR was checked at several DNA dilutions (1, 0.5 and 0.25pl per 15pl reaction mix, respectively). A DNA volume of 0.5pl per 15pl reaction mix yielded sufficient product, a lower concentration did not. All 124 samples were tested, where there was inadequate product on two or more occasions the sample was replaced, either by repeating the DNA extraction process if sufficient tissue was available or by using another appropriately matched patient’s DNA. These replacement samples were also tested.

160 3.3(c)2. M arker concentration

The markers supplied were diluted to a concentration of 5mM and the optimal microsatellite marker concentration for five microsatellite markers (D7S523, D10S211,

D8S549, D11S916, D13S263) was evaluated using 2.5, 5, 8 and 12 picomoles per 15pl reaction (0.5, 1.0, 1.6 and 2.4pl of 5mM microsatellite marker) for each of the sense and the antisense markers (Figure 3.3). Above 8 picomoles per reaction there was considerable marker dimerisation, indicating an inefficient reaction, below this concentration the product yield was impaired. Eight picomoles per 15pl was therefore used in subsequent reactions.

Figure 3.3 Agarose gel of primer D7S523 at 2.5, 5 and 8 picomoles per 15pl reaction in DNA from unmatched patients.

8 picomoles D7S523

The optimal microsatellite marker annealing temperature is affected by the proportion and number of guanine-cytosine (3 hydrogen bonds) and thymine-adenine (2 hydrogen bonds) bonds formed. An annealing temperature of 55, 58, 60 and 62°C was tested;

58°C yielded the greatest product (Figure 3.4).

3.3(c)3. Magnesium chloride concentration

Using the three test microsatellite markers and unmatched patient DNA, the optimum magnesium chloride concentration was established using 22.5, 30, 37.5 and

161 45micromoles per 15pl reaction. The optimal concentration was 37.5 micromoles (1.5pl

of 25mM MgCb per 15pl reaction or 2.5mM) (Figure 3.3).

Figure 3.4 Agarose gel showing optimisation with DNA from 8 unmatched patients

at marker D10S211. MgCh concentration 22.5, 30, 37.5 and 45 pmoles per 15pl

reaction and annealing temperatures of 55, 58, 60 and 62°C were tested. The

optimal concentration of MgCh was 37.5 pmoles and maximum annealing

occurred at 58°C.

MgCh 22.5 30

* m m m ft

55°C •MgCI' 45 no DNA

MM

60°C

62°C

162 Once the optimal conditions were established all 124 samples were amplified, using a

‘cold start’ and 37 microsatellite markers. The PCR reaction conditions and master mix

were as follows:

PCR reaction mix

1 Ox buffer 1.5pl

dNTP (2mM) 1.5pl

MgCl2 (25mM) 1.5|il

Sense primer (5pM) 1.6pl

Antisense primer (5pM) 1.6pl

Water 6.65pl

Taq 0.15pl

DNA 0.5pl

Total reaction mix 15^1

PCR conditions

DNA denaturation/7tf

DNA denaturation 45 secs 95°C

Primer annealing 45 secs 53-58°C

Extension 45 secs 72°C

Final extension 10 min 72°C

Throughout the PCR stages and the genotyping two controls were used, the first where

DNA was replaced by water and the second where DNA was replaced by reaction mix.

163 3.4 Genotyping

A random selection of 8 PCR products from each 96 well plate was run on a 1.5% agarose gel to check sufficient product of the correct size had been generated. This also enabled a ‘by eye’ estimate of the quantity of product for the next stage.

Panels of markers were designed so that PCR products could be pooled and genotyped simultaneously. Products of similar size but with a different fluorescent label (FAM,

HEX or TET) could be placed in the same panel. A gap of at least 20 base pairs was required for products with the same fluorescent label to be in the same panel. Forty microlitres of di-iodinised water was added to a 96 well plate and to this 2-4pl of PCR product was added depending on the yield as assessed on the agarose gel. The plate was sealed, vortexed and centrifuged for 10 secs.

In a 1.5ml Eppendorf™ tube, 1 ml of HiDi was added to 15pl of 400 Genescan® High

Density ROX size standard. ROX is an internal lane size standard and allows high run- to-run precision in sizing DNA fragments by electrophoresis. It consists of 21 single stranded fragments of 50-400bp range. Each fragment is labelled by a single fluorophore, which results in a single peak when run on a polyacrylamide gel in the ABI

A3700.

The HiDi/ROX solution was mixed by vortexing and lOpl aliquots were pipetted into a

MicroAmp Optical 96® well reaction plate. To this lpl of pooled PCR mix was added, the 96 well plate was sealed and centrifuged briefly.

164 The reaction mix was heated at 95°C for 5 minutes. This results in the tangled DNA fragments straightening and the HiDi becoming incorporated within the DNA to maintain its new shape.

The ABI A3 700 software allows each of the 96 lanes to be labelled with a sample name.

The PCR products were therefore arranged in the 96 well plate so that the fluorescent electropherogram generated from the radical prostatectomy DNA could be compared directly with its corresponding lymph node control (Appendix 6).

The ABI A3700 is a high voltage electrophoresis circuit. The machine loads the Hi/Di,

ROX and PCR product mix into a well containing buffer and the cathode. Over a period of 50 seconds this reaction mix is injected into a capillary containing polyacrylamide gel. After the samples have been injected into the capillaries the injection wells are washed and re-filled with buffer. A 1000V potential difference is applied between a loading bar, the cathode, and an electrode, the anode, at the back of the cuvette assembly (Figure 3.4).

Electrode Capillary containing array-fill polymer Cuvette (anode) (+)

DNA fragments

Sheath-flow polymer HIGH VOLTAGE

Loading bar Injection well containing buffer (cathode) (-)

Figure 3.4 The electrophoresis circuit

165 The potential difference results in a current carried by the DNA, buffer and polymer ions. The negatively charged DNA fragments move through the polymer within the capillary towards the anode at varying speeds according primarily to their mass (number of base pairs in the product), but also their charge. When the fragments reach the end of the detection-ends tips, they are moved through a laser light path by the flow of the sheath flow-polymer, which in turn carries the current to the anode and so completes the circuit. As the fluorescently labelled fragments pass through the laser they emit a blue

(FAM), green (TET or HEX) or red (ROX) light that is detected by a camera. The amount of light detected is proportional to the quantity of product and so the amplitude of the curve on the fluorescent electropherogram.

166 3.5 Analysing the data

Preceding the allelic peaks are ‘stutter peaks’ which represent slippage of the Taq polymerase during PCR (Gray et al. 1995) (Figure 3.5). The allelic peak is the highest peak after the stutter peaks. The first allelic peak is usually higher than the second as it is made up of its own product as well as a stutter peak of the second, larger allele. The

ABI A3700 software was set so that the highest peak within the predicted size range, preceded by stutter peaks was labelled (size, amplitude and area under the curve). The next highest peak where present and preceded by stutter peaks was also labelled by the software (Figure 3.5).

The definition of allelic imbalance is based on the relative ratio difference between the two alleles from the control sample and the two alleles from the cancer sample.

However, the ratio that constitutes loss varies in the literature. One of the first definitions of Al using fluorescent electrophoresis was given by Canzian et al. (Canzian et al. 1996). They used the amplitude of the peaks in the equation below

xAI = (At2 x A nl) / (Atl x An2)

Where A=amplitude, t=tumour, n=normal, 1 =shorter allele, 2=longer allele.

A ratio of <0.6 or >1.67 equates to an allelic loss of more than 40% and was classified as Al.

More recently Narla et al. used a similar equation; for each heterozygous locus, the ratio between the smaller peak (usually the second or larger allele) and larger peak (usually the first or smaller allele) in the tumour DNA was compared to the same ratio in the normal DNA.

167 XAI= ([T1/T2]/[N1/N2])

In this way, the X ai varies between 0 (complete allele loss) and 1 (no loss). Narla et al. used an XAi value of <0.7 to define Al in this study (Narla et al. 2001).

In a study in colorectal carcinoma Barratt et al. used the same calculation as above, however, a cut off <0.5 was used to classify Al (Barratt et al. 2002). Other workers have used a lower cut-off, Gray et al. defined Al as a relative amplitude loss of >20% (Gray et al. 1995)

If a sample constitutes pure cancer DNA and one allele has undergone loss of heterozygosity, because of the clonal nature of cancer, all daughter cancer cells would also have Al at this site and so there should be no peak on electropherogram for the lost allele. Contamination by normal benign cells or other cancer clones during the extraction process will result in some normal DNA, without Al, also being amplified during PCR. While there is less normal than cancer DNA to be amplified there may be sufficient quantities to produce a peak on the electropherogram. The cut-off for what defines Al therefore reflects the purity of the cancer DNA. In this study, the majority of samples were not laser microdissected as large quantities of DNA were needed. The master wholemount slides were marked with areas containing at least 70% malignant cells, it is inevitable that some contamination with normal DNA occurred. The cut-off for this project was put at an XAi score of 0.7 to reflect the predicted degree of contamination. This correlates with an allele loss of approximately 40% (Figure 3.5).

168 Figure 3.5 Allelic imbalance is present where the amplitude of the smaller cancer peak divided by the larger cancer peak over the amplitude of the smaller control peak divided by the larger control peak is 0.7 or less.

Cancer Number of base pairs from 5/3’ terminus 1107.51190.40! 11004] [335} Amplitude of peak 103491 I2059!

Control Area under the curve

159.3Q| 167* [806| [797] I r a n i I7975I Stutter peaks Allelic peaks

B/A Al = = < 0.7 D/C

335/1004 = o 3 Al = 797/806

Two researchers both blind to the patient outcome examined each electropherogram independently. The results from both were collated, where discrepancies occurred the electropherograms were examined together and agreed. Where Al occurred in less than

10% of informative cases this was considered background losses and therefore not significant (Dong, Boyd, & Frierson, Jr. 2001; Macoska et al. 1995; Saric et al. 1999).

169 CHAPTER 4

Results

Aim

This chapter will present the results of the allelic imbalance experiments ancf discuss the

findings in relation to the literature. r x 4.1 Allelic Imbalance

Maternal and paternal alleles have naturally occurring sites of two or more nucleotide

repeats, usually outside the coding regions of genes, called microsatellite

polymorphisms. Over 18,000 microsatellite polymorphisms have been identified and are

listed on the Genome database (www.gdb.org). The polymorphic regions of the

maternal and paternal alleles are usually of different lengths, rendering the individual

heterozygous at that locus. Deletion or amplification, which occur during the initiation

and progression of cancer, may be detected by the loss or gain of one of these

polymorphic regions; this is allelic imbalance (Al). Deletions are far more common than

amplification in early prostate cancer accounting for most cases of Al in this study. The

deleted region may contain a tumour suppressor gene and deletion in one allele is

usually preceded by a point mutation in the other allele so that the deletion represents

the second hit of the Knudson’s hypothesis.

If the maternal and paternal microsatellite regions at a particular locus in normal tissue

are of the same length, Al is not detectable and this locus is termed uninformative.

Hence, a normal tissue control is required to evaluate if an individual is heterozygous,

therefore informative at the loci of interest.

170 Allelic imbalance has prognostic and therapeutic implications and has resulted in its use

clinically for certain tumours (Barratt et al. 2002; Bell et al. 1993; Jen et al. 1994). In

patients with Dukes B colorectal carcinoma, allelic imbalance in the region of the DCC

gene (D18S61, 8q22.3) indicated a prognosis similar to a Dukes C lesion (40% 5-year

survival) (Barratt et al. 2002). Conversely, heterozygosity at this locus conferred a

better prognosis (70% 5-year survival). Furthermore, benefit from the chemotherapeutic

agent fluorouracil was significantly greater in patients retaining heterozygosity (Barratt

et al. 2002).

Allelic imbalance in prostate cancer has been consistently reported at several

chromosomal arms, including 7q, 8p, lOq, 12p, 13q, 16q, 18q and 21 q (Gray et al. 1995;

Harkonen et al. 2005; Hugel & Wemert 1999; Kibel et al. 2000; Latil et al. 1995; Latil

et al. 1997; Latil et al. 1999; Leube et al. 2002; Oba et al. 2001; Prasad et al. 1998; Yin

et al. 2001). Several studies have reported an association of Al at certain loci with

histopathological features of the radical prostatectomy specimen, including advanced

stage (Gray et al. 1995; Hugel & Wemert 1999; Kibel et al. 2000; Leube et al. 2002),

higher Gleason score (Hugel & Wemert 1999; Leube et al. 2002) and perineural

invasion (Fromont et al. 2003). However, the frequency of Al differs widely among

studies as tissue from a small number of patients with a range of clinical stages has been

used. Most studies have employed tissue from advanced metastatic tumours so the

association of Al with prognosis in localised prostate cancer remains controversial.

/W hat this study aims to do is to identify microsatellite markers that can be used to

predict biochemical outcome in men undergoing radical prostatectomy.

171 4.2 Results

4.2(a) Interpreting electropherograms

Thirty fluorescently labelled microsatellite markers for prostate cancer associated loci on chromosome arms lq, 2q, 5q, 6q, 7q, 8p, lOp, lOq, lip , 1 lq, 12p, 13q, 15q, 16q,

18q, 19q and 21 q were used. The DNA from the prostate cancer and the lymph node control of the relapse group was amplified in parallel with the DNA from the non­ relapse group who were matched for grade, stage and pre-operative PSA. PCR products were run on a polymer gel and the fluorescent electropherogram assessed by two workers independently. In 79% of cases both workers scored the electropherograms the same (heterozygote, allelic imbalance, homozygote, microsatellite instability or unreadable). Where there was a discrepancy the electropherograms were examined together and an overall score given.

Some examples are shown in Figure 4.1 (See also Figure 3.5 for explanation of figure). Each marker produced a characteristic electropherogram due to ‘stutter peaks’, which precede the allelic peaks. In Figures 4.1a, b and c there is a smaller peak one or two base pairs after the allelic peak, this is as a result of adenine adhering to the PCR product. Adenine does not bind to the sense strand if the complementary nucleotide on the antisense strand is cytosine, guanine or adenine residue. Consequently a peak after the allelic peak is not always seen (Figure 4.Id).

172 Figure 4.1. Shows characteristic cancer and control electropherograms for 5 markers, demonstrating stutter peaks preceding the allelic peak in a., b. and c., heterozygotes in a. and c., a homozygote (uninformative) in b., allelic imbalance in d, and microsatellite instability in e.

cancer cancer 1227 871 1238.021

16475 | 1405S|

control control

184.031 [227.781

Figure 4.1a. D7S523 Figure 4.1b. D5S656 Patient 12 - Heterozygote Patient 12 - Homozygote

cancer cancer

127 1 .6 7 1 |2208| 146551 151501 190291 [626941

control control

1256.271 1271.631 [876 ] 172071 136511 [74721

Figure 4.1c. D16S413 Figure 4.1d.D15S123 Patient 33 - Heterozygote Patient 33 - Allelic imbalance

cancer

Figure 4.1e. D6S314 Patient 2 - Microsatellite instability Extra peak in cancer Allelic peaks

control

173 The appearance of the electropherogram from a homozygote individual or from a heterozygous individual whose maternal and paternal alleles were of significantly different length was used to aid analysis of other electropherograms where the stutter and allelic peaks overlapped (Figure 4.2).

cancer cancer 1158.18| | 1 5 6 .1 4 11160|| 1321] 4 1 6 6 110612) 3 8 1 7 2 1160.121 12472

control control 1 5 6 .1 7 158.11 3 0 6 4 2 8 2 1 0

160.06 112944|

Figure 4.2a. D13S263 Figure 4.2b. D13S263 Patient 39 - Homozygote Patient 32 - Allelic imbalance

Figure 4.2a and b. Electropherograms from patients 39 and 32. Patient 39 is homozygous at the D13S263 locus. Two stutter peaks precede the alleles. This appearance can be used to confirm that patient 32 is heterozygous at this location as there are four peaks, and the cancer sample shows allelic imbalance.

The heterozygosity rate of each marker was similar in this series to the reported ‘hef rate (Table 3.1) ranging between 52-94%. At all loci a mean of 36 (58%) of 62 patient cancer samples were informative i.e. were heterozygotes or had allelic imbalance.

174 For all patients the cumulative allelic imbalance rate for informative cases at all 30 markers was 23%. The cumulative allelic imbalance frequency was greater in the relapse group compared with the non-relapse group, 24% of informative cases in the relapse group and 21% of informative cases in the non-relapse group (Chi square, p>0.05). The results are summarised in Tables 4.1, 4.2, 4.3 and 4.4.

175 4.2(b) Allelic imbalance and PSA outcome

One marker D10S211 at 1 Op 12.1 demonstrated significantly greater allelic imbalance in the relapse patients than in the non-relapse patients (35% vrs 5%, respectively, Fisher's exact test p=0.03). Several loci showed a marked increase in allelic imbalance in the relapse compared with the non-relapse groups. These included D2S222 (29 vrs 8%),

D6S314 (25 vrs 11%), D8S136 (50 vrs 29%), NEFL (71 vrs 55%), D12S1697 (31% vrs

17%), D13S269 (19 vrs 4%), D15S1232 (38 vrs 18%) and D18S541 (25% vrs 0%) in the relapse and non-relapse groups, respectively. Some loci demonstrated an increased rate of allelic imbalance in the non-relapse group. These were: D1S158 (14 vrs 35%),

D1S422 (17 vrs 27%), D5S500 (13 vrs 23%), D11S2000 (6 vrs 31%), D12S89 (14 vrs

29%) and D16S413 (13 vrs 24%) in the relapse and non-relapse groups, respectively.

These differences were not statistically significant.

All the primers gave satisfactory electropherograms for some DNA samples, and all the

DNA samples produced satisfactory electropherograms for some primers. Allelic imbalance was seen in 57 (92%) of the 62 patients. Five patients who did not show allelic imbalance included 3 relapse patients (24, 25, 64) and 2 non-relapse individuals

(30, 61) (Table 4.5).

176 Table 4.1. OVERALL RESULTS - ALL PATIENTS

177 Table 4.2. OVERALL RESULTS - RELAPSE GROUP Table 4.3. OVERALL RESULTS - NON-RELAPSE GROUP

/o/yyyy< /o /o /o yyyy< / / v / o /

/ / V s / £ > / < & / & / / & / & / < & / (fr / *& / r& / / <& / C V > / & / O? / K? / / / & / K* / v / »*s / nV / *y / & / *v>/ Nr / *? / *?/'&/'&/

RELAPSE 14% 17% 35% 5% 29% 13%25% 40% 25% 29% 6% 50% 50% 71% 19% NON-RELAPSE 35% 27% 29% 0% 8 % 23% 23% 47% 11% 25% 6% 29% 55% 25%45% 'Significance, p= 0.12 0.39 0.45 0.49 0.070.30 0.63 0.48 0.41 0.57 0 73 0.53 0.13 0.26 0.63

RELAPSE 6% 35% 5% 6% 14% 31% 19% 25% 28% 38% 13% 25% 0% 21% NON-RELAPSE 10% 5% 0% 31% 31% 29% 17% 4% 33% 28% 18% 24% 0% 5% 19% 'Significance, p= 0.56 0.03 0.56 0.55 0.09 0.29 028 0.5 0.31 0.19 0 25 0.160.54 046

* = Fisher’s Exact test

180 Table 4.5 Status of 30 microsatellite loci in 31 matched pairs of radical prostatectomy patients w ith pT3N0 stage disease.

Non-relapse group

I uni nterpretable heterozygous |allelic Imbalance microsatellite instabilityhomozygous

181 Microsatellite instability was evident in 25 patients, 13 in the relapse group and 12 in the non-relapse group. There was no significant difference in the rate of microsatellite

instability between the relapse and non-relapse groups at any of the 30 markers (Table

4.6).

182 Table 5. INCIDENCE OF MICROSATELLITE INSTABILITY IN THE RELAPSE AND NON-RELPASE GROUPS

RELAPSE NON-RELAPSE Significance, p= 0.51 0.50 0.26 0.75 0.49 0.52 0.44 0.52 0.73 0.48 0.47

RELAPSE NON-RELAPSE Significance, p= 0.51 0.48 0.47 0.57 0.29 0.59 0.52 0.47 0.48 0.26 0.71 0.55 0.67

* = Fisher’s Exact test

183 4.3(a) Discussion

Many sites of allelic imbalance have been identified in prostate cancer though their prognostic significance remains unresolved. The aim of this study was to identify microsatellite markers, which could be used to predict biochemical relapse after radical prostatectomy. Allelic imbalance was seen more frequently in patients who had developed biochemical recurrence within two years of undergoing radical prostatectomy

(24%) compared with patients who were PSA free for at least 3 years after surgery

(21%), this was not significantly different. \ , N i ■ - IL d- f

Allelic loss occurred at one locus on chromosome 10, 1 Op 12.1 (D10S211), significantly more in the relapse group (6 (35%) of 17 informative cases) than the non-relapse cohort

(1 (5%) of 19 informative cases, p= 0.03). Allelic imbalance at 10q25.3 (D10S587) did not exceed 10% in either group and therefore represents background losses (Dong,

Boyd, & Frierson, Jr. 2001; Macoska et al. 1995; Saric et al. 1999). Losses at 10q25.1

(D10S1246), though greater than background did not reach significance between the groups.

Allelic imbalance on chromosome 10 is a common event in several solid tumours, including renal cell cancer (Morita et al. 1991), glioblastoma (Karlbom et al. 1993), meningioma (Rempel et al. 1993), malignant melanoma (Herbst et al. 1994) and endometrial cancer (Peiffer et al. 1995). Loss of 10q22-24, the site of at least two putative tumour suppressor genes - PTEN and MXI1, is the most commonly reported site of deletion on chromosome 10, with rates of 20-50% of informative cases described

(Bergerheim et al. 1991; Macoska et al. 1993; Phillips et al. 1994; Sakr et al. 1994).

Two other sites on chromosome 10, lOp and 10q26, are also deleted in association with or independent of 10q22-24 losses (Ittmann 1996; Karlbom et al. 1993; Komiya et al.

184 1996; Trybus et al. 1996). Trybus et al. examined allelic imbalance in chromosome 10 using 12 markers in 35 radical prostatectomy specimens (Trybus et al. 1996). Overall, 6

(17%) tumours demonstrated loss of lOp loci only, 5 (14%) showed losses of lOq loci only and 14 (40%) were characterised by loses of both lOp and lOq loci. Six (18%) showed loss at D10S211, which is similar to the current series where 7 (19%) of 36 informative demonstrated allelic imbalance at this locus. Trybus et al. found no significant association with 1 Op loses and tumour grade and stage, though patients with

localised tumours had higher frequencies of lOp loss compared to men with extraprostatic tumours and those involving the seminal vesicles or lymph nodes (Trybus et al. 1996).

In a study of 64 clinically localised prostate cancers allelic loss was found in 9 cases in one or more of 12 microsatellite regions examined on chromosome 10 (Ittmann 1996).

Of these 9 cases 6 were informative for D10S211 and 2 (33%) had allelic imbalance at this locus. In this study the incidence of allelic loss was higher than than our series

because the subset of nine individuals who had confirmed Al in chromosome 10 were

selected for analysis.

In a study of 48 Japanese men undergoing either radical prostatectomy or autopsy with

stage B-D prostate cancer, losses on chromosome 10 were reported significantly more

frequently in advanced compared to localised cases (Komiya et al. 1996). In these

studies, unlike our series, tumours from a broad range of cancer stages (T2N0M0-

T4N1M1) were utilised. These studies concluded deletions of chromosome 10 are more

common with higher stage disease. The incidence of Al and biochemical outcome in patients undergoing radical prostatectomy was not correlated in these studies.

185 Several loci demonstrated marked, though not statistically significant differences,

between the relapse and non-relapse groups. Differences in the region of 2q23.1

(D2S222) approached significance (p=0.07), 6 (29%) of 21 relapse patients and 2 (8%) of 25 non-relapse patients region showed allelic imbalance. This region has not been

extensively studied in prostate cancer. Saric et al. reported 1 (6%) of 17 men undergoing

radical prostatectomy for clinically localised disease and 4 (23%) of 23 metastases

samples from autopsy specimens had Al at this locus. Other studies confirm deletion of

2q is associated with metastatic disease in non-small-cell lung carcinoma (Lu et al.

1996). The current study suggests an association with poor outcome after surgery

though further studies are required to confirm these findings.

In the current study 3 (25%) of 12 informative relapse patients and 1 (11%) of 9

informative non-relapse patients demonstrated allelic imbalance at 6q 16.3-27 (marker

D6S314). Cooney et al. have shown that this region is commonly deleted in radical

prostatectomy specimens (Cooney et al. 1996a). However, no association was found

between patients who developed PSA relapse and 6q losses, though only 10 patients

relapsed over a short follow-up of 8 months in this series. In a study using using CGH

analysis, 22% of primary prostate cancers and 44% of recurrent tumours had 6q losses

(Visakorpi et al. 1995b). The region 6 q l6.3-21 has been implicated in hereditary mixed

polyposis syndrome (HMPS), an autosomal dominant disorder predisposing to colonic

polyps and carcinomas (Thomas et al. 1996). A tumour suppressor gene has not yet

been isolated, though several may exist as patients with HMPS do not have increased

risk of prostate cancer development.

Allelic imbalance occurred in 5 (19%) of 26 relapse patients and 1 (4%) of 25 non­

relapse patients in the region 13q21 (marker D13S269). Deletions of 13q are frequently

186 reported in prostate cancer and the region 13q21 is believed to harbour an as yet unidentified TSG important in prostate cancer progression. Studies examining this region have reported deletion in this region is associated with poorly differentiated or metastatic disease (Dong et al. 2000). The results from the current study support these findings; losses at 13q21 appear to be associated with increased risk of biochemical relapse after radical prostatectomy.

The chromosomal region 12pl2-13 contains the TSG p27/kipl and deletion of this region is associated with metastatic potential in prostate cancer (Kibel et al. 2000).

Kibel et al. identified Al in 23% of 60 primary tumours from radical prostatectomy specimens compared with 47% of 19 metastatic lesions from autopsy specimens using polymorphic markers spanning this region, including the marker D12S1697. Of eleven patients undergoing autopsy with multiple metastases, 7 demonstrated Al in this region and the pattern was identical in all the metastases examined from each individual patient as well as in the primary tumour. This study did not examine how many of the patients with 12pl 2-13 losses in the primary tumours at radical prostatectomy went on to develop biochemical relapse or metastases. However, this data suggests genetic inactivation of this region occurs in primary tumours prior to distant spread and may be a critical step in prostate cancer metastasis. In the current study 31% of the relapse patients compared with 17% of the non-relapse men had allelic imbalance at 12pl2.3

(marker D12S1697). This difference was not statistically significant, a larger study may demonstrate allelic imbalance in this region is a predictor of biochemical failure after radical prostatectomy.

Studies looking at allelic imbalance on chromosome 15 in lung, colorectal, renal and ovarian carcinomas have found a low incidence of allelic imbalance in informative

187 cases. However, in breast cancer region 15ql4 was reported to harbour a TSG which plays an important role in the development of metastasis (Wick et al. 1996). There are

no reported series looking at chromosome 15q in a significant number of patients with prostate cancer. The current series demonstrated 6 (38%) of 16 informative relapse men and 3 (18%) of 17 informative non-relapse patients had allelic imbalance at 15q 13.1

(D15S1232). Though this is not statistically significant a TSG important in prostate

cancer progression may exist in this region.

Several sites of allelic loss are consistently associated with prostate cancer metastases. It

might be feasible to identify patients harbouring metastases, before they are clinically

detectable, by examining known deletion patterns associated with metastatic spread.

Three sites frequently deleted in metastatic prostate cancer were evaluated. The tumour

suppressor gene KAI1 gene is located at 1 lpl 1, close to marker D11S902. Deletion of

this region is associated with metastatic prostate cancer (Kawana et al. 1997; Saric et al.

1999) and the KAI1 gene has been shown in metastatic rat models to reduce metastatic

potential (Behrens et al. 1989). Only 1 (3%) of 36 informative cases had allelic

imbalance at this locus, though the patient (79) who had allelic imbalance at this locus

relapsed soon after surgery (by 12 months).

Locus 18q 21.1 (marker D18S541) is also associated with metastatic potential in

prostate cancer (Saric et al. 1999; Ueda et al. 1997). Similar to 11 pi 5, the overall

incidence of allelic imbalance at this locus was low, 2 (10%) of 19 cases, however both

patients with allelic imbalance at this locus developed biochemical relapse, one at 13

months (patient 2) and one at 3 months (patient 43). Patient 43 had good pathological

features after surgery - Gleason 7 disease, specimen confined, surgical margin negative

disease and a preoperative PSA of 5.2ng/ml, and yet developed a rising PSA rapidly,

188 following a PSA nadir, after surgery. Using conventional criteria patient 43 would be expected to have a PSA free survival of over 80% at 5 years (Catalona, Ramos, &

Carvalhal 1999; Pound et al. 1997). The microsatellite marker D18S541 successfully predicted poor biochemical outcome in this individual demonstrating the potentially superior predictive role of molecular markers over conventional phenotypic changes, though this needs to be shown in larger numbers to confirm this does not represent background losses.

Losses at 16q24 are associated with metastatic potential in prostate cancer (Elo et al.

1999; Harkonen et al. 2005; Saric et al. 1999; Suzuki et al. 1996). Al was found in 17% of 18 primary cancers from radical prostatectomy specimens (Saric et al. 1999), which is consistent with an overall allelic imbalance rate of 18% at marker D16S413 in the current series. In metastatic samples Saric et al. noted this rate more than doubled, the rates between the relapse and non-relapse groups in the current series were similar.

Likewise, Al at D16S413 was identified in 27% of 22 primary tumours, in 45% of 22 recurrent prostate cancers and 44% of 9 metastatic lesions (Harkonen et al. 2005).

Furthermore, the incidence of Al at 16q24.3 (marker D16S520) was greater in the recurrent than in the metastatic samples indicating an association with recurrent tumour after androgen withdrawl, rather than metastatic potential.

Losses at 8p play a key role in initiation and progression of prostate cancer and the three markers on the short arm of chromosome 8 had some of the highest rate of allelic imbalance of all 30 markers in this study. At least, two distinct regions of loss have been reported (MacGrogan et al. 1994; Trapman et al. 1994). Loss of 8 p l2-21 is an early event in prostate carcinogenesis, occurring in HGPIN and early invasive disease, whereas loss of 8p22 is a late event, found more commonly in advanced cancers (Abate-

Shen & Shen 2000). Marker D8S136 on 8pl2-22 demonstrated allelic imbalance in

189 39% of informative patients, marker NEFL at 8p21.3 62 % and marker D8S1991 at

8p22 47%. Vocke et a l found 72% Al at D8S136 and 64% at NEFL in the

microdissected prostate cancer of 97 patients with T2a-T3cNl disease. Microdissection

increases the yield of tumour cells compared with normal stromal cells. By reducing the

contamination of the cancer sample, allelic imbalance, when present, will be more

apparent because the amplitude of the deleted allele is less. This together with the

higher number of more advanced cases explains why the Al rates were higher compared

with the current series. In the current series three markers demonstrated a higher rate of

allelic imbalance in the relapse patients than the non-relapse individuals. Macoska et al.

attempted to examine the relationship between Al and PSA outcome after radical

prostatectomy. They looked at 8p deletions in 78 of 135 men. No significant differences

were seen in fractional allelic loss or total 8p loss between those patients with PSA

recurrence (Macoska et al. 1995). Together with the current study there is currently no

evidence of a link between 8p loses and biochemical outcome after radical

prostatectomy (Macoska et al. 1995; Sato et al. 1999). \

4.4 Study Limitations

There are several reasons why with a large cohort of men and a large number of

microsatellite markers more loci were predictive of biochemical outcome. Firstly the

patient factors outlined in Chapter 2 meant that an insufficient number of men could be

matched. To establish the minimum sample size a realistic expectation of what

difference in allelic imbalance between case and control was set at 40%. If either of the \ matched relapse or non-relapse individuals were uninformative or uninterpretable both

would be excluded for matched comparsion. The heterozygosity rate for markers

selected was set at >0.7. The total number of paired matched cases to give 30

informative pairs, assuming a het rate on 0.7, is 62. So a radical prostatectomy cohort of approximately twice the size of ours (ab^ut 2000 cases) would be needed to achieve sufficient patient numbers using this method.

Apart from the difficulty of finding sufficient numbers, the other major limitation of this study was the relatively poor quality of the archival DNA recovered from the samples.

This may be in part due to the relatively slow fixation achieved in the large radical prostatectomy specimens during immersion in formalin. However, all the primers gave satisfactory electropherograms for some DNA samples, and all the DNA samples produced satisfactory electropherograms for some primers. Perhaps by optimising the

PCR conditions for each set of primers and each DNA sample, it would have been possible to increase the success rate, but given the exceptionally large scale of the study and the limited amount of DNA, this approach was not feasible. Recently, improving technology has meant more of the techniques used in this study can now be done using automated robots and large-scale PCR and genotyping machines. Tailoring the PCR conditions for each marker and increasing the number of markers and patients becomes more feasible using this more up-to-date equipment.

Tumours were selected based on their size and Gleason score. Samples were then scraped in the majority of cases using a scalpel blade. Although this technique has been described in the literature it does not provide a pure clone of cancer cells. Prostate cancer is a genetically heterogeneous disease and so scraping large areas of tumour may incorporate several genetically different tumours. Al in one area of tumour may be different from Al in another tumour area. DNA containing several different tumours may not reveal prognostic Al found only in one tumour. Furthermore benign stroma cells will be incorporated (Leube et al. 2002), particularly in the lower grade cancers where the stromaP.epithelial ratio is higher. The presence of normal DNA contaminating

191 cancer DNA diminishes the chance of detecting Al. This problem can be circumvented to a certain extent by laser capture microdissection (LCM) (Emmert-Buck et al. 1995;

Leube et al. 2002; Narla et al. 2001). However with LCM a small area of tumour is extracted for analysis and this may not represent the index or clinically important tumour. Secondly, the DNA yield from LCM is substantially smaller and large quantities of DNA were needed to perform genotyping with 30 markers in the current study. A i i 0 v 1v"° - £ o ) '^ - Vv<' This study involved selecting markers reported by others to show loses in prostate cancer. It is not clear from these studies whether these loses are critical in the pathogenesis of prostate cancer and in particular the transition between extracapsular extension and the developent of metastases. Loses may occur in tumour progression without any change in the cancer behaviour. Similarly prostate cancer progression may occur via many different genetic pathways. Identifying a patient as having a ‘good prognosis’ by demonstrating heterozygosity at a particular locus does not exclude allelic imbalance at another locus, which may represent worse prognosis. Results obtained in this study need to be confirmed in a separate patient cohort.

4.5 Conclusions

Allelic imbalance studies have been used to identify regions of loss, containing critical genes, important in the multi-step process of cancer development and progression.

Identification of specific losses may be used to predict prognosis and treatment response in a number of malignancies. This study demonstrated specific losses on chromosome

1 Op 12.1 were associated with biochemical progression after radical prostatectomy.

Several other loci approached significance and maybe statistically significant in a larger sample. The small number of patients with allelic imbalance at llpll and 18q21.1 all

192 developed biochemical relapse, despite in some case favourable pathological prognostic features. For the minority of individuals with these deletions adjuvant therapy after radical prostatectomy may be of benefit.

This study also illustrates some of the difficulties encountered relating genetic changes in formalin-fixed tissue to clinical outcome. Confirming the Al findings identified in the radical prostatectomy specimen are also present in the diagnostic needle biopsy may, in the future, enable treatment to be tailored to the individual patient.

193 CHAPTER 5

The p21WAF1/clpI gene is inactivated in metastatic prostatic cancer cell lines by promoter methylation.

The aim of this chapter was to describe the role of p21 WAF1 CIPI in cell cycle progression and its potential role as a tumour suppressor gene. The expression of p2JWAFI CIPI in metastatic prostate cancer cell lines was examined and a mechanism of inactivation in prostate cancer is proposed (Bott et al. 2005a).

5.1 Introduction

The Gl/S transition checkpoint in the cell cycle is an important point of control as DNA damage can be repaired before replication occurs (Sherr 1996). Progression through the

Gl/S checkpoint in the cell cycle is controlled by cyclin dependent kinases, whose activity in turn is controlled by cyclin dependent kinase inhibitors, including p21liAFI C,PI. The role of p 2 ]WAFI/CIPI ancj its association with carcinogenesis has been described in section In brief, normally as a cyclin dependent kinase inhibitor

p 2i ^ A F i c i p i a meciiator of p53 induced growth arrest as a modulator of apoptosis.

Inactivation of p21nAF1 CIP1 during prostate cancer development may allow unchecked passage through the cell cycle and p 2 iWAF,/CIPI has been related to several solid malignancies including prostate cancer.

Hypermethylation is the most frequent somatic epigenetic alteration in prostate cancer and may result in changes in gene expression (see section 1.11(d)). It involves the addition of a methyl group to the carbon-5 position of cytosine residues (Figure 5.1),

194 and occurs almost exclusively at cytosine residues that are immediately followed by guanine (CpG dinucleotides). Isolated CpG dinucleotides are relatively rare and are nearly always methylated. In contrast small stretches of DNA rich in CpG dinucleotides, the CpG islands, are usually free of methylation in normal tissue. These

CpG islands are frequently found in the promoter region of human genes and abnormal methylation within the islands has been associated with transcriptional inactivation of the corresponding gene. Alteration in DNA methylation in the promoter region of tumour suppressor genes may be pivotal to the development of cancer (Strathdee &

Brown 2002).

NH

Cytosine

SAM-CH — DNMT SAM

5-Methylcytosine

Figure 5.1. Mechanism of DNA methylation

DNMT = DNA methyltransferase 1,3a or 3b SAM = S-adenosylmethionine

The aim of this study was to examine the expression of p 2 iWAFI/CIPl in metastatic prostate cancer cell lines and to assess the methylation status of the promoter.

195 5.2 Materials and Methods

5.2(a) Cell culture and treatment

Human metastatic prostate cancer cell lines PC3, LNCaP, DU 145 and 1542 NP were obtained from the originators, RD a rhabdomyosarcoma cell line and 1542NP normal prostate epithelial cells were obtained from ATCC (American Type Culture Collection,

VA, USA). LNCaP cells have a slower replication rate than the other cell lines.

500,000 PC3, DU 145, 1542NP, RD cells and 106 cell LNCaP cell were plated in flasks containing standard growth medium, namely RPMI-1640 (Roswell Park Memorial

Institute-1640, Gibco BRL), 10% Foetal Calf Serum (FCS) and 1% of L-Glutamine (G) and cultured in a humidified incubator in 5% carbon dioxide at 37°C. After 16 hours,

500nM of the demethylating agent 5-Aza-2 deoxycytidine (5-Aza-CdR)(Sigma

Chemical Co: A 3656 C 8H 12N4O4) in 15pl of phosphate buffer solution (PBS) was added to the test flasks and 15pl PBS was added to the control flasks. 5AzaCdR sequesters the enzyme DNA methyltransferase and thus prevents DNA methylation during DNA replication. After treatment for 24 hours the test and control media was replaced with standard growth medium. The media was changed every 72 hours.

After 5 days the cells were harvested. The medium was removed and 2mls of trypsin was added and distributed over the monolayer of cells and the flask returned to the

incubator for 5 minutes. Trypsin is a proteolytic enzyme that releases the cells adhering to the flask. Once the cells were detached 8 mls of RPMI and FCS was added to neutralise the enzyme. The suspension was mixed and a sample taken for counting. The remaining cell suspension was alliquoted into a 20ml Universal tube and centrifuged at

1250rpm for 10 minutes at 4°C. The cell pellet was resuspended and washed in 5mls

PBS and centrifuged again. This wash step was repeated. After the second wash step the

196 PBS was aspirated, the cell pellet snap frozen in liquid nitrogen and stored at -80°C.

Cell culture was performed in triplicate.

5.2(b) Cell counting

200pl of the well-mixed single cell suspension from each T80 flask was alliquoted into an Eppendorf ™ tube containing 0.2 ml trypan blue (0.1% solution) and mixed. Trypan blue is a viability stain and is only taken up by dead cells, these can be excluded from the count. A few microlitres of cell suspension mix was applied to the haemocytometer.

Cells were counted using the lOx microscope objective in at least four squares in both of the two chambers on the haemocytometer so that a minimum of 100 cells was counted. An average number of cells was calculated by summing the number of cells in all squares counted and dividing by the number of squares examined. The volume of each square was lO^mls (1mm2 x 0.1mm in depth). The cells were diluted 1:1 with trypan blue, so the cell count was multiplied by 2. This is summarised below:

The cell number/ml of suspension =mean number of cells per square x 2xl04

197 5.2(c) p2J WAFI/c,pl gene expression

5.2(c) 1. RNA extraction

RNA from cell lines was extracted using RNeasy® mini kit (QIAGEN®).

• 600 (xl of buffer RLT was added to the cell pellet and vortexed, buffer RLT

contained p-mercaptoethanol and causes the cell wall disruption and release

RNA.

• The lysate was pipetted onto a QIAshredder spin column, placed in a 2ml

collecting tube and centrifuged for 2mins at 13,000rpm. This homogenisation

step shears the high molecular weight genomic DNA and other high molecular

weight cellular components.

• 600pl of 70% ethanol was added to the homogenised lysate to expose the

phosphate backbone for binding.

• This solution was placed in a RNeasy mini column and centrifuged for 15secs at

13,000rpm, the flow-through was discarded. The spin columns contain a

positively charged silicone membrane that binds to the negatively charged

phosphate backbone of the RNA.

• 700pl of washing buffer RW1 was added to the column and the tube was

centrifuged at 13,000rpm for 15 secs. The wash-through was discarded.

• 500pl washing buffer RPE was added to the column and the tube was

centrifuged at 13,000rpm for 15 secs. The wash-through was discarded.

• 500pl washing buffer RPE was added to the column and the tube was

centrifuged at 13,000rpm for 120 secs to wash and dry the silica-membrane.

• To elute the RNA the column was placed in a 2ml collecting tube and 30pl of

RNase-free water was added to the membrane and centrifuged at 13,000rpm for

60 secs.

• RNA was snap frozen in liquid nitrogen and stored at -70°C.

198 5.2(c)2. cDNA

Single stranded RNA was made into copy DNA (cDNA), as follows. Into a 200pl tube, lpg of RNA made up to 11 pi with 0.1% diethylpyrocarbonate (DEPC) water was alliquoted. The DEPC destroys contaminating RNAses. lpl of oligo(dT) was added, the tube was vortexed and placed in a water bath at 70°C for 10 mins. Oligo(dT) is a primer consisting solely of deoxythymine residues. At 70°C the RNA looses its secondary structure and oligo(dT) anneals to the polyadenylated 3' tail found on most mRNAs.

The tubes were removed and put on ice. To each tube 4 pi buffer, 2 pi deoxythymine triphosphate (DTT) and lpl dNTP (the single nucleotides A,T,G,C) was added. The tubes were placed in a PCR machine at 42°C for 50 minutes. After 2minutes, lpl of

Superscript II reverse transcriptase (RT) was added to all tubes except the ‘no RT’ control. RT hybridises the single nucleotides (dNTP) to their complimentary base sequentially from the oligodT primer, forming cDNA. The tubes were heated to 70°C for 1 Omins to inactivate the RT and placed on ice.

5.2(c)3. Primer design

Primers are complementary to and hybridise with opposite DNA strands either side of the region of interest. The following were applied in primer design.

1. Primers were usually 20-30 base pairs long to allow a reasonably high annealing

temperature, longer than 30 does not improve primer specificity.

2. Primer annealing was helped by having Guanine or Cytosine at the 3’ end, as the

three H+ bond formed by G to C are stronger than the two formed between A

and T.

199 3. Primers were designed so that a total of 50 or more hydrogen bonds form when

the primers anneal and were made up of approximately equal numbers of purines

(A and G) and pyrimidines (T and C).

4. Repetitive sequences and stretches of more than 3 of the same base were

avoided to ensure specificity. Complimentary sequences either within a primer

or between a primer pair were minimised to prevent the formation of primer

dimmers - an artifact where the primer itself acts as a template resulting in a

short length product.

5. The distance between the sense and the antisense primer i.e. the length of

desired product did not exceed 10 kilobases, the polymerase enzyme was less

effective at amplifying the cDNA after 3kb.

6. When performing an RT-PCR there was a risk of cDNA contamination by

genomic DNA. This was overcome by designing primers in adjacent exons

where there was a significant length of intron in between. The genomic DNA

segment including the intron would be too large to amplify. However, the intron

is removed in the formation of mRNA so the cDNA can be amplified.

Intragenic primers for p21 and GAPDH were designed and are summarised in

Table 5.1.

200 Table 5.1 Intragenic primers for p21 WAF,/CIP1 an(j GAPDH

Primer Sequence Annealing Product

temperature size (bp)

p21 sense tgaccctgaagtgagcacagcc 60°C 834 p21 anti sense gccgagagaaaacagtccaggc

aaccaccactttgtcaagctc GAPDH sense 60°C 210 GAPDH gtccaggggtcttactccttg anti sense

5.2(c)4. Reverse Transcription - Polymerase Chain Reaction (RT-PCR)

The reaction mix used in these experiments is shown in Table 5.2.

Table 5.2 PCR master mix for lx 50pl reaction

5 pi 1 Ox buffer

5 pi dNTP (2mM)

3 pi MgCl2 (25mM)

2 pi sense primer (lOpmol/pl)

2 pi antisense primer (lOpmol/pl)

29.5 pi H2O (variable)

0.5 pi Taq enzyme

cDNA was amplified using intragenic primers at p21WAF1/CIP\ spanning a region -990

to -67 relative to the transcription start site and to GAPDH. Reactions were performed

201 in a Perkin Elmer-Gene Amp 2400 thermal cycler (Perkin Elmer Ltd., UK), at 94°C for

150 s, then 34 cycles at 94°C for 45 s, 60°C for 45 s and 72°C for 45 s, followed by an extension step at 72°C for 5 mins. The PCR products were electrophoresed through a

1.5% agarose gel

5.2(c)5. Agarose gel electrophoresis

Agarose gel electrophoresis was used to separate DNA fragments according to their size and charge. 1.5g of agarose was added to lOOmls of lxTAE buffer (Tris acetate electrophoresis buffer) and heated until the agarose had completely dissolved. 5pi of ethidium bromide (at 5mg/ml) was added per lOOmls of gel solution and mixed.

Ethidium bromide incorporates into DNA strands and is photosensitive. The agarose was poured into a gel tray containing a gel comb and allowed to set.

A stock lx solution of gel loading buffer was used and 6pl added and pipette mixed with 25pi of PCR product. 20pl of PCR product/gel loading buffer mix was loaded along with 4pl size standard 1Kb ladder (mixed 3:2 by volume with gel loading buffer), to ensure correct product size.

A potential difference of 135V was placed across the gel tank. Migration across the gel was observed using the movement of the blue gel-loading buffer until the products had separated sufficiently when an image under ultraviolet light was recorded.

202 5.2(d) Promoter sequencing

5.2(d)l. DNA extraction

The frozen cell pellet was thawed on ice and re-suspended in PBS. DNA extraction was performed using the DNeasy Tissue kit ™ and the extraction process for cultured cells is similar to that of paraffin embedded sections described in Chapter 3.

The DNA content of the eluate was measured by diluting 2pl in 498pl water and using optical densometry (OD). Final DNA concentrations were 259-1500ng/ml.

The sample DNA was stored at -25°C.

5.2(d)2. Sodium bisulphite modification

Sodium bisulphite deaminates cytosine, removing the NH 2 group from carbon-4

(Figure 5.2) converting cytosine to uracil, unless the cytosine is methylated in the carbon-5 position. Metastatic prostate cancer cell line, RD (positive control), and NP

(negative control) DNA was treated with sodium bisulphite.

Figure 5.2. Sodium bisulphite treatment of unmethylated and methylated DNA

M| MM| | 5’ggg gcg gac cgc 5’ggg gcggac cgc

NaBiSO, NaBiSO,

1r r 5’ggg gug gau ugu 5’ggg gcggau cgc 1 1 1 M MM

Unmethylated region Methylated region

203 The modification protocol requires 0.001-lpg of DNA. The CgGenome™ DNA modification kit was used. In a screw-cap 2ml centrifuge tube, 7pl of 3M sodium hydroxide was added to Ipg of PC3, DU145, LNCaP, RD, and 1542NP DNA dissolved

in lOOpl of sterile water, mixed and placed in a water bath at 37°C, for lOmins. The

NaOH renders the DNA single stranded. 550pl of Modification Reagent I was added and the sample vortexed. The samples were incubated at 50°C for 18 hours in a water bath. Reagent I contains sodium bisulphite, which is acidic, the pH was titrated to 5.0 by the addition of 3M NaOH.

To purify the modified DNA, 750pl of reagent II and 5pl of reagent III was added and left at room temperature for 10 minutes. The samples were centrifuged at 8000rpm for

lOsecs and the supernatant discarded. The pellets were washed using 1ml of 70% ethanol, the sample vortexed and centrifuged at 8000 rpm for 10 seconds, the supernatant was discarded. This was repeated twice. A final wash step using 1ml 90% ethanol was performed and the supernatant removed. 25 pi of buffer TE was added to the sample pellet, the tube vortexed and incubated in a water bath for 15mins at 55°C.

The tube was centrifuged at 8,000 rpm for 3mins and the supernatant, containing the modified DNA drawn off. The DNA was stored at -20°C.

5.2(d)3. PCR

Primers for the promoter region of p2 ; WAF1/api Were designed as outlined above.

‘Modified’ primers were designed assuming cytosine was converted to uracil and

‘Unmodified’ primers were also designed assuming cytosine has not been deaminated.

Primers were designed so they did not anneal over a CpG dinucleotide. Primers yielded product containing the CpG site at -692 - the SIE-1 STAT binding site.

204 Table 5.3 p21 WAF1/CIPI primers

Primer sequence Annealing Primer temp p21 M l-sense TTT GTT GT AT G ATTT G AGTT AG 60°C p21 Ml-antisense TAATCCCTCACTAAATCACCTC p21 M2-sense GTTTT GTT GGGGT GTT AGGT G 60°C p21 M2-antisense CACACCTCAACTAACACAACT p21 U1 -sense ATCATTCTGGCCTCAAGATGC 62°C p21 U1-antisense CGGCTCCACAAGGAACTGAC

Primers 1 and 2 denote different sites in the promoter region in cell, M = primers for modified DNA, U = primers for unmodified DNA.

PCR was performed on PC3, DU 145, LNCaP using primers Ml, M2, and U1 in a

Perkin Elmer-Gene Amp 2400 thermal cycler (Perkin Elmer Ltd., UK), at 95°C for

150s, then 36 cycles at 95°C for 45 s, 60-62°C for 30 s and 72°C for 45 s, followed by an extension step at 72°C for 10 mins. The PCR products were electrophoresed on a

1.5% agarose gel. Product was cut out from the gel using a number 11 scalpel blade and placed in a 1.5ml Eppendorf™ tube and weighed.

205 5.2(d)4. DNA extraction from agarose gel

A 3x volume of buffer QC to gel was added to the section cut out from agarose gel. For example if 50mg of gel was cut out 150ml of buffer QC was added. The tube was placed at 50°C for 10 mins and vortexed every 3 mins until fully dissolved.

A lx volume 100% isopropanol was added and the sample vortexed (if 50mg of gel, then 50ml of isopropanol was added). The sample was pipetted onto a QIAspin column and centrifuged at 13,000rpm for lmin and the eluate discarded. DNA binds to the silicone membrane of the column. 500pl buffer QG was added and the sample centrifuged at 13,000rpm for lmin. Buffer QG was a wash buffer.

The sample was washed by adding 750pl of buffer PE and centrifuged at 13,000rpm for

1 min. The washing buffer was discarded and the column spun for a further 3minutes to dry the sample. The column was placed in an Eppendorf™ and 30pl buffer EB was added and left for lmin. Buffer EB alters the charge distribution of DNA so that the

DNA no longer adheres to the silicone membrane. The sample was centrifuged at

8,000rpm so that the eluate contained the extracted DNA. 3pl of DNA mixed with 2pl of lx gel loading buffer was run on a 1.5% agarose gel to check DNA extraction and to estimate DNA concentration for sequencing.

206 5.2(d)5. Sequencing reaction

The protocol for a sequence reaction was the same as a PCR reaction except 1) in addition to standard NTP, fluorescently labelled dideoxy NTPs are added (A-green, C- blue, G-black and T-red), these emitted a specific wavelength of light when excited by a laser. The dideoxyNTPs prevent binding of further NTPs and so stop further elongation of the DNA strand 2) only the sense or the antisense primer is included.

The sequence reaction involves a primer annealing to the sense strand of the DNA fragment and NTPs elongating from the primer. When a dideoxy NTP binds the elongation phase is stopped. The result is many copies of parts of the sense strand ranging form lbp to the product size length e.g.220bp, each with a fluorescently labelled dinucleotide at the 3’ end. To confirm the results the same technique can be performed using an antisense primer, which will generate product with a labelled 5’end.

The reaction was set up as follows.

The DNA yield from the gel extraction process was classified very dim to bright and was added to DEPC water in a 200pl tube as below:

Gel appearance Volume DNA (pi) Volume water (pi) Bright 3 11 Moderate 5 9 Dim 7 7 Very dim 10 4

lpl of sense primer at 4pmol//jal and 5 pi of solution containing NTP, ddNTP and polymerase enzyme was added (Terminator Ready Reaction Mix from the ABI

PRISM® dRhodamine Terminator Cycle Sequencing Ready Reaction Kit) and the

207 sample mixed. Oil was added to prevent evaporation during heating. The solution was then placed in a Perkin Elmer-Gene Amp 2400 thermal cycler (Perkin Elmer Ltd., UK) and programmed as below:

96°C lmin

96°C 30secs

50°C 15secs

60°C 4mins

4°C hold

The oil was removed and the sample transferred to 500pl tube containing 2pl of 3M sodium acetate and 50pl of 95% ethanol, mixed by vortexing and placed on ice for 15 mins.

Precipitation of PCR products was subsequently performed by incubating with a mixture of 50pl of 95% ethanol and 2pl of 3M sodium acetate on ice for 15 minutes.

The sample was centrifuged for 25minutes at 13,000rpm and the supernatant discarded.

The remaining DNA pellet was washed with 70% ethanol (lOOpl). The pellet was stored in the dark. Automated sequencing was performed by running the fluorescently labelled

DNA fragments on the ABI PRISM® 377 DNA Sequencer. The PCR products were loaded into the top of a polyacrylamide gel with product separating into distinct bands according to size and charge. Near the base of the gel the bands passed through the beam of a constantly scanning laser, which excited the fluorescent label attached to

DNA bands. The wavelengths of the resulting light emissions were identified and the signal quantitated. Thus the wavelengh identified the terminal base characteristic of each band, and the quantitation of each band’s fluorescent signal gave the relative number of fragments within each band. In addition to nucleotide sequence text files the automated sequencer also provided trace diagrams.

208 5.3 Results

5.3(a) Cell Culture

The results of the cell counting after 5-Aza-CdR treatment are shown below.

Figure 5.3 The number of cells harvested after treatment with 5- Aza-CdR or PBS (control)

Number of cells

PC3 LNCaP DU145 RD NP

Cell line

□ 5-Aza-CdR ■ PBS

Both Figures 5.3 and 5.4 demonstrate treatment with 5-Aza-CdR has a profound effect on the rate of cell division in the tumour cells. The rate of cell turnover is 1.9x in PC3,

1.8x in LNCaP, 1.3x in Du 145 and 21x in RD greater in the untreated cells. There was also a dramatic difference in the size and morphology of the cells treated with 5-Aza-

CdR as shown in PC3 cells in Figures 5.4a and 5.4b.

209 Figure 5.4a PC3 ceils after 24 hours treatment with 5-Aza-CdR and 5 days growth in standard media (xlO objective).

Figure 5.4b PC3 cells after 24 hours treatment with PBS and 5 days growth in standard media (xlO objective).

210 5.3(b) RT-PCR

RT-PCR was performed on cDNA derived in triplicate from RNA of the treated and untreated cell lines. In all three metastatic prostate cancer cell lines p21WAFi/CIPI expression was either low or undetectable in the untreated populations. In the prostate cancer cell lines treated with 5-Aza-CdR, p 2 lWAF1/CIPI PCR product of 834 base pairs was seen (Figure 5.5). This implies 5-Aza-CdR treatment results in the re-expression of p2JHAFI ( irl gene, which is otherwise inactivated in these prostate cancer cell lines. In the untreated RD cell line p21WAFI CIPI was expressed at a low level, following 5-Aza-

CdR treatment gene expression increased. In the 1542NP cells p 2 lWAFI/CIP1 expression was present in untreated and treated cells. The GAPDH positive control confirmed equivalent quantities of cDNA in the treated and untreated pairs. No product was seen in the ‘no cDNA’ control.

Figure 5.5 Agarose gel showing RT-PCR of 5-Aza-CdR treated (+) and untreated

(-) cell line DNA. mma )p21w‘ I ■ o TtwiH

b) GAPDH gene ■ ■ ■

211 5.3(c) Sequencing the p21 WAF,/C,P1 promoter

To provide further evidence that inactivation of p21WAFIC,pl in the prostate cancer cell lines was due to methylation in the promoter, cell line DNA was modified by sodium bisulphite treatment and sequenced. In all three prostate cancer cell lines (PC3, LNCaP,

DU 145) cytosine was conserved following sodium bisulphite treatment at positions -

896, -856 and -692 in the p21WAFI/CIPI promoter, confirming methylation at these sites

(Figure 5.6). In the RD cell line, a positive control, cytosine was also conserved at the same positions (-896, -856 and -692) (Figure 5.7a). indicating that these sites are methylated, in agreement with previous studies (Chen et al. 2000) (Table 5.4).

Figure 5.6 shows part of the p21nAFI/CIPI promoter region in modified PC3 cells.

The cytosine residue is conserved at position -896 (a), -856 (b) and at the SIE-1 binding site at -692 (c), the latter shown here using the antisense primer.

Figure 5.6a Figure 5.6b Figure 5.6c SIE-1 -896 -ovzr Q'j A C G G G T C C G A

The cytosine nucleotide within the SIE-1 site is shown more clearly in Figure 5.6c, where the antisense primer was used. This confirms the cytosine is conserved in the

SIE-1 binding site, suggesting it is methylated in the C-5 position. Similar findings were

212 observed in LNCaP and DU 145 cell lines, cytosine was conserved in the -896, -856 and

-692 (SIE binding site) positions.

In contrast, DNA from the normal prostate epithelial cell line, 1542NP, used as a negative control, was not methylated at the positions -896 and -856. This suggests that methylation at these sites is cancer specific and not an artefact due to cell culture.

Figure 5.7 Positive and negative controls: a) Preservation of the cytosine residue in

RD at -896 and b) conversion to thymine in NP.

Figure 5.7a RD positive Figure 5.7b NP negative control at -896 control at -896

213 Table 5.4 Results of p 2 lWAFI/CIP1 promoter sequencing after sodium bisulphite DNA

DNA Site 5’-3’

-896 -856 -692 -471 -377 -278-78

PC3 MM MNDND ND

LNCaP MM MMMU

DU145 M M M MM U

RD M M MNDNDND

1542NP UUNDNDNDND

Binding site cap SIE-1 cap

modification.

(M=methylated, U=unmethylated, M/U=partial methylation, ND=not determined)

In all samples, the cytosine residues that were not part of a CG site were converted to thymine by sodium bisulphite modification, confirming successful modification.

Cytosine methylation was also seen at two other sites (-471 and -377) in LNCaP and

DU 145. The sequence for PC3 DNA at these sites could not be reliably interpreted.

There is a CpG island in the p 2 lWAF,/C,PI gene promoter downstream of the sites

discussed. The following sites were sequenced in LNCaP and DU 145 cell line DNA

after modification and were not methylated: - 278, -259, -251, -245, -206, -192, -172, -

152, -149, -123, -114, -108, -104, -101, -99, -91, -87 and -78.

214 5.4 Discussion

The pivotal role of CDKIs and in particular p2]HAF1/CIPI in control of the cell cycle as well as modulators of apoptosis and differentiation implies a tumour suppressor function (Hirama & Koeffler 1995; Toyoshima & Hunter 1994). The aim of this study was to assess the expression of p21WAFI/CIPl in metastatic prostatic cell lines and to determine if the low levels of expression observed result from methylation of the p21 promoter. \ /

\] f e j W A F i c i p i gene eXpressjon was iow or undetectable using RT-PCR in the metastatic prostate cancer cell lines and RD control but was highly expressed in normal prostate epithelial cell lines (1542NP). Following treatment with the demethylating agent 5-Aza-

CdR, p21l*AFI'CIP1 was re-expressed in the metastatic prostate cancer cell lines. This suggests p2 7 WAF,/C1P1 is inactivated by methylation. Low levels of expression were present in PC3, LNCaP and RD untreated cells presumably because either methylation is incomplete or because methylation does not completely inhibit gene expression.

SIE-1 (Sis-inducible-element-1) sites (TTCNNNGAA) are found at -692, -2557 and -

4232 (SIE-1, SIE-2 and SIE-3, respectively), in the p21 WAFI/CIPI promoter. STAT1, a cytoplasmic protein which binds to SIE sites (Darnell, Jr. 1997), is an essential component of the IFN-y signal transduction pathway (Xu et al.). Activation of STAT1 in response to IFN-y results in up-regulation of the p2Jlj'Abl/CIP] gene and inhibition of cell growth (Xu et al.). Methylation at the SIE-1 binding site (-692) inhibits binding of

STAT1 and decreases p 2 lWAFl/CIP1 gene expression (Chen et al. 2000). The SIE site at -

692 was methylated in all 3 prostate cancer cell lines. Inhibition of ST AT binding by methylation may be the mechanism by which the p 2 iWAbI/CIPI gene is silenced in prostate cancer cell lines.

215 CG sites at -896, and -856 were methylated in all three prostate cancer cell lines but not in the normal prostate epithelial cell line, suggesting that methylation at these sites is cancer specific. The CG site at -896 is not a recognised transcription factor binding site

(TFSEARCH: Searching Transcription Factor Binding Sites (ver 1.3) 2004). A myeloid zinc finger protein MZF1 does bind 7 base pairs downstream at -889 (TFSEARCH:

Searching Transcription Factor Binding Sites (ver 1.3) 2004). MZF-1 is a transcription factor of the Kruppel family of zinc finger proteins (Gaboli et al. 2001). In in vitro transient transfection experiments, MZF-1 can alter transcription in haemopoietic and non-haemopoietic cells (Hromas et al. 1996; Morris et al.). It is feasible that methylation close to the MZF-1 binding site alters the structure of the binding domain affecting transcription. Another Kruppel-like transcription factor, KLF6, is mutated in prostate cancer (Narla et al. 2001). Up regulation of wild type KLF6 increases expression of the p 2 iWAF,/CIPI gene fivefold and significantly reduces cell proliferation.

Unlike the wild type mutant KLF6 does not up-regulate p 2 iWAF,/C,PI gene, suggesting

KLF6 is a tumour suppressor gene in prostate cancer. A germline KLF6 single nucleotide polymorphism is significantly associated with an increased relative risk of prostate cancer in men, regardless of family history of disease (Narla et al. 2005). KLF6 interacts with DNA through a GC box promoter element, not in the regions examined in this study (Narla et al. 2001). The work by Narla et al. on KLF6 does however confirm a role for Krupel-like factors in the development and progression of prostate cancer.

Site -856 of the p 2 lWAFI/c,PI gene promoter was methylated in all three cancer cell lines and is a transcription factor binding site (TFSEARCH: Searching Transcription Factor

Binding Sites (ver 1.3) 2004). The transcription factor ‘cap’ binds to sequence

TCACTCGT and triggers transcription initiation (Bucher 1990). The CG at -377 was

216 methylated in LNCaP and DU 145 cancer cells and is also a cap binding site (Bucher

1990), in this instance binding to TCAGCGGT. Little is known about cap in normal or malignant cells.

The 18 CpG sites downstream of the area studied were unmethylated in two cell lines; this implies the p2J^AFFCIPl promoter is not inactivated by methylation at these sites. p2JliAF, c,PI is one of many cell cycle regulators and fits into the complex pathways of cell cycle control. It may be inactivated in several ways during tumourigenesis and methylation of upstream transcription factors may account for the reactivation of p 2 ]UAf-iciPi gene following treatment with 5-Aza-CdR. However sequencing experiments confirmed several sites on the p2JWAFIC,PI promoter were methylated and this suggests inactivation by methylation maybe the mechanism of p 2 jM'AFIC,PI inactivation in the cell lines examined.

Further studies are required to examine whether these findings are present in human prostate cancer samples. Results from several preclinical studies have demonstrated that decitabine (5-AzaCdR) effectively demethylated and reactivated different tumour suppressor genes in cell-lines from breast cancer, mesothelioma, lung, gastric, colorectal cancer and haematological malignancies. The role of decitabine in prostate cancer is under evaluation (NCI-T95-0070). Re-activation of tumour suppressor genes including p21WAFI C,PI by demethylation may play a role in prostate cancer therapy in the future.

217 5.5 Conclusions

The p 2 iHAF1 ap/ gene product has a number of roles involving cell cycle control and apoptosis. With this in mind it is unsurprising that this gene has been implicated as a tumour suppressor gene in a variety of solid and haematological malignancies. This study demonstrates p 2 lWAFI/C,PI is poorly expressed in metastatic prostate cancer cell lines and sequestration of the DNA methyltransferase enzyme results in re-expression.

Furthermore, several cytosine residues in the promoter region are methylated, in particular the site of ST ATI and cap binding. The inhibition of the STAT1 signalling pathway by methylation of the p21 promoter may inactivate thep21WAFI:C,PI tumour

suppressor gene in prostate cancer.

218 Chapter 6

Conclusions and Future Directions

The natural history of prostate cancer is so variable that reliable predictors of disease outcome are desperately needed to target patients who need radical and adjuvant treatment to prevent disease progression while avoiding unnecessary treatment in those with indolent disease. Current clinico-pathological markers, including Gleason grade,

stage, pre-operative PSA and margin status are inadequate at predicting biochemical outcome in all men. In this thesis we have used a battery of markers to try and identify critical areas of allelic imbalance, which may be used to predict PSA failure after radical prostatectomy.

Markers were selected from across the spectrum of prostate cancer biology, ranging

from 8p 12-21 deletions found in high-grade prostatic intra-epithelial neoplasia and early

invasive tumours to lip and 18q21 losses seen in metastatic and hormone refractory

disease. To predict PSA outcome after radical prostatectomy, patients were matched for the best current markers of disease progression. One area of deletion at 1 Op 12.1 was

significantly more common in patients who developed biochemical relapse. Markers

associated with metastatic disease on 1 lp and 18q21 were more commonly seen in men

who relapsed, though this did not reach significance a larger study may demonstrate

these to be important, albeit infrequent, predictors of PSA failure.

Future studies could repeat the allelic imbalance experiments on a larger cohort of patients focussing on markers, including D10S211, which showed the greatest

differences between the two groups. Alternatively a prospective study could be

undertaken using a select group of these markers to examine the allelic imbalance of

219 men at the time of their radical prostatectomy in an attempt to predict which men develop PSA recurrence.

Ultimately the goal of molecular staging is to identify patients who will benefit from radical treatment before they undergo that treatment and which patients would be better managed by less invasive techniques. Comparing the genetic changes in radical prostatectomy specimens with the corresponding prostate biopsies will demonstrate how reliably genetic changes can be identified in biopsy tissue. As prostate cancer is a multifocal heterogeneous tumour the biopsy findings may not mirror the tumour characteristics of the whole prostate. The disparity between biopsy and radical Gleason

scores demonstrates this variability. This disparity may be further exacerbated, as there appears to be greater genetic variation than Gleason variation within one radical prostatectomy specimen. Future studies could compare the allelic imbalance findings of the radical prostatectomy specimen with the Al in the pre-operative biopsy tissue. If

markers are readily identified in both biopsy and radical prostate tissue this would have

an important impact in directing appropriate radical treatment.

In clinical practice there is a pressing need for new therapeutic agents to treat hormone

relapse disease, as there is a lack of treatment options and the prognosis for a man with

androgen independent prostate cancer is less than a year. Improving our understanding

of how a prostate cancer clone develops androgen independence will allow the

development of such agents. Meanwhile targeting genome wide aberrant methylation

with agents such as decitabine may improve the outcome of men with metastatic

prostate cancer.

220 Despite some knowledge of the genetic alterations found in prostate cancer the molecular mechanisms leading to the initiation and progression of the disease is poorly understood. In colorectal cancer a model of tumorigenesis in which specific genetic changes result in the stepwise progression from benign to malignant epithelium was reported over a decade ago. Prostate cancer appears much more complicated in its aetiology and pathogenesis. Only a small number of mutations have been identified, and no mutations have been reported in early stage disease. This is because only a tiny fraction of the 30,000 human genes have so far been screened for prostate cancer mutations. The Sanger Institute’s Cancer Genome Project is systematically screening for mutations in a number of malignancies, including prostate, and this should shed more light on these changes in prostate cancer. Another difficulty in determining the pathogenesis of prostate cancer is the critical changes may occur outside the coding sequence and it is difficult to predict the consequences of sequence variation in these regions. Furthermore genetic variation other than sequence variation may affect the oncogenes and tumour suppressor genes. Hypermethylation is proposed in this thesis as the mechanism of inactivation of the tumour suppressor gene p 2 lWAFI C,PI. Finally the loss of one tumour suppressor gene may be sufficient for tumour suppressor gene inactivation - haploinsufficiency, as reported in NKX3.1 and PTEN. The difficulty is we do not know how common haploinsufficiency is and so how important the loss of a single gene may be in prostate cancer development.

Although molecular staging of radical prostatectomy patients has been under study for more than a decade, all techniques remain research tools. Still, this area holds great promise for improving the accuracy of staging, providing a more accurate prognosis and tailoring treatment for the individual man with clinically localised prostate cancer.

Likewise identifying critical genetic or epigenetic changes will allow the development of novel therapeutic tools in the treatment of all stages of prostate cancer.

221 222 APPENDICES

Appendix 1. Power calculation: based on the McNemar test. Assuming a significance level of 5% (p=0.05) Proportion with category Number of cases Number of controls Power (%) Case (%) Control (%) 70 40 44 70 30 68 60 30 44 60 20 69 20 20 40 20 26 40 10 55 30 10 33 20 5 28 70 40 56 70 30 81 60 30 56 60 83 20 40 20 40 20 34 40 10 69 30 10 42 20 5 36 70 40 62 70 30 86 60 30 62 60 87 30 30 20 40 20 37 40 10 74 30 10 47 20 5 40 70 40 75 70 30 95 60 30 75 60 95 30 60 20 40 20 47 40 10 86 30 10 59 20 5 51 70 40 75 70 30 95 60 30 75 60 40 40 20 95 40 20 48 40 10 86 30 10 59 20 5 51 70 40 87 70 30 99 60 30 87 60 20 99 40 80 40 20 60 40 10 95 30 10 72 20 5 63

223 Appendix 2. Preoperative features of matched patients

Bx Longest Bx Longest Bx No of length Biopsy Bx No of length Biopsy Random PreO p Bx No of positive involved Bx Highest G leason Random PreO p Bx No of positive involved Bx Highest G leason num ber Age (yrs) PSA co res co res (mm) % in core grade num ber Age (yrs) PSA c o res co res (mm) % in core grade 1 62 14.2 9 6 4+3 5 72 7.9 6 4 6 50 3+3 2 58 12 2 2 8 40 3+3 9 63 8 NA 3 59 18 6 3 16.1 95 4+4 10 73 16.7 6 2 2 15 3+4 4 69 10.1 NA 12 55 8.8 5 4 9 30 3+3 6 63 16.7 6 1 10 50 3+4 13 59 5.9 NA 7 55 14.7 4 1 14 100 3+3 17 65 5.1 5 2 8 100 2+3 11 54 4.3 6 1 12 75 3+3 19 59 8.3 3 2 7 15 3+3 15 68 9.1 NA 20 63 6.4 6 3 9 55 3+4 16 61 8 6 6 14 100 3+5 21 59 16.3 4 1 15 100 4+3 18 57 18.6 4 4 9 100 3+3 23 68 17.8 NA 22 57 8.6 3 1 8 44 3+5 26 61 12 6 3 13 65 3+4 24 60 16 6 6 18 100 4+3 28 57 7 NA 25 59 8.6 NA 29 59 7.4 NA 27 53 7.5 8 3 30 69 8.8 6 2 6 30 3+3 35 65 5.1 6 4 10 90 4+3 31 63 9.7 7 1 16 80 4+3 38 62 10.2 3 2 6 70 2+4 32 71 11.3 NA 39 49 18 NA 33 61 16 6 2 2 13 2+3 42 53 15.7 6 1 1 8 4+4 34 66 9 6 2 8 15 2+3 43 56 5.2 6 3 6 60 4+4 36 67 13.4 6 3 5 50 3+4 45 51 9.8 NA 37 70 13 NA 50 70 15.7 6 2 20 100 4+3 44 60 5.5 6 2 5 15 3+3 52 72 5.9 6 4 15 100 4+4 48 68 12.9 NA 55 56 14 6 2 12 65 5+4 49 73 10 NA 58 56 16 NA 51 52 17 6 4 1.3 100 3+4 59 58 8 7 6 3 20 5+3 53 62 16.6 4 4 3+3 60 57 8.7 NA 54 57 11 NA 64 54 14.4 6 4 10 50 4+3 56 65 16.6 50 74 54 9.9 1 4 4+? 57 65 10.5 NA 75 66 15.8 NA 61 66 15.1 NA 78 71 12.4 6 3 20 100 3+3 62 53 9.3 4 2 10 100 3+3 79 62 9.7 11 4 5 35 3+5 77 71 7.5 6 5 6 33 3+3 Mean 59.6 Mean 64 (sd6.1) (sd 6.2) Appendix 3. Radical prostatectomy pathological findings of in matched patients (Relapse)

PATHOLOGICAL OUTCOME - RELAPSE

Randomis m Radical Lym pho / ation number 1 12-Jan-98 3+4 7 4 32 0 CIP 0 0 0 0 1 0 0 2 03-Feb-98 4+3 7 3.6 25 0 0 0 1 0 0 1 0 0 3 13-May-98 4+3 7 2.837 0 0 0 1 0 0 1 0 0 4 01-Apr-98 4+3 7 2 149 0 0 0 1 0 0 1 0 0 6 10-Dec-97 4+3 7 1.8 48 0 0 0 1 0 0 1 0 0 7 04-Jul-94 2+3 5 2.3 52 0 CEP 0 0 0 0 1 0 0 11 17-Oct-97 3+3 6 1.2 29 0 CEP 0 0 0 0 0 0 0 15 27-Nov-95 2+4 6 1.7 52 0 0 0 1 0 0 0 0 0 16 23-Mar-99 4+3 7 3.2 46 0 0 0 1 0 0 1 0 0 18 13-Jan-95 3+2 5 2.2 29 0 CEP 0 0 0 0 1 0 0 22 02-May-95 2+3 5 2.8 57 0 CIP 0 0 ? 0 1 0 0 24 16-Sep-98 4+3 7 4 32 1 0 0 1 0 0 1 0 0 25 08-Jul-98 4+3 7 5.1 54 0 CEP 0 0 0 0 1 0 0 27 21-Jul-97 4+3 7 2.6 40 1 CEP 0 0 0 0 1 0 0 35 09-Sep-96 3+2 5 5.7 48 0 CIP 0 0 0 0 1 0 0 38 18-Oct-94 2+4 6 10 92 1 CEP 0 0 0 0 1 0 0 39 09-Sep-96 4+3 7 2.1 29 0 CEP 0 0 0 0 0 1 0 42 31-Dec-97 4 + 4 8 1.5 32 1 0 0 1 0 0 0 0 0 43 21-Sep-98 4+3 7 5.2 31 1 0 0 1 0 0 0 0 0 45 05-Aug-96 3+4 7 1.6 28 0 CEP 0 0 0 0 0 0 50 28-Oct-97 4+3 7 3.5 27 0 0 0 1 0 0 0 0 0 52 09-Sep-98 4+3 7 5.5 30 1 0 0 1 0 0 0 NS 0 55 19-Jan-98 4+3 7 5.6 49 0 0 0 1 0 0 1 0 0 0 58 17-Oct-95 4+3 7 5.8 36 0 0 0 1 0 0 0 — c 59 26-Jul-00 4+3 7 1.9 35 1 0 0 1 0 0 1 ... .0...... 0 60 15-Nov-00 4+3 7 0.5 38 1 0 0 1 0 0 0 0 0 64 13-Sep-00 5+3 8 10.5 104 1 0 0 1 0 0 1 0 0 74 01-Jun-94 3+4 7 3.7 53 0 0 0 1 0 0 0 0 0 75 17-Jul-96 4+3 7 3 24 0 CIP 0 0 0 1 0 0 78 15-Dec-99 4+3 7 3.9 49 0 0 0 1 0 0 0 1 0 A I I - , , O . C o A A QQ n fi 1 n o o o o 79 10-May-00 3+5 8 1.4 o o u U u 1 V Mean 6.74 Median Median 38 3 (sd 0.8) (0.5-10.5) (24-149) CEP = Transverse extraprostatic positive margin, CIP = Transverse intraprostatic positive margin, NS - Not Submitted 225 Appendix 3. Radical prostatectomy pathological findings in matched patients (Non-relapse)

Randomis ation num ber 5 04-Feb-98 3+4 7 2.5 62 0 0 0 1 0 0 1 0 0 9 14-Dec-95 3+4 7 2.3 45 0 0 0 1 0 0 1 0 0 10 02-Dec-97 4+3 7 2.8 82 0 CEP 0 0 0 1 0 0 12 29-Feb-96 4+3 7 1.3 52 0 0 0 1 0 0 0 0 0 13 04-Aug-97 3+4 7 1.2 48 0 0 0 1 0 0 0 0 0 17 11-Mar-97 3+3 6 0.9 45 0 0 0 1 0 0 1 0 0 19 02-Apr-96 3+4 7 3.4 47 0 0 0 0 0 1 0 0 20 31-Oct-95 3+4 7 2.2 48 0 0 0 1 0 0 0 0 0 21 10-Dec-97 4+4 8 1.7 37 0 0 0 1 0 0 0 0 0 23 02-Oct-96 2+3 5 8.4 50 0 0 0 1 0 0 1 0 0 26 22-Jan-97 4+3 7 5.4 41 0 0 0 1 0 0 1 0 0 28 27-Sep-93 2+3 5 3 39 0 CIP 0 0 0 1 1 0 29 27-Jan-98 3+4 7 1.6 50 0 0 0 1 0 0 0 0 0 30 21-Apr-98 3+4 7 1.1 39 0 0 0 1 0 0 0 0 0 31 22-Sep-98 3+4 7 0.9 17 0 0 0 1 0 0 0 0 0 32 27-Apr-98 3+4 7 2.6 27 0 0 0 1 0 0 1 0 0 33 09-Oct-95 3+2 5 1.5 57 0 0 0 1 0 0 1 0 0 34 12-NOV-96 3+2 5 1.7 59 0 0 0 1 0 0 1 1 0 36 14-0ct-97 4+3 7 3.2 64 1 0 0 1 0 0 1 1 0 37 10-Dec-91 3+4 7 1.7 NA 1 0 0 1 0 0 0 0 NS 44 09-Feb-98 3+4 7 1.1 NA 0 CIP 0 0 0 0 0 0 48 24-Apr-96 4+2 6 1 62 0 0 0 1 0 0 1 0 0 49 08-Sep-97 4+3 7 5.6 37 0 CEP 0 0 0 0 0 0 51 28-Jul-98 3+4 7 2.2 38 0 0 0 1 0 0 1 0 0 53 18-Sep-95 4+3 7 5 6 47 0 0 0 1 0 0 1 1 0 54 20-May-98 3+4 7 0.6 60 0 0 0 1 0 0 0 0 0 56 06-May-97 3+4 7 1.1 32 0 0 0 1 0 0 0 0 0 57 20-Aug-97 3+4 7 4.1 25 0 CIP 0 0 0 1 0 0 61 21-Aug-96 4+4 8 1.7 83 0 CIP 0 0 0 1 0 0 62 22-Aug-95 3+3 6 1.9 33 0 0 0 1 0 0 0 0 0 77 12-Nov-97 4+4 8 0.8 38 0 0 0 1 0 0 0 0 0 Mean 6.74 Median 1.8 Median 47

(sd 0.8) (0.8-8.4) (27-83)

CEP = Transverse extraprostatic positive margin, CIP = Transverse intraprostatic positive margin, NS - Not Submitted 226 Appendix 4. PSA follow-up in matched patients - Relapse

Random Date of Post Post PSA 1 Post PSA Post PSA 2 Post PSA Post PSA 3 Post PSA Post P§a 7 Post PSA Post PSA 5 number Surgery PSA1 Date 2 Date 3 Date 4 Date 5 Date 1 09-Jan-98 0.5 20-Jan-98 0.1 27-Feb-98 0.1 02-Mar-98 0.1 03-Jun-98 0.3 17-Sep-98 2 03-Feb-98 0.2 13-Feb-98 0.05 03-Apr-98 0.05 14-May-98 0.1 17-Oct-98 0.1 16-Nov-98 3 13-May-98 0.4 31-May-98 0.1 01-Jul-98 0.4 01-Sep-98 0.1 05-Nov-98 0.2 01-Dec-98 4 01-Apr-98 0.2 14-Apr-98 0.1 26-May-98 0.1 03-Sep-98 0.1 09-Oct-98 0.2 12-Apr-99 6 10-Dec-97 0.3 22-Dec-97 0.1 17-Mar-98 1.2 29-Jun-98 1.3 AA 01-Jul-98 0.3 DXR 08-Jan-99 7 04-Jul-94 <0.5 03-Oct-94 <0.5 16-Jan-95 0.4 13-Apr-95 0.8 01-Nov-95 1.1 26-Jan-96 11 17-Oct-97 0.2 30-Oct-97 0.1 25-Nov-97 0.5 30-Jan-98 0.1 10-Feb-98 0.2 08-Apr-98 15 27-Nov-95 0.1 01-May-96 0.3 19-Jan-98 0.4 06-Aug-98 0.6 12-Nov-98 0.6 09-Feb-99 16 23-Mar-99 0.1 13-May-99 0.1 14-Jun-99 0.1 09-Aug-99 0.1 10-Nov-99 0.2 21-Feb-00 18 13-Jan-95 0.3 31-Jan-95 1.4 01-Jan-97 AA/DXR 23-May-00 0.04 13-Sep-01 22 02-May-95 1.3 01-Jul-96 0.9 20-Apr-99 0.8 21-Oct-99 0.8 03-Apr-00 0.8 05-Jan-01 24 16-Sep-98 0.4 29-Sep-98 0.1 22-Oct-98 0.1 27-Jan-99 0.1 27-Apr-99 0.1 09-Sep-99 25 08-Jul-98 0.1 20-Jul-98 0.1 04-Sep-98 0.1 26-Nov-98 0.2 01-Mar-99 0.4 07-Jun-99 27 21-Jul-97 0.3 31-Jul-97 0.1 01-Oct-97 0.1 01-Jan-98 0.1 01-Apr-98 0.4 14-Nov-98 35 09-Sep-96 0.2 01-Dec-96 0.5 01-Mar-97 0.6 01-Jun-97 0.9 01-Sep-97 1 29-Jun-98 38 18-Oct-94 0.5 01-Jan-95 0.5 01-May-95 0.7 01-Oct-95 0.7 01-Feb-96 1 01-Jul-96 39 09-Sep-96 0.2 09-Dec-96 0.4 18-Mar-97 0.5 04-Apr-97 0.6 02-Sep-97 0.6 09-Dec-97 42 31-Dec-97 0.2 01-Mar-98 0.1 01-Jun-98 0.5 05-Jul-99 43 21-Sep-98 0.5 08-Oct-98 0.2 09-Nov-98 0.4 16-Dec-98 0.7 28-Apr-99 0.4 25-M ay-99 45 05-Aug-96 0.2 17-Dec-96 0.4 07-Apr-97 0.4 06-Jun-97 0.8 13-Aug-97 0.5 19-Sep-97 50 28-Oct-97 0.7 10-Nov-97 0.1 18-Nov-97 0.4 05-Nov-99 0.6 02-Jan-98 0.1 31-Mar-98 52 09-Sep-98 0.1 10-Oct-98 0.5 09-Feb-99 0.5 20-May-99 2.6 01-Aug-00 55 19-Jan-98 0.1 01-Apr-98 0.3 01-Jul-98 58 17-Oct-95 0.3 01-Nov-96 0.5 01-Jul-98 1.5 02-Dec-98 1.4 14-Mar-99 1.82 26-Jan-00 59 26-Jul-00 0.2 07-Jan-01 0.6 03-Apr-01 60 15-Nov-00 No nadir 64 13-Sep-OO 0.2 30-Aug-01 AA 74 01-Jun-94 1.4 17-Mar-95 0.8 23-Jun-95 0.8 31-Aug-95 75 17-Jul-96 1.6 01-Dec-96 3.3 01-Apr-97 78 15-Dec-99 0.1 20-Jan-00 0.4 04-May-00 1.2 8-00 bone mets 08-Sep-00 79 10-May-00 0.1 22-May-00 0.1 10-Jul-00 0.2 13-Nov-OO 0.18 24-Jan-01 0.33 02-May-01 AA = Antiandrogens, DXR = Radiotherapy Appendix 4. PSA follow-up in matched patients - Relapse

POSTOPERATIVE PSA 2. - RELAPSE

R andom Date of PostPSA 6 P o st PSA Post PSA 7 P o st PSA P o st PSA8 P o st PSA P o st PSA 9 P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA n um ber Surgery PostPSA 6 Date 7 Date 8 Date 9 Date 10 10 Date 11 11 Date 12 1 12-Jan-98 0.3 01-Dec-98 1 01-Aug-99 0.9 06-Aug-99 0.59 AA 21-Sep-99 0.3 05-Jan-00 0.4 05-Apr-00 0.4 2 03-Feb-98 0.3 04-Mar-99 0.25 20-Sep-99 0.42 03-Dec-99 0.5 26-Apr-00 0.6 26-Jul-00 0.7 03-0ct-00 0.5 3 13-May-98 0.1 01-Feb-99 0.1 23-Mar-99 0.1 21-Jun-99 0 3 08-Sep-99 0.4 28-Oct-99 0.5 16-Dec-99 0.5 4 01-Apr-98 0.5 26-Oct-99 1.1 03-Apr-00 0.1 Cas 06-Jul-00 0.2 21-Nov-00 0.2 17-May-01 0.3 05-Nov-01 6 10-Dec-97 0.35 09-Jul-99 7 04-Jul-94 0.3AA 26-Apr-96 0.4 21-Oct-96 11 17-Oct-97 0.1AA/DXF 28-Sep-98 0.1 25-Mar-99 0.02 26-Apr-99 0.04 24-Apr-01 Died AML 15 27-Nov-95 0.6 13-May-99 0.8 07-Jun-99 0.9 08-Aug-99 0.5 18-NOV-99 0.8 09-Feb-00 0.7 10-May-00 0.6 16 23-Mar-99 0.2 18-Apr-00 0.4 11-Aug-00 0.2 13-Dec-00 0.2 14-Mar-01 0.1 18-Jul-01 0.1 14-Nov-01 18 13-Jan-95 22 02-May-95 0.7 31-Jul-01 0.4 18-Jan-02 24 16-Sep-98 0.2 22-Mar-00 0.5 29-Jun-00 0.1 AA 27-Sep-00 0.1 27-Mar-01 0.1 20-Sep-01 0.1 03-Oct-01 25 08-Jul-98 6.4 AA 26-Jul-99 2.1 AA 08-Nov-99 17.8 16-Mar-00 18.9 31-Mar-00 6.58 16-May-00 5.5 24-Aug-00 42 27 21-Jul-97 0.5 13-Feb-99 0.3 17-Jul-99 0.3 15-Jan-00 0.3 15-Jul-00 0.4 20-Jan-01 0.5 23-Jun-01 1.1 35 09-Sep-96 38 18-Oct-94 2.1 01-Jan-97 0.5 01-May-99 0.5 09-Aug-99 0.5 01-Dec-99 0.5 08-Feb-00 0.1 29-Mar-00 0.1 39 09-Sep-96 0.6 13-Mar-98 1.1 23-Jun-98 1.1 01-Oct-98 0.2 02-Dec-98 0.2 19-Feb-99 0.2 24-NOV-99 0.3 42 31-Dec-97 43 21-Sep-98 45 05-Aug-96 50 28-Oct-97 0.1 30-Jun-98 0.1 29-Sep-98 0.3 29-Mar-99 0.3 08-Apr-99 0.3 01-Jul-99 0.4 04-Oct-99 0.6 52 09-Sep-98 55 19-Jan-98 58 17-Oct-95 59 26-Jul-00 60 15-Nov-00 64 13-Sep-00 74 01-Jun-94 75 17-Jul-96 78 15-Dec-99 79 10-M ay-00 0.26 06-Jun-01

AA = Antiandrogens, DXR = Radiotherapy Appendix 4. PSA follow-up in matched patients - Relapse

Random Date of P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA Post PSA 18 P o st PSA P o st PSA P o st PSA num ber Surgery 14 14 Date 15 15 Date 16 16 Date 17 17 Date 18 Date 19 19 Date 20 1 09-Jan-98 2 12-Jan-01 0.8 02-Mar-01 2 03-Feb-98 0.04 DXR 12-Nov-01 3 13-May-98 1.7 10-Jul-00 2.2 16-Oct-OO 2.5 21-Nov-00 6.2 26-Jan-01 3.4 Cas 29-Mar-01 146 25-Jun-01 156.5 4 01-Apr-98 6 10-Dec-97 7 04-Jul-94 11 17-Oct-97 15 27-NOV-95 0.5 06-Feb-01 0.5 05-Jun-01 Died CaP 5/02 16 23-Mar-99 18 13-Jan-95 22 02-May-95 24 16-Sep-98 25 08-Jul-98 Died CaP 27 21-Jul-97 35 09-Sep-96 38 18-Oct-94 0.1 16-Oct-OO 0.1 12-Feb-01 0.1 14-May-01 0.1 11-Sep-01 39 09-Sep-96 1.3 04-Jun-01 0.3 28-Aug-01 42 31-Dec-97 43 21-Sep-98 45 05-Aug-96 50 28-Oct-97 0.8 13-Jul-00 0.8 31-Oct-OO 1.1 01-Feb-01 0.3 14-Sep-01 0.3 03-Dec-01 52 09-Sep-98 55 19-Jan-98 58 17-Oct-95 59 26-Jul-00 60 15-Nov-00 64 13-Sep-OO 74 01-Jun-94 75 17-Jul-96 78 15-Dec-99 79 10-May-00

AA = Antiandrogens, DXR = Radiotherapy Appendix 4. PSA follow-up in matched patients - Non-relapse ------

Random Date of Post Post PSA 1 Post PSA Post PSA 2 Post PSA Post PSA 3Post PSA Post PSA 4 Post PSA Post PSA 5 number Surgery PSA1 Date 2 Date 3 Date 4 Date 5 Date 5 04-Feb-98 0.1 23-Feb-98 0.1 31-Mar-98 0.1 09-Jul-98 0.1 15-Oct-98 0.1 19-Jan-99 9 14-Dec-95 0.6 22-Dec-95 0.1 01-Dec-96 0.1 01-Aug-98 0.1 07-Jan-99 0.1 08-Jun-99 10 02-Dec-97 0.3 17-Dec-97 0.1 05-Mar-98 0.1 25-Jun-98 0.1 25-Jun-98 Never >0.2 8/01 12 29-Feb-96 0.7 04-Mar-96 0.2 12-Mar-96 0.1 18-Mar-96 0.1 01-May-96 Never >0.2 8/01 13 04-Aug-97 0.1 01-Nov-97 0.1 01-Feb-98 0.1 01-May-98 0.1 29-Jun-98 0.1 22-Aug-98 17 11-Mar-97 0.5 28-Jul-97 0.5 20-Oct-97 0.5 14-Apr-98 0.5 10-Oct-98 0.5 26-Apr-99 19 02-Apr-96 0.5 20-May-96 0.5 22-Jul-96 0.5 21-Oct-96 0.5 21-Apr-97 0.5 20-Mar-98 20 31-Oct-95 0.5 18-Dec-95 0.5 19-Feb-96 0.5 01-Jul-96 0.5 02-Dec-96 0.5 12-May-97 21 10-Dec-97 0.9 22-Dec-97 0.1 02-Feb-98 0.1 05-May-98 0.1 10-Sep-98 0.1 18-Nov-98 23 02-Oct-96 0.1 29-Oct-96 0.1 11-Mar-97 0.05 22-Sep-97 0.05 14-Mar-98 0.05 01-Sep-98 26 22-Jan-97 1.2 03-Feb-97 0.1 01-Apr-97 0.1 01-Jul-97 0.1 01-Oct-97 0.1 08-Jul-98 28 27-Sep-93 0.5 22-May-95 0.5 20-Nov-95 0.5 20-May-96 0.5 18-Nov-96 0.5 17-Nov-97 29 27-Jan-98 0.5 02-Mar-98 0.5 15-Jun-98 0.5 11-Feb-99 0.5 23-Aug-99 0.5 09-Dec-99 30 21-Apr-98 0.1 05-May-98 0.1 11-Jun-98 0.1 16-Jul-98 0.1 03-Sep-98 0.1 13-Mar-99 31 22-Sep-98 0.1 16-Nov-98 0.1 15-Mar-99 0.1 17-Jun-99 0.1 28-Sep-99 0.1 27-Mar-00 32 27-Apr-98 never>0.2 01-Aug-01 33 09-Oct-95 never>0.2 01-Aug-01 34 12-Nov-96 0.5 13-Jan-97 0.5 24-Feb-97 0.5 28-Jul-97 0.5 26-Jan-98 0.5 25-Jan-99 36 14-Oct-97 0.5 02-Feb-98 0.5 05-May-98 0.5 09-Nov-98 0.5 11-Nov-98 0.5 01-May-99 37 10-Dec-91 0.5 07-Dec-92 0.5 21-Jun-93 0.1 01-Sep-95 44 09-Feb-98 0.5 23-Mar-98 0.5 08-Jun-98 0.5 12-Oct-98 0.5 18-Jan-99 0.5 01-Jul-99 48 24-Apr-96 0.1 01-Dec-96 0.5 01-Jun-97 never >0.2 01-Aug-01 49 08-Sep-97 0.5 24-Oct-97 0.5 27-Feb-98 never >0.2 02-Sep-02 51 28-Jul-98 0.02 07-Sep-98 0.2 17-Nov-98 0.02 03-Feb-99 0.02 07-Jun-99 Rose 6/01 53 18-Sep-95 0.1 01-Dec-95 0.1 03-Apr-96 0.1 01-Jun-96 0.10 01-Sep-96 0.1 01-Jan-97 54 20-May-98 never>0.2 01-Aug-01 <0.2 08-Aug-02 56 06-May-97 0.2 19-May-97 0.1 01-Aug-97 0.1 01-May-98 57 20-Aug-97 0.6 29-Aug-97 0.1 07-Oct-97 0.1 15- Jan-98 0.1 22-Apr-98 0.1 17-Aug-98 61 21-Aug-96 1.2 29-Aug-96 0.1 01-Nov-96 0.1 01-Feb-97 0.1 01-Aug-97 62 22-Auq-95 0.5 01-Nov-95 never>0.2 01-Aug-01 77 12-Nov-97 0.1 25-Nov-97 0.2 01-Feb-98 0.1 01-May-98 0.1 01-Oct-98 0.1 21-Apr-99 AA = Antiandrogens, DXR = Radiotherapy 230 Appendix 4. PSA follow-up in matched patients - Non-relapse ------

R andom Date of PostPSA 6 P o st PSA Post PSA 7 P o st PSA P o st PSA8 P o st PSA P o st PSA 9 P o st PSA P o st PSA P o st PSA P o st PSA P o st PSA n um ber _ S “ rcjery__ PostPSA 6 Date 7 Date 8 Date 9 Date 10 10 Date 11 11 Date 12 5 04-Feb-98 0.1 20-Jul-99 0.1 03-Jul-00 0.1 24-Jul-01 9 14-Dec-95 0.1 15-Dec-99 0.1 01-Aug-00 10 02-Dec-97 12 29-Feb-96 13 04-Aug-97 0.1 02-Mar-99 0.12 08-Nov-99 Not >0.2 08-Jan-02 17 11-Mar-97 0.5 25-Apr-00 0.7 30-Apr-01 0.6 29-Apr-02 No treatment 19 02-Apr-96 0.5 23-Sep-98 0.5 25-Mar-99 0.1 13-Apr-00 0.1 24-Apr-01 0.1 17-Apr-02 20 31-Oct-95 0.5 10-Nov-97 0.5 01-Aug-98 0.5 01-Jul-99 0.5 01-Oct-99 0.1 01-Sep-00 21 10-Dec-97 0.1 12-Mar-99 0.1 26-May-99 0.1 31-Aug-99 0.1 20-Dec-99 0.1 03-Jul-00 0.1 28-Nov-00 0.1 23 02-Oct-96 0.05 01-Sep-98 0.05 19-Apr-99 0.05 10-Sep-99 0.05 24-Oct-01 26 22-Jan-97 0.1 02-Aug-99 0.1 01-Sep-00 1.5 09-Nov-00 0.1 27-Mar-01 0.1 26-Jun-01 0.1 01-Oct-01 0.1 28 27-Sep-93 0.5 16-Nov-98 0.5 15-Nov-99 0.1 11-Dec-00 29 27-Jan-98 0.1 26-Feb-01 30 21-Apr-98 0.1 03-Jul-01 31 22-Sep-98 0.1 04-Sep-01 32 27-Apr-98 33 09-Oct-95 34 12-Nov-96 0.5 01-Jul-99 36 14-Oct-97 0.5 01-May-00 <0.1 13-Jun-01 <0.5 10-Dec-01 0.3 01-Jul-02 37 10-Dec-91 44 09-Feb-98 0.5 18-Jan-00 <0.1 17-Jul-00 <0.1 02-Aug-01 <0 1 17-Jul-02 48 24-Apr-96 49 08-Sep-97 51 28-Jul-98 53 18-Sep-95 0.1 01-May-97 never >0.2 54 20-May-98 56 06-May-97 57 20-Aug-97 0.1 04-Feb-99 0.1 05-Aug-99 0.1 25-Jan-00 0.2 07-Aug-00 0.1 24-Aug-00 0.1 14-Nov-OO 0.2 61 21-Aug-96 62 22-Aug-95 77 12-Nov-97 0.1 17-Sep-99 0.1 31-Mar-00 0.1 20-Sep-00 0.1 12-Sep-01

AA = Antiandrogens, DXR = Radiotherapy 231 Appendix 5. Procedure for staining with haematoxylin and eosin

Procedure Time 1. D ew ax sections in X ylene 3 mins 2. Repeat dewax sections in Xylene 3 mins 3. Wash in 100% ethanol 1 mins 4. Repeat wash in 100% ethanol 1 mins 5. Wash 70% ethanol 1 mins 6. Rehydrate in tap water 1 mins 7. Stain with Haemotoxylin 5 mins 8. W a s h in running tap w ater 3 mins 9. Differentiate in 1% acid-alcohol 6 secs 10. Wash Scott's tap water 5 mins 11. W a s h in tap w ater 30 secs 12. Stain with 1% Eosin 5 mins 13. Wash in running tap water 1 mins 14. Dry in 70% ethanol 15 secs 15. Dry in 100% ethanol 30 secs 16. Dry in 100% ethanol 30 secs

232 Appendix 6. The ABgene™ DNA block 96 well plates arranged for genotyping

1 2 3 4 5 6 7 8 9 10 11 12 A 1R 3R 10R 12R 19R 21R 27R 29R 35R 37R 45R 49R B 5R 7 R 15R 17R 23R 25R 31R 33R 39R 43R 51R 53R C 1L 3L 10L 12L 19L 21L 27L 29L 35L 37L 45L 49L D 5L 7L 15L 17L 23L 25L 31L 33L 39L 43L 51L 53L E 2R 4R 11R 13R 20R 22 R 28R 30R 36R 38R 48R 50R F 6R 9R 16R 18R 24R 26R 32R 34R 42R 44R 52R G 2L 4L 11L 13L 20L 22L 28L 30L 36L 38L 48L 50L H 6L 9L 16L 18L 24L 26L 32L 34L 42L 44L 52L water

1 2 3 4 5 6 7 8 9 10 11 12 A 54R 56R 62R 74R B 58R 60R 77R 79R C 54L 56L 62L 74L D 58L 60L 77L 79L E 55R 57R 64R 75R F 59R 61R 78R G 55L 57L 64L 75L F 59L 61L 78L water

Number denotes patient number (matched patients only) R = Radical prostatectomy specimen (cancer) L = Lymph node or seminal vesicle specimen (contol)

233 Appendix 7 Features of the polymorphic markers used in the allelic imbalance stud] / T ype o f Primer G ene Frequency Tissue Technique Reference repeat D1S422 Familial prostate cancer Di 15% of D1S414 ATF3 Di D1S158 H PC 1 (hereditary) Oi 5% of 21 T3 and T4 autoradiography- loss of 60% by eye Latini JM J Urol 2001; 166: 1931-36 D1S305 PSMBA Di

D2S222 not known Di 6% of 17 (informative cases) primary from RP autoradiography-formally assessed Saric T Int J Cancer 1999; 81: 219-224

D5S656 close to APC Di D5S500 close to CTNNA1 Di

D6S501 not known Tetra 11 % of 52 (informative cases) RP specimens autoradiography - 50% loss by eye Cooney KA Cancer Res 1996; 56: 4150-53 D6S314 not known Di 2% of 52 (informative cases) RP specimens autoradiography - 50% loss by eye Cooney KA Cancer Res 1996; 56: 4150-53

D7S480 close to MET Di 0% primary, mets or HRCAP Autopsy HRCAP autoradiography - 50% formally assessed Kawana Y Prostate 2002, 53:60 D7S480 close to MET Di 17% of 46 (informative cases) RP specimens autoradiography - 1 .5x formally assessed Takahashi S Cancer Res 1995; 55: 4114-9 D7S480 close to MET Di 46% of 33 (informative cases) T1-3, NO-1, MO-1 autoradiography - loss of >30% formally assessed Latil A. Clin Cancer Res 1995; 1: 1385-9 D7S480 close to MET Di 27% of 11 (informative cases) RP or LN if LN+ at RP autoradiography - loss <30% formally assessed Zenklusen J Cancer Res 1994; 54: 6370-3 D7S523 close to MET Di 3% of 47 (informative cases) RP and TURP autoradiography - by eye Huge) A Br J Cancer 1999; 79: 551-57 D7S523 close to MET Di 24% of 42 (informative cases) RP specimens autoradiography -1 5x formally assessed Takahashi S Cancer Res 1995; 55: 4114-9 D7S523 close to MET Di 28% of 7 (infomative casesO RP or LN if LN+ at RP autoradiography - loss < 30% formally assessed Zenklusen J Cancer Res 1994; 54: 6370-3

D8S136 NKX3.1 Di 18% of 17 (informative cases) primary from RP autoradiography-formally assessed Saric T Int J Cancer 1999; 81: 219-224 D8S136 NKX3.1 Di 72% of 25 RP and TURP, any grade autoradiography-loss or faint to eye Emmert-Buck MR Can Res 1995; 55: 2959 D8S1991 MSR Di homozygous deletion in single pt RP LN mets in 1 pt multiplex PCR Bova GS Genomics 1996; 35: 46-54 NEFL not kncwn Di 27% of 60 RP patients with LN+ 8% Autoradiography loss of >50% Emmert-Buck MR Cancer Res 1995; 55: 2959 NEFL not known Di 72% of 29(homozygous deletion present) T2-T4N1M1 Autoradiography loss of >50% Macoska JA Cancer Res 1995; 55: 5390-5, NEFL not kncwn Di 63% of 56 (informative cases) T2-T3bN1M1 Autoradiography-near or complete loss by eye VockeC Cancer Res 1996; 56; 2411-16 NEFL not kncwn Di 35% of 29 Cancer, 21% of 29 HGPIN RP specimens Autoradiography-<50% densometry Haggman M. Urology 1997; 50: 643-47 NEFL not known Di 54% of 45 RP specimens Autoradiography-<50% by eye Prasad M Gen Chrom Can 1998; 23:255

234 T ype o f Primer G en e F requency Tissue Technique R eferen ce repeat

D10S211 DI 17% of 35 RP spedm ens T2-T3N1 Autoradiography loss of >50% Trybus T et al Cancer Res 1996; 56: 2263 D10S211 DI 2 cases of 64 ?hcw many informative T2-4.N0-1 Autotoradiography-loss >50% formally assesed Ittmann M Cancer Res 1996, 56: 2143-47 D10S587 close to MXI1 Di 0% of 17 primary (informative cases) RP and mets autoradiography-formally assessed Saric T Int J Cancer 1999; 81: 219-224 D10S587 close to MXI1 DI 19% of 37 T1M0-T4M1 Fluorescent allele typing > 20% loss Gray IC et al Cancer Res 1995; 55: 4800-03 D10S587 close to MXI1 DI 1 cases of 64 ?how many informative T2-4.N0-1 Autotoradiography-loss >50% formally assesed Ittmann M Cancer Res 1996; 56: 2143-47 D10S1246 close to MXI1 Tetra

D11S902 close to KAI1 Di no I oh in primary or mets RP and autopsy mets of CaP death Autotoradiography-loss >50% formally assesed Kawana Y Prostate 1997; 32: 205-13 D11S916 close to GSTP1 Tetra D11S2000

D12S1697 p27/kip1 Di 14% of 14 primary, 30% of 10 HRCAP mets Autopsy of CaP mets pts autoradiography - 50% loss formally Kawana Y Prostate 2002; 53: 60-64 D12S89 p27/kip1 Di 10% of 10 primary, 0% of 7 HRCAP mets Autopsy of CaP mets pts autoradiography - 50% loss formally Kawana Y Prostate 2002; 53: 60-64

D13S269 Di 27% of 41 TURP - 78% Gleason 8-10, all T4 autoradiography scanned <50% Dong JT Prostate 2001; 49: 166-171 D13S263 d o se to RB1 Di 32% of 41 TURP-78% Gleason 8-10, allT4 autoradiography scanned <50% Dong JT Prostate 2001; 49: 166-171 D13S171 intragenic to BRAC1 Di

D15S1232 Tetra

D16S413 close c-mye, H- and M-cadherin Di 17% of 18 primary RP and mets autoradiography-formally assessed Saric T Int J Cancer 1999; 81: 219-224

D18S541 close to DPC4 and DCC Tetra 22% of 47 RP AND TURP autoradiography - by eye Hugel A Br J Cancer 1999; 79: 551-57

D19S223 close to C-CAM Di 100% of 25 RP specimens, all grades immunohistchemistry Kleinerman DI Cancer Res 1995; 55: 1215

D21S156 not known Di 35% of 17 primary RP and mets autoradiography-formally assessed Saric T Int J Cancer 1999; 81: 219-224

235 References

Abate-Shen, C. & Shen, M. M. 2000, ’’Molecular genetics of prostate cancer”, Genes Dev., vol. 14, no. 19, pp. 2410-2434.

Abbou, C. C., Salomon, L., Hoznek, A., Antiphon, P., Cicco, A., Saint, F., Alame, W., Bellot, J., & Chopin, D. K. 2000, ’’Laparoscopic radical prostatectomy: preliminary results”, Urology, vol. 55, no. 5, pp. 630-634.

Abuzallouf, S., Dayes, 1., & Lukka, H. 2004, ’’Baseline staging of newly diagnosed prostate cancer: a summary of the literature”, J.Urol., vol. 171, no. 6 Pt 1, pp. 2122-2127.

Adolfsson, J. 1994, ’’Prognostic value of deoxyribonucleic acid content in prostate cancer: a review of current results", Int.J.Cancer , vol. 58, no. 2, pp. 211-216.

Agus, D. B., Scher, H. I., Higgins, B., Fox, W. D., Heller, G., Fazzari, M., Cordon-Cardo, C., & Golde, D. W. 1999, "Response of prostate cancer to anti-Her-2/neu antibody in androgen- dependent and -independent human xenograft models”, Cancer Res., vol. 59, no. 19, pp. 4761-4764.

Ahonen, M. H., Tenkanen, L., Teppo, L., Hakama, M., & Tuohimaa, P. 2000, "Prostate cancer risk and prediagnostic serum 25-hydroxyvitamin D levels (Finland)”, Cancer Causes Control, vol. 11, no. 9, pp. 847-852.

Albertsen, P. C., Fryback, D. G., Storer, B. E., Kolon, T. F., & Fine, J. 1995, "Long-term survival among men with conservatively treated localized prostate cancer”, JAMA, vol. 274, no. 8, pp. 626-631.

Albertsen, P. C., Fryback, D. G., Storer, B. E., Kolon, T. F., & Fine, J. 1996, "The impact of co-morbidity on life expectancy among men with localized prostate cancer”, J.Urol., vol. 156, no. 1, pp. 127-132.

Albertsen, P. C., Hanley, J. A., Gleason, D. F., & Barry, M. J. 1998, "Competing risk analysis of men aged 55 to 74 years at diagnosis managed conservatively for clinically localized prostate cancer”, JAMA, vol. 280, no. 11, pp. 975-980.

Alers, J. C., Rochat, J., Krijtenburg, P. J., Hop, W. C., Kranse, R., Rosenberg, C., Tanke, H. J., Schroder, F. H., & van Dekken, H. 2000, "Identification of genetic markers for prostatic cancer progression”, Lab Invest, vol. 80, no. 6, pp. 931-942.

Alizadeh, A. A., Eisen, M. B., Davis, R. E., Ma, C., Lossos, I. S., Rosenwald, A., Boldrick, J. C., Sabet, H., Tran, T., Yu, X., Powell, J. I., Yang, L., Marti, G. E., Moore, T., Hudson, J., Jr., Lu, L., Lewis, D. B., Tibshirani, R.,

236 Sherlock, G., Chan, W. C., Greiner, T. C., Weisenburger, D. D., Armitage, J. O., Warnke, R., & Staudt, L. M. 2000, ’’Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling", Nature , vol. 403, no. 6769, pp. 503-511.

American Urological Association (AUA) 2000, "Prostate-specific antigen (PSA) best practice policy.’’, Oncology (Huntingt), vol. 14, no. 2, pp. 267-8, 280.

Anderson, D. E. & Badzioch, M. D. 1992, "Breast cancer risks in relatives of male breast cancer patients", J.NatLCancer Inst, vol. 84, no. 14, pp. 1114- 1117.

Anderson, J. 2003, "The role of antiandrogen monotherapy in the treatment of prostate cancer", BJU.Int, vol. 91, no. 5, pp. 455-461.

Anscher, M. S., Clough, R., & Dodge, R. 2000, "Radiotherapy for a rising prostate-specific antigen after radical prostatectomy: the first 10 years", Int J RadiatOncolBioLPhys., vol. 48, no. 2, pp. 369-375.

Anscher, M. S. & Prosnitz, L. R. 1991, "Multivariate analysis of factors predicting local relapse after radical prostatectomy—possible indications for postoperative radiotherapy", Int.J.Radiat.Oncol.Biol.Phys., vol. 21, no. 4, pp. 941-947.

Aprikian, A. G., Sarkis, A. S., Fair, W. R., Zhang, Z. F., Fuks, Z., & Cordon- Cardo, C. 1994, "Immunohistochemical determination of p53 protein nuclear accumulation in prostatic adenocarcinoma", J.U rol, vol. 151, no. 5, pp. 1276- 1280.

Aus, G., Abbou, C. C., Heidenreich, A., Schmid, H.-P., Van Poppel, H., Wolff, J. M., & Zattoni, F. 2003, "Prostate Cancer", European Association of Urology Guidelines.

Ayala, A. G., Ro, J. Y., Babaian, R., Troncoso, P., & Grignon, D. J. 1989, "The prostatic capsule: does it exist? Its importance in the staging and treatment of prostatic carcinoma", Am. J.Surg.Pathol., vol. 13, no. 1, pp. 21- 27.

Babaian, R. J., Troncoso, P., Bhadkamkar, V. A., & Johnston, D. A. 2001, "Analysis of clinicopathologic factors predicting outcome after radical prostatectomy", Cancer , vol. 91, no. 8, pp. 1414-1422.

Badalament, R. A., Miller, M. C., Peller, P. A., Young, D. C., Bahn, D. K., Kochie, P., O'Dowd, G. J., & Veltri, R. W. 1996, "An algorithm for predicting nonorgan confined prostate cancer using the results obtained from sextant core biopsies with prostate specific antigen level", J UroL, vol. 156, no. 4, pp. 1375-1380.

Barocas, D. A., Han, M., Epstein, J. I., Chan, D. Y., Trock, B. J., Walsh, P. C., & Partin, A. W. 2001, "Does capsular incision at radical retropubic prostatectomy affect disease-free survival in otherwise organ-confined prostate cancer?", Urology, vol. 58, no. 5, pp. 746-751.

237 Barratt, P. L., Seymour, M. T., Stenning, S. P., Georgiades, I., Walker, C., Birbeck, K., & Quirke, P. 2002, "DNA markers predicting benefit from adjuvant fluorouracil in patients with colon cancer: a molecular study", Lancet , vol. 360, no. 9343, pp. 1381-1391.

Bastian, P. J., Palapattu, G. S., Lin, X., Yegnasubramanian, S., Mangold, L. A., Trock, B., Eisenberger, M. A., Partin, A. W., & Nelson, W. G. 2005, "Preoperative serum DNA GSTP1 CpG island hypermethylation and the risk of early prostate-specific antigen recurrence following radical prostatectomy", Clin.Cancer Res., vol. 11, no. 11, pp. 4037-4043.

Bates, T. S., Gillatt, D. A., Cavanagh, P. M., & Speakman, M. 1997, "A comparison of endorectal magnetic resonance imaging and transrectal ultrasonography in the local staging of prostate cancer with histopathological correlation", Br J Urol, vol. 79, no. 6, pp. 927-932.

Bedford, M. T. & van Helden, P. D. 1987, "Hypomethylation of DNA in pathological conditions of the human prostate", Cancer Res., vol. 47, no. 20, pp. 5274-5276.

Beer, D. G., Kardia, S. L., Huang, C. C., Giordano, T. J., Levin, A. M., Misek, D. E., Lin, L., Chen, G., Gharib, T. G., Thomas, D. G., Lizyness, M. L., Kuick, R., Hayasaka, S., Taylor, J. M., Iannettoni, M. D., Orringer, M. B., & Hanash, S. 2002, "Gene-expression profiles predict survival of patients with lung adenocarcinoma", N atM ed, vol. 8, no. 8, pp. 816-824.

Behrens, J., Mareel, M. M., Van Roy, F. M., & Birchmeier, W. 1989, "Dissecting tumor cell invasion: epithelial cells acquire invasive properties after the loss of uvomorulin-mediated cell-cell adhesion", J.Cell Biol., vol. 108, no. 6, pp. 2435-2447.

Bell, S. M., Scott, N., Cross, D., Sagar, P., Lewis, F. A., Blair, G. E., Taylor, G. R., Dixon, M. F., & Quirke, P. 1993, "Prognostic value of p53 overexpression and c-Ki-ras gene mutations in colorectal cancer", Gastroenterology, vol. 104, no. 1, pp. 57-64.

Bellido, T., O'Brien, C. A., Roberson, P. K., & Manolagas, S. C. 1998, "Transcriptional activation of the p21(WAFl,CIPl,SDIl) gene by interleukin-6 type cytokines. A prerequisite for their pro-differentiating and anti-apoptotic effects on human osteoblastic cells", / Biol.Client., vol. 273, no. 33, pp. 21137-21144.

Bennett, C. L., Tosteson, T. D., Schmitt, B., Weinberg, P. D., Ernstoff, M. S., & Ross, S. D. 1999, "Maximum androgen-blockade with medical or surgical castration in advanced prostate cancer: A meta-analysis of nine published randomized controlled trials and 4128 patients using flutamide", Prostate Cancer Prostatic.Dis., vol. 2, no. 1, pp. 4-8.

Bergerheim, U. S., Kunimi, K., Collins, V. P., & Ekman, P. 1991, "Deletion mapping of chromosomes 8,10, and 16 in human prostatic carcinoma", Genes Chromosomes.Cancer, vol. 3, no. 3, pp. 215-220.

238 Bernard, D., Pourtier-Manzanedo, A., Gil, J., & Beach, D. H. 2003, "Myc confers androgen-independent prostate cancer cell growth", J.ClinJnvest, vol. 112, no. 11, pp. 1724-1731.

Berthon, P., Valeri, A., Cohen-Akenine, A., Drelon, E., Paiss, T., Wohr, G., Latil, A., Millasseau, P., Mellah, L, Cohen, N., Blanche, H., Bellane- Chantelot, C., Demenais, F., Teillac, P., Le Due, A., de Petriconi, R., Hautmann, R., Chumakov, I., Bachner, L., Maitland, N. J., Lidereau, R., Vogel, W., Fournier, G., Mangin, P., & Cussenot, 0 . 1998, "Predisposing gene for early-onset prostate cancer, localized on chromosome lq42.2-43", Am. J.Hum.Genet., vol. 62, no. 6, pp. 1416-1424.

Bhatia-Gaur, R., Donjacour, A. A., Sciavolino, P. J., Kim, M., Desai, N., Young, P., Norton, C. R., Gridley, T., Cardiff, R. D., Cunha, G. R., Abate- Shen, C., & Shen, M. M. 1999, "Roles for Nkx3.1 in prostate development and cancer", Genes Dev., vol. 13, no. 8, pp. 966-977.

Bieche, I. & Lidereau, R. 1995, "Genetic alterations in breast cancer", Genes Chromosomes.Cancer, vol. 14, no. 4, pp. 227-251.

Billis A, Freitas LLL, & Magna LA 2002, "Vas deferens involvement in radical prostatectomy. Prevelence and significance.", United States and Canadian Academy of Pathology - Annual meeting.

Bittner, M., Meltzer, P., Chen, Y., Jiang, Y., Seftor, E., Hendrix, M., Radmacher, M., Simon, R., Yakhini, Z., Ben Dor, A., Sampas, N., Dougherty, E., Wang, E., Marincola, F., Gooden, C., Lueders, J., Glatfelter, A., Pollock, P., Carpten, J., Gillanders, E., Leja, D., Dietrich, K., Beaudry, C., Berens, M., Alberts, D., & Sondak, V. 2000, "Molecular classification of cutaneous malignant melanoma by gene expression profiling", Nature , vol. 406, no. 6795, pp. 536-540.

Blaveri, E., Simko, J. P., Korkola, J. E., Brewer, J. L., Baehner, F., Mehta, K., Devries, S., Koppie, T., Pejavar, S., Carroll, P., & Waldman, F. M. 2005, "Bladder cancer outcome and subtype classification by gene expression", Clin.Cancer Res., vol. 11, no. 11, pp. 4044-4055.

Blazer, D. G., III, Umbach, D. M., Bostick, R. M., & Taylor, J. A. 2000, "Vitamin D receptor polymorphisms and prostate cancer", Mol.Carcinog., vol. 27, no. 1, pp. 18-23.

Blute, M. L., Bostwick, D. G., Bergstralh, E. J., Slezak, J. M., Martin, S. K., Amling, C. L., & Zincke, H. 1997, "Anatomic site-specific positive margins in organ-confined prostate cancer and its impact on outcome after radical prostatectomy", Urology, vol. 50, no. 5, pp. 733-739.

Blute, M. L., Bostwick, D. G., Seay, T. M., Martin, S. K., Slezak, J. M., Bergstralh, E. J., & Zincke, H. 1998, "Pathologic classification of prostate carcinoma: the impact of margin status", Cancer , vol. 82, no. 5, pp. 902-908.

Bostwick, D. G., Grignon, D. J., Hammond, M. E., Amin, M. B., Cohen, M., Crawford, D., Gospadarowicz, M., Kaplan, R. S., Miller, D. S., Montironi, R., Pajak, T. F., Pollack, A., Srigley, J. R., & Yarbro, J. W. 2000, "Prognostic

239 factors in prostate cancer. College of American Pathologists Consensus Statement 1999", Arch.Pathol.Lab Med., vol. 124, no. 7, pp. 995-1000.

Bostwick, D. G., Shan, A., Qian, J., Darson, M., Maihle, N. J., Jenkins, R. B., & Cheng, L. 1998, "Independent origin of multiple foci of prostatic intraepithelial neoplasia: comparison with matched foci of prostate carcinoma", Cancer, vol. 83, no. 9, pp. 1995-2002.

Botchkina, G. I., Kim, R. H., Botchkina, I. L., Kirshenbaum, A., Frischer, Z., & Adler, H. L. 2005, "Noninvasive detection of prostate cancer by quantitative analysis of telomerase activity", Clin.Cancer Res., vol. 11, no. 9, pp. 3243-3249.

Bott SRJ, Thomas K, Foley C, Besarani D, & Kirby R 2002, "Cancer control, continence and potency after radical prostatectomy - a UK series", British Association of Urological Surgeons - annual meeting, abstract 117.

Bott, S. R. 2004, "Management of recurrent disease after radical prostatectomy", Prostate Cancer Prostatic Diseases, vol. 7, no. 3, pp. 211-216.

Bott, S. R., Arya, M., Kirby, R. S., & Williamson, M. 2005a, "The p21(WAFl/CIPl) gene is inactivated in metastatic prostatic cancer cell lines by promoter methylation", Prostate Cancer and Prostatic Diseases, vol. in press.

Bott, S. R., Arya, M., Shergill, I. S., & Williamson, M. 2005b, "Molecular changes in prostatic cancer", Surg.OncoL, vol. 14, no. 2, pp. 91-104.-

Bott, S. R., Freeman, A. A., Stenning, S., Cohen, J., & Parkinson, M. C. 2005c, "Radical prostatectomy: pathology findings in 1001 cases compared with other major series and over time", BJU.Int., vol. 95, no. 1, pp. 34-39.

Bott, S. R., Young, M. P., Kellett, M. J., & Parkinson, M. C. 2002, "Anterior prostate cancer: is it more difficult to diagnose?", BJU.Int, vol. 89, no. 9, pp. 886-889.

Bott, S. R. J., Freeman, A., Foley, C. L., Stenning, S., & Parkinson, M. C. 2003a, "Gleason grade at biopsy and radical prostatectomy: does it correlate?", British Association of Urological Surgeons - annual meeting, abstract 81.

Bott, S. R. J., Freeman, A., Parkinson, M. C., Kirby, R. S., Sohaib, S. A., & Husband, J. 2003b, "Does MRI using pelvic phased array coils better predict localised prostate cancer?", British Association of Urological Surgeons - annual meeting, abstract 17.

Bott, S. R. J. & Kirby, R. S. 2002, "Avoidance and management of positive surgical margins before, during and after radical prostatectomy", Prostate Cancer and Prostatic Diseases, vol. 5, no. 4, pp. 252-263.

Bott, S. R. J., Parkinson, M. C., Sydes, M. R., & Emberton, M. 2003c, "Predictive nonograms for prostate cancer - a UK series", British Association of Urological Surgeons - annual meeting, abstract 82.

240 Bott, S. R. J., Winstanley, A., Parkinson, M. C., & Kirby, R. S. 2003d, ’’The significance of a positive apical margin after radical prostatectomy”, British Association of Urological Surgeons - annual meeting, abstract 85.

Bowen, C., Spiegel, S., & Gelmann, E. P. 1998, ’’Radiation-induced apoptosis mediated by retinoblastoma protein”, Cancer Res., vol. 58, no. 15, pp. 3275- 3281.

Brawley, O. W. 1997, "Prostate carcinoma incidence and patient mortality: the effects of screening and early detection”, Cancer , vol. 80, no. 9, pp. 1857- 1863.

Brewster, S. F., Oxley, J. D., Trivella, M., Abbott, C. D., & Gillatt, D. A. 1999, ’’Preoperative p53, bcl-2, CD44 and E-cadherin immunohistochemistry as predictors of biochemical relapse after radical prostatectomy”, J Urol, vol. 161, no. 4, pp. 1238-1243.

Brooks, J. D., Bova, G. S., Ewing, C. M., Piantadosi, S., Carter, B. S., Robinson, J. C., Epstein, J. I., & Isaacs, W. B. 1996, "An uncertain role for p53 gene alterations in human prostate cancers”, Cancer Res., vol. 56, no. 16, pp. 3814-3822.

Brooks, J. D., Weinstein, M., Lin, X., Sun, Y., Pin, S. S., Bova, G. S., Epstein, J. I., Isaacs, W. B., & Nelson, W. G. 1998, ”CG island methylation changes near the GSTP1 gene in prostatic intraepithelial neoplasia”, Cancer Epidemiol.Biomarkers Prev., vol. 7, no. 6, pp. 531-536.

Brothman, A. R., Maxwell, T. M., Cui, J., Deubler, D. A., & Zhu, X. L. 1999, "Chromosomal clues to the development of prostate tumors”, Prostate, vol. 38, no. 4, pp. 303-312.

Bubendorf, L., Kononen, J., Koivisto, P., Schraml, P., Moch, H., Gasser, T. C., Willi, N., Mihatsch, M. J., Sauter, G., & Kallioniemi, O. P. 1999, "Survey of gene amplifications during prostate cancer progression by high- throughout fluorescence in situ hybridization on tissue microarrays”, Cancer Res., vol. 59, no. 4, pp. 803-806.

Bucher, P. 1990, "Weight matrix descriptions of four eukaryotic RNA polymerase II promoter elements derived from 502 unrelated promoter sequences”,J Mol Biol., vol. 212, no. 4, pp. 563-578.

Byar, D. P. & Mostofi, F. K. 1972, "Carcinoma of the prostate: prognostic evaluation of certain pathologic features in 208 radical prostatectomies. Examined by the step-section technique”, Cancer, vol. 30, no. 1, pp. 5-13.

Cadeddu, J. A., Partin, A. W., DeWeese, T. L., & Walsh, P. C. 1998, "Long­ term results of for prostate cancer recurrence following radical prostatectomy”, J.Urol., vol. 159, no. 1, pp. 173-177.

Cairns, P., Esteller, M., Herman, J. G., Schoenberg, M., Jeronimo, C., Sanchez-Cespedes, M., Chow, N. H., Grasso, M., Wu, L., Westra, W. B., & Sidransky, D. 2001, "Molecular detection of prostate cancer in urine by GSTP1 hyper methylation", Clin.Cancer Res., vol. 7, no. 9, pp. 2727-2730.

241 Cairns, P., Okami, K., Halachmi, S., Halachmi, N., Esteller, M., Herman, J. G., Jen, J., Isaacs, W. B., Bova, G. S., & Sidransky, D. 1997, "Frequent inactivation of PTEN/MMAC1 in primary prostate cancer", Cancer Res., vol. 57, no. 22, pp. 4997-5000.

Califano, J., Ahrendt, S. A., Meininger, G., Westra, W. H., Koch, W. M., & Sidransky, D. 1996, "Detection of telomerase activity in oral rinses from head and neck squamous cell carcinoma patients", Cancer Res., vol. 56, no. 24, pp. 5720-5722.

Canizares, F., Sola, J., Perez, M., Tovar, I., De Las, H. M., Salinas, J., Penafiel, R., & Martinez, P. 2001, "Preoperative values of CA 15-3 and CEA as prognostic factors in breast cancer: a multivariate analysis", Tumour.Biol. pp. 273-281.

Canzian, F., Salovaara, R., Hemminki, A., Kristo, P., Chadwick, R. B., Aaltonen, L. A., & de la, C. A. 1996, "Semiautomated assessment of loss of heterozygosity and replication error in tumors", Cancer Res., vol. 56, no. 14, pp. 3331-3337.

Carlson, G. D., Calvanese, C. B., Kahane, H., & Epstein, J. 1 .1998, "Accuracy of biopsy Gleason scores from a large uropathology laboratory: use of a diagnostic protocol to minimize observer variability", Urology, vol. 51, no. 4, pp. 525-529.

Carter, B. S., Bova, G. S., Beaty, T. H., Steinberg, G. D., Childs, B., Isaacs, W. B., & Walsh, P. C. 1993, "Hereditary prostate cancer: epidemiologic and clinical features", J.Urol., vol. 150, no. 3, pp. 797-802.

Carter, B. S., Ewing, C. M., Ward, W. S., Treiger, B. F., Aalders, T. W., Schalken, J. A., Epstein, J. I., & Isaacs, W. B. 1990, "Allelic loss of chromosomes 16q and lOq in human prostate cancer", Proc.Natl*Acad.ScuU.S.A, vol. 87, no. 22, pp. 8751-8755.

Carter, H. B., Epstein, J. I., & Partin, A. W. 1999, "Influence of age and prostate-specific antigen on the chance of curable prostate cancer among men with nonpalpable disease", Urology, vol. 53, no. 1, pp. 126-130.

Catalona, W. J. & Bigg, S. W. 1990, "Nerve-sparing radical prostatectomy: evaluation of results after 250 patients", J.Urol., vol. 143, no. 3, pp. 538-543.

Catalona, W. J., Ramos, C. G., & Carvalhal, G. F. 1999, "Contemporary results of anatomic radical prostatectomy", CA Cancer J Clin., vol. 49, no. 5, pp. 282-296.

Catalona, W. J. & Smith, D. S. 1998, "Cancer recurrence and survival rates after anatomic radical retropubic prostatectomy for prostate cancer: intermediate-term results", J Urol., vol. 160, no. 6 Pt 2, pp. 2428-2434.

Cathelineau, X., Cahill, D., Widmer, H., Rozet, F., Baumert, H., & Vallancien, G. 2004, "Transperitoneal or extraperitoneal approach for laparoscopic radical prostatectomy: a false debate over a real challenge", J.Urol, vol. 171, no. 2 Pt 1, pp. 714-716.

242 Chamberlain, N. L., Driver, E. D., & Miesfeld, R. L. 1994, "The length and location of CAG trinucleotide repeats in the androgen receptor N-terminal domain affect transactivation function", Nucleic Acids Res., vol. 22, no. 15, pp. 3181-3186.

Chang, M., Tsuchiya, K., Batchelor, R. H., Rabinovitch, P. S., Kulander, B. G., Haggitt, R. C., & Burmer, G. C. 1994, "Deletion mapping of chromosome 8p in colorectal carcinoma and arising in ulcerative colitis, prostatic carcinoma, and malignant fibrous histiocytomas", Am. J Pathol., vol. 144, no. 1, pp. 1-6.

Chen, B., He, L., Saveli, V. H., Jenkins, J. J., & Parham, D. M. 2000, "Inhibition of the interferon-gamma/signal transducers and activators of transcription (STAT) pathway by hypermethylation at a STAT-binding site in the p21WAFl promoter region", Cancer Res., vol. 60, no. 12, pp. 3290- 3298.

Cheng, L., Darson, M. F., Bergstralh, E. J., Slezak, J., Myers, R. P., & Bostwick, D. G. 1999, "Correlation of margin status and extraprostatic extension with progression of prostate carcinoma", Cancer, vol. 86, no. 9, pp. 1775-1782.

Cheng, L., Lloyd, R. V., Weaver, A. L., Pisansky, T. M., Cheville, J. C., Ramnani, D. M., Leibovich, B. C., Blute, M. L., Zincke, H., & Bostwick, D. G. 2000, "The cell cycle inhibitors p21WAFl and p27KIPl are associated with survival in patients treated by salvage prostatectomy after radiation therapy", Clin.Cancer Res., vol. 6, no. 5, pp. 1896-1899.

Cher, M. L., Bianco, F. J., Jr., Lam, J. S., Davis, L. P., Grignon, D. J., Sakr, W. A., Banerjee, M., Pontes, J. E., & Wood, D. P., Jr. 1998a, "Limited role of radionuclide bone scintigraphy in patients with prostate specific antigen elevations after radical prostatectomy", J Urol., vol. 160, no. 4, pp. 1387- 1391.

Cher, M. L., Bova, G. S., Moore, D. H., Small, E. J., Carroll, P. R., Pin, S. S., Epstein, J. I., Isaacs, W. B., & Jensen, R. H. 1996, "Genetic alterations in untreated metastases and androgen-independent prostate cancer detected by comparative genomic hybridization and allelotyping", Cancer Res., vol. 56, no. 13, pp. 3091-3102.

Cher, M. L., Lewis, P. E., Banerjee, M., Hurley, P. M., Sakr, W., Grignon, D. J., & Powell, I. J. 1998b, "A similar pattern of chromosomal alterations in prostate cancers from African-Americans and Caucasian Americans", Clin.Cancer Res., vol. 4, no. 5, pp. 1273-1278.

Chesire, D. R., Ewing, C. M., Sauvageot, J., Bova, G. S., & Isaacs, W. B. 2000, "Detection and analysis of beta-catenin mutations in prostate cancer", Prostate, vol. 45, no. 4, pp. 323-334.

Chin, Y. E., Kitagawa, M., Kuida, K., Flavell, R. A., & Fu, X. Y. 1997, "Activation of the ST AT signaling pathway can cause expression of caspase 1 and apoptosis", Mol Cell BioL pp. 5328-5337.

243 Chodak, G. W., Thisted, R. A., Gerber, G. S., Johansson, J. E., Adolfsson, J., Jones, G. W., Chisholm, G. D., Moskovitz, B., Livne, P. M., & Warner, J. 1994, "Results of conservative management of clinically localized prostate cancer", N.Engl. J.Med., vol. 330, no. 4, pp. 242-248.

Coetzee, L. J., Hars, V., & Paulson, D. F. 1996, "Postoperative prostate- specific antigen as a prognostic indicator in patients with margin-positive prostate cancer, undergoing adjuvant radiotherapy after radical prostatectomy", Urology, vol. 47, no. 2, pp. 232-235.

Connolly, J. A., Shinohara, K., Presti, J. C., Jr., & Carroll, P. R. 1996, "Local recurrence after radical prostatectomy: characteristics in size, location, and relationship to prostate-specific antigen and surgical margins", Urology, vol. 47, no. 2, pp. 225-231.

Cooney, K. A., McCarthy, J. D., Lange, E., Huang, L., Miesfeldt, S., Montie, J. E., Oesterling, J. E., Sandler, H. M., & Lange, K. 1997, "Prostate cancer susceptibility locus on chromosome lq: a confirmatory study", J.NatLCancer Inst., vol. 89, no. 13, pp. 955-959.

Cooney, K. A., Wetzel, J. C., Consolino, C. M., & Wojno, K. J. 1996a, "Identification and characterization of proximal 6q deletions in prostate cancer", Cancer Res. , vol. 56, no. 18, pp. 4150-4153.

Cooney, K. A., Wetzel, J. C., Merajver, S. D., Macoska, J. A., Singleton, T. P., & Wojno, K. J. 1996b, "Distinct regions of allelic loss on 13q in prostate cancer", Cancer Res., vol. 56, no. 5, pp. 1142-1145.

Corder, E. H., Guess, H. A., Hulka, B. S., Friedman, G. D., Sadler, M., Vollmer, R. T., Lobaugh, B., Drezner, M. K., Vogelman, J. H., & Orentreich, N. 1993, "Vitamin D and prostate cancer: a prediagnostic study with stored sera", Cancer EpidemioLBiomarkers Prev., vol. 2, no. 5, pp. 467-472.

Cornud, F., Belin, X., Fromont, G., Chretien, Y., Helenon, O., Boisrond, L., Casanova, J. M., Meuriot, T., Moreau, J. F., & Dufour, B. 1992, "(What may be expected from endorectal echography and magnetic resonance imaging in the evaluation of local extension of cancer of the prostate?]", J Radiol., vol. 73, no. 11, pp. 617-627.

Costello, L. C., Franklin, R. B., & Narayan, P. 1999, "Citrate in the diagnosis of prostate cancer", Prostate, vol. 38, no. 3, pp. 237-245.

Cote, R. J., Shi, Y., Groshen, S., Feng, A. C., Cordon-Cardo, C., Skinner, D., & Lieskovosky, G. 1998, "Association of p27Kipl levels with recurrence and survival in patients with stage C prostate carcinoma", JNatl.Cancer Inst. pp. 916-920.

Cox, J. D., Gallagher, M. J., Hammond, E. H., Kaplan, R. S., & Schellhammer, P. F. 1999, "Consensus statements on radiation therapy of prostate cancer: guidelines for prostate re-biopsy after radiation and for radiation therapy with rising prostate-specific antigen levels after radical prostatectomy. American Society for Therapeutic Radiology and Oncology Consensus Panel", J.Clin.Oncol., vol. 17, no. 4, p. 1155.

244 Craft, N., Shostak, Y., Carey, M., & Sawyers, C. L. 1999, ”A mechanism for hormone-independent prostate cancer through modulation of androgen receptor signaling by the HER-2/neu tyrosine kinase", NatM ed., vol. 5, no. 3, pp. 280-285.

Cunningham, J. M., Shan, A., Wick, M. J., McDonnell, S. K., Schaid, D. J., Tester, D. J., Qian, J., Takahashi, S., Jenkins, R. B., Bostwick, D. G., & Thibodeau, S. N. 1996, "Allelic imbalance and microsatellite instability in prostatic adenocarcinoma", Cancer Res., vol. 56, no. 19, pp. 4475-4482.

D'Amico, A. V., Chen, M. H., Roehl, K. A., & Catalona, W. J. 2004a, "Preoperative PSA velocity and the risk of death from prostate cancer after radical prostatectomy", N.Engl.J.Med., vol. 351, no. 2, pp. 125-135.

D'Amico, A. V., Moul, J., Carroll, P. R., Sun, L., Lubeck, D., & Chen, M. H. 2004b, "Prostate specific antigen doubling time as a surrogate end point for prostate cancer specific mortality following radical prostatectomy or radiation therapy", J.Urol., vol. 172, no. 5 Pt 2, p. S42-S46.

D'Amico, A. V., Whittington, R., Malkowicz, B., Schnall, M., Schultz, D., Cote, K., Tomaszewski, J. E., & Wein, A. 2000a, "Endorectal magnetic resonance imaging as a predictor of biochemical outcome after radical prostatectomy in men with clinically localized prostate cancer", J.Urol., vol. 164, no. 3 Pt 1, pp. 759-763.

D'Amico, A. V., Whittington, R., Malkowicz, S. B., Schultz, D., Fondurulia, J., Chen, M. H., Tomaszewski, J. E., Renshaw, A. A., Wein, A., & Richie, J. P. 2000b, "Clinical utility of the percentage of positive prostate biopsies in defining biochemical outcome after radical prostatectomy for patients with clinically localized prostate cancer", J Clin.Oncol., vol. 18, no. 6, pp. 1164- 1172.

D'Amico, A. V., Whittington, R., Malkowicz, S. B., Schultz, D., Schnall, M., Tomaszewski, J. E., & Wein, A. 1995, "A multivariate analysis of clinical and pathological factors that predict for prostate specific antigen failure after radical prostatectomy for prostate cancer", J Urol., vol. 154, no. 1, pp. 131- 138.

Dahiya, R., Lee, C., McCarville, J., Hu, W., Kaur, G., & Deng, G. 1997a, "High frequency of genetic instability of microsatellites in human prostatic adenocarcinoma", Int J Cancer, vol. 72, no. 5, pp. 762-767.

Dahiya, R., McCarville, J., Hu, W., Lee, C., Chui, R. M., Kaur, G., & Deng, G. 1997b, "Chromosome 3p24-26 and 3p22-12 loss in human prostatic adenocarcinoma", Int.J.Cancer, vol. 71, no. 1, pp. 20-25.

Dahiya, R., McCarville, J., Lee, C., Hu, W., Kaur, G., Carroll, P., & Deng, G. 1997c, "Deletion of chromosome llp l5 , p i2, q22, q23-24 loci in human prostate cancer", Int. J.Cancer, vol. 72, no. 2, pp. 283-288.

Darnell, J. E., Jr. 1997, "STATs and gene regulation", Science, vol. 277, no. 5332, pp. 1630-1635.

245 Datto, M. B., Li, Y., Panus, J. F., Howe, D. J., Xiong, Y., & Wang, X. F. 1995, "Transforming growth factor beta induces the cyclin-dependent kinase inhibitor p21 through a p53-independent mechanism", Proc.Natl.Acad.ScL U.S.A, vol. 92, no. 12, pp. 5545-5549.

Deasy, N. P., Conry, B. G., Lewis, J. L., Ford, T. F., Russell, G. A., Basu, R., & Flanagan, J. J. 1997, "Local staging of prostate cancer with 0.2 T body coil MRI", Clin.Radiol., vol. 52, no. 12, pp. 933-937.

DeGrado, T. R., Coleman, R. E., Wang, S., Baldwin, S. W., Orr, M. D., Robertson, C. N., Polascik, T. J., & Price, D. T. 2001, "Synthesis and evaluation of 18F-labeled choline as an oncologic tracer for positron emission tomography: initial findings in prostate cancer", Cancer Res., vol. 61, no. 1, pp. 110-117.

Di Cristofano, A., De Acetis, M., Koff, A., Cordon-Cardo, C., & Pandolfi, P. P. 2001, "Pten and p27KIPl cooperate in prostate cancer tumor suppression in the mouse", NatGeneL , vol. 27, no. 2, pp. 222-224.

Di Cristofano, A. & Pandolfi, P. P. 2000, "The multiple roles of PTEN in tumor suppression", Cell, vol. 100, no. 4, pp. 387-390.

Diblasio, C. J. & Kattan, M. W. 2003, "Use of nomograms to predict the risk of disease recurrence after definitive local therapy for prostate cancer", Urology, vol. 62, no. 6 Suppl 2, pp. 9-18.

Doherty, A. P., Bower, M., Smith, G. L., Miano, R., Mannion, E. M., Mitchell, H., & Christmas, T. J. 2000, "Undetectable ultrasensitive PSA after radical prostatectomy for prostate cancer predicts relapse-free survival", Br J Cancer , vol. 83, no. 11, pp. 1432-1436.

Dong, J. T., Boyd, J. C., & Frierson, H. F., Jr. 2001, "Loss of heterozygosity at 13ql4 and 13q21 in high grade, high stage prostate cancer", Prostate, vol. 49, no. 3, pp. 166-171.

Dong, J. T., Chen, C., Stultz, B. G., Isaacs, J. T., & Frierson, H. F., Jr. 2000, "Deletion at 13q21 is associated with aggressive prostate cancers", Cancer Res., vol. 60, no. 14, pp. 3880-3883.

Dong, J. T., Lamb, P. W., Rinker-Schaeffer, C. W., Vukanovic, J., Ichikawa, T., Isaacs, J. T., & Barrett, J. C. 1995, "KAI1, a metastasis suppressor gene for prostate cancer on human chromosome llp ll.2" , Science, vol. 268, no. 5212, pp. 884-886.

Dong, J. T., Sipe, T. W., Hyytinen, E. R., Li, C. L., Heise, C., McClintock, D. E., Grant, C. D., Chung, L. W., & Frierson, H. F., Jr. 1998, "PTEN/MMAC1 is infrequently mutated in pT2 and pT3 carcinomas of the prostate", Oncogene, vol. 17, no. 15, pp. 1979-1982.

Dong, Y., Walsh, M. D., McGuckin, M. A., Gabrielli, B. G., Cummings, M. C., Wright, R. G., Hurst, T., Khoo, S. K., & Parsons, P. G. 1997, "Increased expression of cyclin-dependent kinase inhibitor 2 (CDKN2A) gene product P16INK4A in ovarian cancer is associated with progression and unfavourable prognosis", Int J Cancer, vol. 74, no. 1, pp. 57-63.

246 Dotto, G. P. 2000, "p21(WAFl/Cipl): more than a break to the cell cycle?”, Biochim.Biophys.Acta, vol. 1471, no. 1, p. M43-M56.

Drobnjak, M., Osman, I., Scher, H. I., Fazzari, M., & Cordon-Cardo, C. 2000, "Overexpression of cyclin DI is associated with metastatic prostate cancer to bone”, Clin.Cancer Res., vol. 6, no. 5, pp. 1891-1895.

Dunsmuir, W. D., Edwards, S. M., Lakhani, S. R., Young, M., Corbishley, C., Kirby, R. S., Dearnaley, D. P., Dowe, A., Ardern-Jones, A., Kelly, J., & Eeles, R. A. 1998, "Allelic imbalance in familial and sporadic prostate cancer at the putative human prostate cancer susceptibility locus, HPCl. CRC/BPG UK Familial Prostate Cancer Study Collaborators. Cancer Research Campaign/British Prostate Group”, Br.J.Cancer , vol. 78, no. 11, pp. 1430- 1433.

Eagle, L. R., Yin, X., Brothman, A. R., Williams, B. J., Atkin, N. B., & Prochownik, E. V. 1995, "Mutation of the MXI1 gene in prostate cancer", Nat.Genet, vol. 9, no. 3, pp. 249-255.

Edwards, J., Krishna, N. S., Grigor, K. M., & Bartlett, J. M. 2003, "Androgen receptor gene amplification and protein expression in hormone refractory prostate cancer", Br.J.Cancer , vol. 89, no. 3, pp. 552-556.

Eeles, R. A., Durocher, F., Edwards, S., Teare, D., Badzioch, M., Hamoudi, R., Gill, S., Biggs, P., Dearnaley, D., Ardern-Jones, A., Dowe, A., Shearer, R., McLennan, D. L., Norman, R. L., Ghadirian, P., Aprikian, A., Ford, D., Amos, C., King, T. M., Labrie, F., Simard, J., Narod, S. A., Easton, D., & Foulkes, W. D. 1998, "Linkage analysis of chromosome lq markers in 136 prostate cancer families. The Cancer Research Campaign/British Prostate Group U.K. Familial Prostate Cancer Study Collaborators", Am.J.Hum.Genet., vol. 62, no. 3, pp. 653-658.

Eeles, R. A. & the UK Familial Prostate Study Group & the CRC/BPG UK Familial Prostate Cancer Study Collaborators 1999, "Genetic predisposition to prostate cancer", Prostate cancer and prostatic disease , vol. 2, pp. 9-15.

el Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin, D., Mercer, W. E., Kinzler, K. W., & Vogelstein, B. 1993, "WAF1, a potential mediator of p53 tumor suppression", Cell, vol. 75, no. 4, pp. 817- 825.

Ellis, W. J., Vessella, R. L., Corey, E., Arfman, E. W., Oswin, M. M., Melchior, S., & Lange, P. H. 1998, "The value of a reverse transcriptase polymerase chain reaction assay in preoperative staging and followup of patients with prostate cancer", J.Urol., vol. 159, no. 4, pp. 1134-1138.

Elo, J. P., Harkonen, P., Kyllonen, A. P., Lukkarinen, O., & Vihko, P. 1999, "Three independently deleted regions at chromosome arm 16q in human prostate cancer: allelic loss at 16q24.1-q24.2 is associated with aggressive behaviour of the disease, recurrent growth, poor differentiation of the tumour and poor prognosis for the patient", Br.J.Cancer , vol. 79, no. 1, pp. 156-160.

247 Elo, J. P. & Visakorpi, T. 2001, "Molecular genetics of prostate cancer", Ann.Med., vol. 33, no. 2, pp. 130-141.

Emmert-Buck, M. R., Vocke, C. D., Pozzatti, R. O., Duray, P. H., Jennings, S. B., Florence, C. D., Zhuang, Z., Bostwick, D. G., Liotta, L. A., & Linehan, W. M. 1995, "Allelic loss on chromosome 8pl2-21 in microdissected prostatic intraepithelial neoplasia", Cancer Res. , vol. 55, no. 14, pp. 2959-2962.

Engelbrecht, M. R., Jager, G. J., Laheij, R. J., Verbeek, A. L., van Lier, H. J., & Barentsz, J. O. 2002, "Local staging of prostate cancer using magnetic resonance imaging: a meta-analysis", Eur.Radiol., vol. 12, no. 9, pp. 2294- 2302.

Epstein, J. 1 .1990, "Evaluation of radical prostatectomy capsular margins of resection. The significance of margins designated as negative, closely approaching, and positive", Am.J Surg.Pathol., vol. 14, no. 7, pp. 626-632.

Epstein, J. I., Carmichael, M., Partin, A. W., & Walsh, P. C. 1993, "Is tumor volume an independent predictor of progression following radical prostatectomy? A multivariate analysis of 185 clinical stage B adenocarcinomas of the prostate with 5 years of followup", J.Urol., vol. 149, no. 6, pp. 1478-1481.

Epstein, J. I., Partin, A. W., Sauvageot, J., & Walsh, P. C. 1996, "Prediction of progression following radical prostatectomy. A multivariate analysis of 721 men with long-term follow-up", AnuJ.Surg.PathoL, vol. 20, no. 3, pp. 286- 292.

Epstein, J. I., Pizov, G., & Walsh, P. C. 1993, "Correlation of pathologic findings with progression after radical retropubic prostatectomy", Cancer, vol. 71, no. 11, pp. 3582-3593.

Epstein, J. I., Pound, C. R., Partin, A. W., & Walsh, P. C. 1998, "Disease progression following radical prostatectomy in men with Gleason score 7 tumor", J.Urol., vol. 160, no. 1, pp. 97-100.

Epstein, J. I. & Sauvageot, J. 1997, "Do close but negative margins in radical prostatectomy specimens increase the risk of postoperative progression?", J Urol, vol. 157, no. 1, pp. 241-243.

Epstein, J. I., Walsh, P. C., & Brendler, C. B. 1994, "Radical prostatectomy for impalpable prostate cancer: the Johns Hopkins experience with tumors found on transurethral resection (stages T1A and TIB) and on needle biopsy (stage TIC)", J.Urol., vol. 152, no. 5 Pt 2, pp. 1721-1729.

Epstein, J. I., Walsh, P. C., Carmichael, M., & Brendler, C. B. 1994, "Pathologic and clinical findings to predict tumor extent of nonpalpable (stage Tic) prostate cancer", JAMA, vol. 271, no. 5, pp. 368-374.

Erbersdobler, A., Fritz, H., Schnoger, S., Graefen, M., Hammerer, P., Huland, H., & Henke, R. P. 2002a, "Tumour grade, proliferation, apoptosis, microvessel density, p53, and bcl-2 in prostate cancers: differences between tumours located in the transition zone and in the peripheral zone", Eur.Urol., vol. 41, no. 1, pp. 40-46.

248 Erbersdobler, A., Huhle, S., Palisaar, J., Graefen, M., Hammerer, P., Noldus, J., & Huland, H. 2002b, "Pathological and clinical characteristics of large prostate cancers predominantly located in the transition zone", Prostate Cancer Prostatic Dis ., vol. 5, no. 4, pp. 279-284.

Eschrich, S., Yang, I., Bloom, G., Kwong, K. Y., Boulware, D., Cantor, A., Coppola, D., Kruhoffer, M., Aaltonen, L., Orntoft, T. F., Quackenbush, J., & Yeatman, T. J. 2005, "Molecular staging for survival prediction of colorectal cancer patients", J.Clin.Oncol., vol. 23, no. 15, pp. 3526-3535.

Faith, D., Han, S., Lee, D. K., Friedl, A., Hicks, J. L., De Marzo, A. M., & Jarrard, D. F. 2005, "pl6 Is upregulated in proliferative inflammatory atrophy of the prostate", Prostate.

Fearon, E. R. & Vogelstein, B. 1990, "A genetic model for colorectal tumorigenesis", Cell, vol. 61, no. 5, pp. 759-767.

Feigl, P., Blumenstein, B., Thompson, I., Crowley, J., Wolf, M., Kramer, B. S., Coltman, C. A., Jr., Brawley, O. W., & Ford, L. G. 1995, "Design of the Prostate Cancer Prevention Trial (PCPT)", Control Clin.Trials, vol. 16, no. 3, pp. 150-163.

Feilotter, H. E., Nagai, M. A., Boag, A. H., Eng, C., & Mulligan, L. M. 1998, "Analysis of PTEN and the 10q23 region in primary prostate carcinomas", Oncogene, vol. 16, no. 13, pp. 1743-1748.

Fidler, I. J. 1990, "Critical factors in the biology of human cancer metastasis: twenty- eighth G.H.A. Clowes memorial award lecture", Cancer Res., vol. 50, no. 19, pp. 6130-6138.

Foley, C. L., Bott, S. R. J., Parkinson, M. C., & Kirby, R. S. 2002, "The role of anastomotic biopsy in the evaluation of patients with a rising PSA after radical prostatectomy.", British Association of Urological Surgeons Annual meeting.

Fowler, F. J., Jr., Barry, M. J., Lu-Yao, G., Roman, A., Wasson, J., & Wennberg, J. E. 1993, "Patient-reported complications and follow-up treatment after radical prostatectomy. The National Medicare Experience: 1988-1990 (updated June 1993)", Urology, vol. 42, no. 6, pp. 622-629.

Franks, L. M. 1956, "The natural history of prostatic cancer", Lancet , vol. 2, pp. 1037-1039.

Freedland, S. J., Aronson, W. J., Kane, C. J., Terris, M. K., Presti, J. C., Jr., Trock, B., & Amling, C. L. 2004, "Biochemical outcome after radical prostatectomy among men with normal preoperative serum prostate-specific antigen levels", Cancer , vol. 101, no. 4, pp. 748-753.

Freedland, S. J., Presti, J. C., Jr., Amling, C. L., Kane, C. J., Aronson, W. J., Dorey, F., & Terris, M. K. 2003, "Time trends in biochemical recurrence after radical prostatectomy: results of the SEARCH database", Urology, vol. 61, no. 4, pp. 736-741.

249 Fromont, G., Joulin, V., Chantrel-Groussard, K., Vallancien, G., Guillonneau, B., Valid ire, P., Latil, A., & Cussenot, 0 . 2003, "Allelic losses in localized prostate cancer: association with prognostic factors", J.Urol., vol. 170, no. 4 Pt 1, pp. 1394-1397.

Gaboli, M., Kotsi, P. A., Gurrieri, C., Cattoretti, G., Ronchetti, S., Cordon- Cardo, C., Broxmeyer, H. E., Hromas, R., & Pandolfi, P. P. 2001, "Mzfl controls cell proliferation and tumorigenesis", Genes Dev., vol. 15, no. 13, pp. 1625-1630.

Gann, P. H., Hennekens, C. H., & Stampfer, M. J. 1995, "A prospective evaluation of plasma prostate-specific antigen for detection of prostatic cancer", JAMA, vol. 273, no. 4, pp. 289-294.

Gao, A. C., Lou, W., Dong, J. T., & Isaacs, J. T. 1997, "CD44 is a metastasis suppressor gene for prostatic cancer located on human chromosome llpl3", Cancer Res., vol. 57, no. 5, pp. 846-849.

Gao, H., Ouyang, X., Banach-Petrosky, W., Borowsky, A. D., Lin, Y., Kim, M., Lee, H., Shih, W. J., Cardiff, R. D., Shen, M. M., & Abate-Shen, C. 2004, "A critical role for p27kipl gene dosage in a mouse model of prostate carcinogenesis", Proc.NatLAcad.ScLU.S.A, vol. 101, no. 49, pp. 17204-17209.

Gao, X., Porter, A. T., & Honn, K. V. 1997, "Involvement of the multiple tumor suppressor genes and 12-lipoxygenase in human prostate cancer. Therapeutic implications", Adv.Exp.Med.BioL, vol. 407, pp. 41-53.

Gao, X., Zacharek, A., Grignon, D. J., Sakr, W., Powell, I. J., Porter, A. T., & Honn, K. V. 1995, "Localization of potential tumor suppressor loci to a < 2 Mb region on chromosome 17q in human prostate cancer", Oncogene, vol. 11, no. 7, pp. 1241-1247.

Gartel, A. L. & Tyner, A. L. 1999, "Transcriptional regulation of the P21((WAF1/CIP1)) gene", Exp Cell Res., vol. 246, no. 2, pp. 280-289.

Gerber, G. S., Thisted, R. A., Scardino, P. T., Frohmuller, H. G., Schroeder, F. H., Paulson, D. F., Middleton, A. W., Jr., Rukstalis, D. B., Smith, J. A., Jr., Schellhammer, P. F., Ohori, M., & Chodak, G. W. 1996, "Results of radical prostatectomy in men with clinically localized prostate cancer", JAMA, vol. 276, no. 8, pp. 615-619.

Gervasi, L. A., Mata, J., Easley, J. D., Wilbanks, J. H., Seale-Hawkins, C., Carlton, C. E., Jr., & Scardino, P. T. 1989, "Prognostic significance of lymph nodal metastases in prostate cancer", J Urol., vol. 142, no. 2 Pt 1, pp. 332-336.

Giovannucci, E., Stampfer, M. J., Krithivas, K., Brown, M., Dahl, D., Brufsky, A., Talcott, J., Hennekens, C. H., & Kantoff, P. W. 1997, "The CAG repeat within the androgen receptor gene and its relationship to prostate cancer", Proc.NatLAcad.ScLU.S.A, vol. 94, no. 7, pp. 3320-3323.

Gleason, D. F. 1992, "Histologic grading of prostate cancer: a perspective", HunuPathoL, vol. 23, no. 3, pp. 273-279.

250 Gleason, D. F. & Mellinger, G. T. 1974, ’’Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging”, J.UroL, vol. I ll, no. 1, pp. 58-64.

Gleave, M., Tolcher, A., Miyake, H., Nelson, C., Brown, B., Beraldi, E., & Goldie, J. 1999, ’’Progression to androgen independence is delayed by adjuvant treatment with antisense Bcl-2 oligodeoxynucleotides after castration in the LNCaP prostate tumor model”, Clin.Cancer Res., vol. 5, no. 10, pp. 2891-2898.

Goessl, C., Muller, M., Heicappell, R., Krause, H., Straub, B., Schrader, M., & Miller, K. 2001, ”DNA-based detection of prostate cancer in urine after prostatic massage”, Urology, vol. 58, no. 3, pp. 335-338.

Gold, P. & Freedman, S. 0 . 1975, ’’Tests for carcinoembryonic antigen. Role in diagnosis and management of cancer”, JAMA, vol. 234, no. 2, pp. 190-192.

Goldenberg, S. & et al 1997, ”A randomised trial of neoadjuvant androgen withdrawl therapy prior to radical prostatectomy: 24 month post-treatment PSA results", J Urol, vol. 157 (part2), p. 92.

Goldenberg, S. L., Klotz, L. H., Srigley, J., Jewett, M. A., Mador, D., Fradet, Y., Barkin, J., Chin, J., Paquin, J. M., Bullock, M. J., & Laplante, S. 1996, "Randomized, prospective, controlled study comparing radical prostatectomy alone and neoadjuvant androgen withdrawal in the treatment of localized prostate cancer. Canadian Urologic Oncology Group”, J Urol., vol. 156, no. 3, pp. 873-877.

Graefen, M., Hammerer, P., Michl, U., Noldus, J., Haese, A., Henke, R. P., Huland, E., & Huland, H. 1998, "Incidence of positive surgical margins after biopsy-selected nerve- sparing radical prostatectomy”, Urology, vol. 51, no. 3, pp. 437-442.

Gray, I. C., Phillips, S. M., Lee, S. J., Neoptolemos, J. P., Weissenbach, J., & Spurr, N. K. 1995, "Loss of the chromosomal region 10q23-25 in prostate cancer", Cancer Res. , vol. 55, no. 21, pp. 4800-4803.

Grignon, D. J., Caplan, R., Sarkar, F. H., Lawton, C. A., Hammond, E. H., Pilepich, M. V., Forman, J. D., Mesic, J., Fu, K. K., Abrams, R. A., Pajak, T. F., Shipley, W. U., & Cox, J. D. 1997, ”p53 status and prognosis of locally advanced prostatic adenocarcinoma: a study based on RTOG 8610”, J.NalLCancer Inst, vol. 89, no. 2, pp. 158-165.

Grignon, D. J. & Sakr, W. A. 1994, "Zonal origin of prostatic adenocarcinoma: are there biologic differences between transition zone and peripheral zone adenocarcinomas of the prostate gland?”, J Cell BiochenuSuppl, vol. 19, pp. 267-269.

Grossfeld, G. D., Chang, J. J., Broering, J. M., Miller, D. P., Yu, J., Flanders, S. C., Henning, J. M., Stier, D. M., & Carroll, P. R. 2000a, "Impact of positive surgical margins on prostate cancer recurrence and the use of secondary cancer treatment: data from the CaPSURE database”, J.UroL, vol. 163, no. 4, pp. 1171-1177.

251 Grossfeld, G. D., Latini, D. M., Lubeck, D. P., Broering, J. M., Li, Y. P., Mehta, S. S., & Carroll, P. R. 2002, "Predicting disease recurrence in intermediate and high-risk patients undergoing radical prostatectomy using percent positive biopsies: results from CaPSURE”, Urology, vol. 59, no. 4, pp. 560-565.

Grossfeld, G. D., Tigrani, V. S., Nudell, D., Roach, M., Ill, Weinberg, V. K., Presti, J. C., Jr., Small, E. J., & Carroll, P. R. 2000b, "Management of a positive surgical margin after radical prostatectomy: decision analysis", J.UroL, vol. 164, no. 1, pp. 93-99.

Gu, K., Mes-Masson, A. M., Gauthier, J., & Saad, F. 1998, "Analysis of the pl6 in early-stage prostate cancer", MoLCarcinog., vol. 21, no. 3, pp. 164-170.

Guillonneau, B. & Vallancien, G. 2000, "Laparoscopic radical prostatectomy: the Montsouris experience", J.UroL, vol. 163, no. 2, pp. 418- 422.

Gumbiner, L. M., Gumerlock, P. H., Mack, P. C., Chi, S. G., deVere White, R. W., Mohler, J. L., Pretlow, T. G., & Tricoli, J. V. 1999, "Overexpression of cyclin D1 is rare in human prostate carcinoma", Prostate, vol. 38, no. 1, pp. 40-45.

Guo, Y., Sklar, G. N., Borkowski, A., & Kyprianou, N. 1997, "Loss of the cyclin-dependent kinase inhibitor p27(Kipl) protein in human prostate cancer correlates with tumor grade", Clin.Cancer Res., vol. 3, no. 12 Pt 1, pp. 2269-2274.

Habuchi, T., Liqing, Z., Suzuki, T., Sasaki, R., Tsuchiya, N., Tachiki, H., Shimoda, N., Satoh, S., Sato, K., Kakehi, Y., Kamoto, T., Ogawa, O., & Kato, T. 2000a, "Increased risk of prostate cancer and benign prostatic hyperplasia associated with a CYP17 gene polymorphism with a gene dosage effect", Cancer Res., vol. 60, no. 20, pp. 5710-5713.

Habuchi, T., Suzuki, T., Sasaki, R., Wang, L., Sato, K., Satoh, S., Akao, T., Tsuchiya, N., Shimoda, N., Wada, Y., Koizumi, A., Chihara, J., Ogawa, O., & Kato, T. 2000b, "Association of vitamin D receptor gene polymorphism with prostate cancer and benign prostatic hyperplasia in a Japanese population", Cancer Res., vol. 60, no. 2, pp. 305-308.

Haese, A., Graefen, M., Becker, C., Noldus, J., atz, J., Cagiannos, I., attan, M. W., cardino, P. T., uland, E., uland, H., & ilja, H. 2003, "The role of human glandular kallikrein 2 for prediction of pathologically organ confined prostate cancer.", Prostate, vol. 54 , no. (3), pp. 181-186.

Haggman, M., Nordin, B., Mattson, S., & Busch, C. 1997a, "Morphometric studies of intra-prostatic volume relationships in localized prostatic cancer", Br.J Urol, vol. 80, no. 4, pp. 612-617.

Haggman, M. J., Wojno, K. J., Pearsall, C. P., & Macoska, J. A. 1997b, "Allelic loss of 8p sequences in prostatic intraepithelial neoplasia and carcinoma", Urology, vol. 50, no. 4, pp. 643-647.

252 Halvorsen, O. J., Haukaas, S. A., & Akslen, L. A. 2003, "Combined loss of PTEN and p27 expression is associated with tumor cell proliferation by Ki-67 and increased risk of recurrent disease in localized prostate cancer", Clin.Cancer Res., vol. 9, no. 4, pp. 1474-1479.

Halvorsen, O. J., Hostmark, J., Haukaas, S., Hoisaeter, P. A., & Akslen, L. A. 2000, "Prognostic significance of pl6 and CDK4 proteins in localized prostate carcinoma", Cancer , vol. 88, no. 2, pp. 416-424.

Han, M., Partin, A. W., Piantadosi, S., Epstein, J. 1., & Walsh, P. C. 2001, "Era specific biochemical recurrence-free survival following radical prostatectomy for clinically localized prostate cancer", J.UroL, vol. 166, no. 2, pp. 416-419.

Han, M., Partin, A. W., Zahurak, M., Piantadosi, S., Epstein, J. 1., & Walsh, P. C. 2003, "Biochemical (prostate specific antigen) recurrence probability following radical prostatectomy for clinically localized prostate cancer", J Urol, vol. 169, no. 2, pp. 517-523.

Han, M., Snow, P. B., Epstein, J. I., Chan, T. Y., Jones, K. A., Walsh, P. C., & Partin, A. W. 2000, "A neural network predicts progression for men with gleason score 3+4 versus 4+3 tumors after radical prostatectomy", Urology, vol. 56, no. 6, pp. 994-999.

Harden, S. V., Sanderson, H., Goodman, S. N., Partin, A. A., Walsh, P. C., Epstein, J. I., & Sidransky, D. 2003, "Quantitative GSTP1 methylation and the detection of prostate adenocarcinoma in sextant biopsies", J.NatLCancer Inst., vol. 95, no. 21, pp. 1634-1637.

Harisinghani, M. G., Barentsz, J., Hahn, P. F., Deserno, W. M., Tabatabaei, S., van de Kaa, C. H., de la, R. J., & Weissleder, R. 2003, "Noninvasive detection of clinically occult lymph-node metastases in prostate cancer", N.EngLJMed., vol. 348, no. 25, pp. 2491-2499.

Harkonen, P., Kyllonen, A. P., Nordling, S., & Vihko, P. 2005, "Loss of heterozygosity in chromosomal region 16q24.3 associated with progression of prostate cancer", Prostate, vol. 62, no. 3, pp. 267-274.

Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, K., & Elledge, S. J. 1993, "The p21 Cdk-interacting protein Cipl is a potent inhibitor of G1 cyclin- dependent kinases", Cell, vol. 75, no. 4, pp. 805-816.

Herbst, R. A., Weiss, J., Ehnis, A., Cavenee, W. K., & Arden, K. C. 1994, "Loss of heterozygosity for 10q22-10qter in malignant melanoma progression", Cancer Res. , vol. 54, no. 12, pp. 3111-3114.

Hirama, T. & Koeffler, H. P. 1995, "Role of the cyclin-dependent kinase inhibitors in the development of cancer", Blood, vol. 86, no. 3, pp. 841-854.

Hirst, A. E. & Bergman, R. T. 1954, "Carcinoma of the prostate in men 80 or more years old", Cancer , vol. 7, pp. 136-141.

253 Hogg, A., Bia, B., Onadim, Z., & Cowell, J. K. 1993, "Molecular mechanisms of oncogenic mutations in tumors from patients with bilateral and unilateral retinoblastoma", Proc.Natl.Acad.ScLU.S^49 vol. 90, no. 15, pp. 7351-7355.

Hoh, C. K., Seltzer, M. A., Franklin, J., deKernion, J. B., Phelps, M. E., & Belldegrun, A. 1998, "Positron emission tomography in urological oncology", J UroL, vol. 159, no. 2, pp. 347-356.

Hollstein, M., Sidransky, D., Vogelstein, B., & Harris, C. C. 1991, "p53 mutations in human cancers", Science, vol. 253, no. 5015, pp. 49-53.

Hricak, H., Schoder, H., Pucar, D., Lis, E., Eberhardt, S. C., Onyebuchi, C. N., & Scher, H. I. 2003, "Advances in imaging in the postoperative patient with a rising prostate-specific antigen level", Semin.OncoU vol. 30, no. 5, pp. 616-634.

Hromas, R., Davis, B., Rauscher, F. J., Ill, Klemsz, M., Tenen, D., Hoffman, S., Xu, D., & Morris, J. F. 1996, "Hematopoietic transcriptional regulation by the myeloid zinc finger gene, MZF-1", Curr. Top.Microbiol.Immunol., vol. 211, pp. 159-164.

Hsieh, C. L., Oakley-Girvan, I., Gallagher, R. P., Wu, A. H., Kolonel, L. N., Teh, C. Z., Halpern, J., West, D. W., Paffenbarger, R. S., Jr., & Whittemore, A. S. 1997, "Re: prostate cancer susceptibility locus on chromosome lq: a confirmatory study", J.NatLCancer Inst, vol. 89, no. 24, pp. 1893-1894.

Hsing, A. W., Gao, Y. T., Wu, G., Wang, X., Deng, J., Chen, Y. L., Sesterhenn, I. A., Mostofi, F. K., Benichou, J., & Chang, C. 2000, "Polymorphic CAG and GGN repeat lengths in the androgen receptor gene and prostate cancer risk: a population-based case-control study in China", Cancer Res. , vol. 60, no. 18, pp. 5111-5116.

Huang, A., Gandour-Edwards, R., Rosenthal, S. A., Siders, D. B., Deitch, A. D., & White, R. W. 1998, "p53 and bcl-2 immunohistochemical alterations in prostate cancer treated with radiation therapy", Urology, vol. 51, no. 2, pp. 346-351.

Hugel, A. & Wernert, N. 1999, "Loss of heterozygosity (LOH), malignancy grade and clonality in microdissected prostate cancer", Br.J.Cancer , vol. 79, no. 3-4, pp. 551-557.

Hugosson, J., Aus, G., Becker, C., Carlsson, S., Eriksson, H., Lilja, H., Lodding, P., & Tibblin, G. 2000, "Would prostate cancer detected by screening with prostate-specific antigen develop into clinical cancer if left undiagnosed? A comparison of two population-based studies in Sweden", BJU.Inl., vol. 85, no. 9, pp. 1078-1084.

Hull, G. W., Rabbani, F., Abbas, F., Wheeler, T. M., Kattan, M. W., & Scardino, P. T. 2002, "Cancer control with radical prostatectomy alone in 1,000 consecutive patients", J Urol., vol. 167, no. 2 Pt 1, pp. 528-534.

Humphrey, P. A. & Vollmer, R. T. 1997, "Percentage carcinoma as a measure of prostatic tumor size in radical prostatectomy tissues", M odPathoL, vol. 10, no. 4, pp. 326-333.

254 Huncharek, M. & Muscat, J. 1996, ’’Serum prostate-specific antigen as a predictor of staging abdominal/pelvic computed tomography in newly diagnosed prostate ca n cer Abdonulmaging, vol. 21, no. 4, pp. 364-367.

Ikonen, T., Palvimo, J. J., Kallio, P. J., Reinikainen, P., & Janne, O. A. 1994, "Stimulation of androgen-regulated transactivation by modulators of protein phosphorylation", Endocrinology, vol. 135, no. 4, pp. 1359-1366.

Ingles, S. A., Coetzee, G. A., Ross, R. K., Henderson, B. E., Kolonel, L. N., Crocitto, L., Wang, W., & Haile, R. W. 1998, "Association of prostate cancer with vitamin D receptor haplotypes in African-Americans", Cancer Res., vol. 58, no. 8, pp. 1620-1623.

Irvine, R. A., Yu, M. C., Ross, R. K., & Coetzee, G. A. 1995, "The CAG and GGC microsatellites of the androgen receptor gene are in linkage disequilibrium in men with prostate cancer", Cancer Res., vol. 55, no. 9, pp. 1937-1940.

Isaacs, W., De Marzo, A., & Nelson, W. G. 2002, "Focus on prostate cancer", Cancer Cell, vol. 2, no. 2, pp. 113-116.

Ishii, H., Baffa, R., Numata, S. I., Murakumo, Y., Rattan, S., Inoue, H., Mori, M., Fidanza, V., Alder, H., & Croce, C. M. 1999, "The FEZ1 gene at chromosome 8p22 encodes a leucine-zipper protein, and its expression is altered in multiple human tumors", Proc.Natl.Acad.ScuU.SA., vol. 96, no. 7, pp. 3928-3933.

Ittmann, M. 1996, "Allelic loss on chromosome 10 in prostate adenocarcinoma", Cancer Res., vol. 56, no. 9, pp. 2143-2147.

Ittmann, M. M. & Wieczorek, R. 1996, "Alterations of the retinoblastoma gene in clinically localized, stage B prostate adenocarcinomas", HunuPathoL, vol. 27, no. 1, pp. 28-34.

Iwao, K., Matoba, R., Ueno, N., Ando, A., Miyoshi, Y., Matsubara, K., Noguchi, S., & Kato, K. 2002, "Molecular classification of primary breast tumors possessing distinct prognostic properties", HunuMoLGenet., vol. 11, no. 2, pp. 199-206.

Jack, G. S., Cookson, M. S., Coffey, C. S., Vader, V., Roberts, R. L., Chang, S. S., Smith, J. A., Jr., & Shappell, S. B. 2002, "Pathological parameters of radical prostatectomy for clinical stages Tic versus T2 prostate adenocarcinoma: decreased pathological stage and increased detection of transition zone tumors", J Urol., vol. 168, no. 2, pp. 519-524.

Jaffe, J. M., Malkowicz, S. B., Walker, A. H., MacBride, S., Peschel, R., Tomaszewski, J., Van Arsdalen, K., Wein, A. J., & Rebbeck, T. R. 2000, "Association of SRD5A2 genotype and pathological characteristics of prostate tumors", Cancer Res., vol. 60, no. 6, pp. 1626-1630.

Jarrard, D. F., Bova, G. S., Ewing, C. M., Pin, S. S., Nguyen, S. H., Baylin, S. B., Cairns, P., Sidransky, D., Herman, J. G., & Isaacs, W. B. 1997b, "Deletional, mutational, and methylation analyses of CDKN2 (p 16/MTS 1) in

255 primary and metastatic prostate cancer”, Genes Chromosomes.Cancer, vol. 19, no. 2, pp. 90-96.

Jarrard, D. F., Bova, G. S., Ewing, C. M., Pin, S. S., Nguyen, S. H., Baylin, S. B., Cairns, P., Sidransky, D., Herman, J. G., & Isaacs, W. B. 1997a, ”Deletional, mutational, and methylation analyses of CDKN2 (pl6/MTSl) in primary and metastatic prostate cancer”, Genes Chromosomes.Cancer, vol. 19, no. 2, pp. 90-96.

Jarrard, D. F., Modder, J., Fadden, P., Fu, V., Sebree, L., Heisey, D., Schwarze, S. R., & Friedl, A. 2002, ’’Alterations in the pl6/pRb cell cycle checkpoint occur commonly in primary and metastatic human prostate cancer”, Cancer Lett, vol. 185, no. 2, pp. 191-199.

Jassem, E., Bigda, J., Dziadziuszko, R., Schlichtholz, B., Le Roux, D., Grodzki, T., Rzyman, W., Konopa, K., Poberezna, M., Dobrzanska, Z., Skokowski, J., Soussi, T., & Jassem, J. 2001, ’’Serum p53 antibodies in small cell lung cancer: the lack of prognostic relevance”, Lung Cancer , vol. 31, no. 1, pp. 17-23.

Jen, J., Kim, H., Piantadosi, S., Liu, Z. F., Levitt, R. C., Sistonen, P., Kinzler, K. W., Vogelstein, B., & Hamilton, S. R. 1994, "Allelic loss of chromosome 18q and prognosis in colorectal cancer”, N.EngLJ Med, vol. 331, no. 4, pp. 213-221.

Johansson, J. E., Holmberg, L., Johansson, S., Bergstrom, R., & Adami, H. 0 . 1997, ’’Fifteen-year survival in prostate cancer. A prospective, population- based study in Sweden”, JAMA, vol. 277, no. 6, pp. 467-471.

Johnstone, P. A., Tarman, G. J., Riffenburgh, R., & et al 1997, ’’Yield of imaging and scintigraphy assessing biochemical failure in prostate cancer patients.”, Urol Oncol, vol. 3, p. 108.

Kallioniemi, O. P. & Visakorpi, T. 1996, ’’Genetic basis and clonal evolution of human prostate cancer”, Adv.Cancer Res., vol. 68, pp. 225-255.

Karlbom, A. E., James, C. D., Boethius, J., Cavenee, W. K., Collins, V. P., Nordenskjold, M., & Larsson, C. 1993, "Loss of heterozygosity in malignant gliomas involves at least three distinct regions on chromosome 10”, HunuGenet., vol. 92, no. 2, pp. 169-174.

Kattan, M. W., Eastham, J. A., Stapleton, A. M., Wheeler, T. M., & Scardino, P. T. 1998, ”A preoperative nomogram for disease recurrence following radical prostatectomy for prostate cancer", JNatLCancer Inst., vol. 90, no. 10, pp. 766-771.

Kattan, M. W., Shariat, S. F., Andrews, B., Zhu, K., Canto, E., Matsumoto, K., Muramoto, M., Scardino, P. T., Ohori, M., Wheeler, T. M., & Slawin, K. M. 2003, "The addition of interleukin-6 soluble receptor and transforming growth factor betal improves a preoperative nomogram for predicting biochemical progression in patients with clinically localized prostate cancer”, J Clin.OncoL, vol. 21, no. 19, pp. 3573-3579.

256 Katz, M. S., Zelefsky, M. J., Venkatraman, E. S., Fuks, Z., Hummer, A., & Leibel, S. A. 2003, "Predictors of biochemical outcome with salvage conformal radiotherapy after radical prostatectomy for prostate cancer", J Clin. Oncol, vol. 21, no. 3, pp. 483-489.

Kawamata, N., Park, D., Wilczynski, S., Yokota, J., & Koeffler, H. P. 1996, "Point mutations of the Mxil gene are rare in prostate cancers", Prostate, vol. 29, no. 3, pp. 191-193.

Kawana, Y., Komiya, A., Ueda, T., Nihei, N., Kuramochi, H., Suzuki, H., Yatani, R., Imai, T., Dong, J. T., Imai, T., Yoshie, O., Barrett, J. C., Isaacs, J. T., Shimazaki, J., Ito, H., & Ichikawa, T. 1997, "Location of KAI1 on the short arm of human chromosome 11 and frequency of allelic loss in advanced human prostate cancer", Prostate, vol. 32, no. 3, pp. 205-213.

Kazemi-Esfarjani, P., Trifiro, M. A., & Pinsky, L. 1995, "Evidence for a repressive function of the long polyglutamine tract in the human androgen receptor: possible pathogenetic relevance for the (CAG)n-expanded neuronopathies", HunuMoLGenet, vol. 4, no. 4, pp. 523-527.

Kibel, A. S., Faith, D. A., Bova, G. S., & Isaacs, W. B. 2000, "Loss of heterozygosity at 12P12-13 in primary and metastatic prostate adenocarcinoma", J Urol, vol. 164, no. 1, pp. 192-196.

Kibel, A. S., Schutte, M., Kern, S. E., Isaacs, W. B., & Bova, G. S. 1998, "Identification of 12p as a region of frequent deletion in advanced prostate cancer", Cancer Res., vol. 58, no. 24, pp. 5652-5655.

Kibel, A. S., Suarez, B. K., Belani, J., Oh, J., Webster, R., Brophy-Ebbers, M., Guo, C., Catalona, W. J., Picus, J., & Goodfellow, P. J. 2003, "CDKN1A and CDKN1B polymorphisms and risk of advanced prostate carcinoma", Cancer Res., vol. 63, no. 9, pp. 2033-2036.

Kiel, H. J., Wieland, W. F., & Rossler, W. 2000, "Local control of prostate cancer by transrectal HIFU-therapy", Arch.Ital.UrolAndrol, vol. 72, no. 4, pp. 313-319.

Kinoshita, H., Shi, Y., Sandefur, C., Meisner, L. F., Chang, C., Choon, A., Reznikoff, C. R., Bova, G. S., Friedl, A., & Jarrard, D. F. 2000, "Methylation of the androgen receptor minimal promoter silences transcription in human prostate cancer", Cancer Res., vol. 60, no. 13, pp. 3623-3630.

Kleihues, P., Schauble, B., zur, H. A., Esteve, J., & Ohgaki, H. 1997, "Tumors associated with p53 germline mutations: a synopsis of 91 families", Am .J Pathol., vol. 150, no. 1, pp. 1-13.

Knudson, A. G., Jr. 1971, "Mutation and cancer: statistical study of retinoblastoma", Proc.Natl.Acad.ScL U.S.A, vol. 68, no. 4, pp. 820-823.

Koivisto, P., Kononen, J., Palmberg, C., Tammela, T., Hyytinen, E., Isola, J., Trapman, J., Cleutjens, K., Noordzij, A., Visakorpi, T., & Kallioniemi, O. P. 1997, "Androgen receptor gene amplification: a possible molecular mechanism for androgen deprivation therapy failure in prostate cancer", Cancer Res., vol. 57, no. 2, pp. 314-319.

257 Kolvenbag, G. J., Iversen, P., & Newling, D. W. 2001, ’’Antiandrogen monotherapy: a new form of treatment for patients with prostate cancer”, Urology, vol. 58, no. 2 Suppl 1, pp. 16-23.

Komiya, A., Suzuki, H., Ueda, T., Yatani, R., Emi, M., Ito, H., & Shimazaki, J. 1996, "Allelic losses at loci on chromosome 10 are associated with metastasis and progression of human prostate cancer”, Genes Chromosomes.Cancer, vol. 17, no. 4, pp. 245-253.

Koppie, T. M., Grossfeld, G. D., Nudell, D. M., Weinberg, V. K., & Carroll, P. R. 2001, "Is anastomotic biopsy necessary before radiotherapy after radical prostatectomy?”, J Urol, vol. 166, no. 1, pp. 111-115.

Kupelian, P., Katcher, J., Levin, H., Zippe, C., & Klein, E. 1996, "Correlation of clinical and pathologic factors with rising prostate-specific antigen profiles after radical prostatectomy alone for clinically localized prostate cancer”, Urology, vol. 48, no. 2, pp. 249-260.

Kwabi-Addo, B., Giri, D., Schmidt, K., Podsypanina, K., Parsons, R., Greenberg, N., & Ittmann, M. 2001, "Haploinsufficiency of the Pten tumor suppressor gene promotes prostate cancer progression”, Proc.Natl.Acad.ScLU.Sji , vol. 98, no. 20, pp. 11563-11568.

Lange, P. H., Ercole, C. J., Lightner, D. J., Fraley, E. E., & Vessella, R. 1989, "The value of serum prostate specific antigen determinations before and after radical prostatectomy”, J UroL, vol. 141, no. 4, pp. 873-879.

Latil, A., Bieche, I., Pesche, S., Volant, A., Valeri, A., Fournier, G., Cussenot, O., & Lidereau, R. 1999, "Loss of heterozygosity at chromosome arm 13q and RBI status in human prostate cancer”, Hum.Pathol., vol. 30, no. 7, pp. 809-815.

Latil, A., Cussenot, O., Fournier, G., Baron, J. C., & Lidereau, R. 1995, "Loss of heterozygosity at 7q31 is a frequent and early event in prostate cancer”, Clin.Cancer Res., vol. 1, no. 11, pp. 1385-1389.

Latil, A., Cussenot, O., Fournier, G., Driouch, K., & Lidereau, R. 1997, "Loss of heterozygosity at chromosome 16q in prostate adenocarcinoma: identification of three independent regions”, Cancer Res., vol. 57, no. 6, pp. 1058-1062.

Lee, H. H., Warde, P., & Jewett, M. A. 1999, "Neoadjuvant hormonal therapy in carcinoma of the prostate”, BJU.Int., vol. 83, no. 4, pp. 438-448.

Lee, W. H., Isaacs, W. B., Bova, G. S., & Nelson, W. G. 1997, ”CG island methylation changes near the GSTP1 gene in prostatic carcinoma cells detected using the polymerase chain reaction: a new prostate cancer biomarker”, Cancer EpidemioLBiomarkers Prev., vol. 6, no. 6, pp. 443-450.

Lerner, S. E., Blute, M. L., Bergstralh, E. J., Bostwick, D. G., Eickholt, J. T., & Zincke, H. 1996, "Analysis of risk factors for progression in patients with pathologically confined prostate cancers after radical retropubic prostatectomy”, J.Urol., vol. 156, no. 1, pp. 137-143.

258 Leube, B., Drechsler, M., Muhlmann, K., Schafer, R., Schulz, W. A., Santourlidis, S., Anastasiadis, A., Ackermann, R., Visakorpi, T., Muller, W., & Royer-Pokora, B. 2002, "Refined mapping of allele loss at chromosome 10q23-26 in prostate cancer”, Prostate, vol. 50, no. 3, pp. 135-144.

Li, C., Berx, G., Larsson, C., Auer, G., Aspenblad, U., Pan, Y., Sundelin, B., Ekman, P., Nordenskjold, M., van Roy, F., & Bergerheim, U. S. 1999, "Distinct deleted regions on chromosome segment 16q23-24 associated with metastases in prostate cancer", Genes Chromosomes.Cancer, vol. 24, no. 3, pp. 175-182.

Li, C., Larsson, C., Futreal, A., Lancaster, J., Phelan, C., Aspenblad, U., Sundelin, B., Liu, Y., Ekman, P., Auer, G., & Bergerheim, U. S. 1998, "Identification of two distinct deleted regions on chromosome 13 in prostate cancer", Oncogene, vol. 16, no. 4, pp. 481-487.

Li, F. P. & Fraumeni, J. F., Jr. 1969, "Soft-tissue , breast cancer, and other neoplasms. A familial syndrome?", Ann.Intern.Med., vol. 71, no. 4, pp. 747-752.

Li, R., Waga, S., Hannon, G. J., Beach, D., & Stillman, B. 1994, "Differential effects by the p21 CDK inhibitor on PCNA-dependent DNA replication and repair", Nature , vol. 371, no. 6497, pp. 534-537.

Liaw, D., Marsh, D. J., Li, J., Dahia, P. L., Wang, S. I., Zheng, Z., Bose, S., Call, K. M., Tsou, H. C., Peacocke, M., Eng, C., & Parsons, R. 1997, "Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome", NatGenet ., vol. 16, no. 1, pp. 64-67.

Lin, Y., Uemura, H., Fujinami, K., Hosaka, M., Harada, M., & Kubota, Y. 1997, "Telomerase activity in primary prostate cancer", J.UroL, vol. 157, no. 3, pp. 1161-1165.

Linja, M. J., Savinainen, K. J., Saramaki, O. R., Tammela, T. L., Vessella, R. L., & Visakorpi, T. 2001, "Amplification and overexpression of androgen receptor gene in hormone- refractory prostate cancer", Cancer Res., vol. 61, no. 9, pp. 3550-3555.

Linzer, D. G., Stock, R. G., Stone, N. N., Ratnow, R., Ianuzzi, C., & Unger, P. 1996, "Seminal vesicle biopsy: accuracy and implications for staging of prostate cancer", Urology, vol. 48, no. 5, pp. 757-761.

Loda, M., Cukor, B., Tam, S. W., Lavin, P., Fiorentino, M., Draetta, G. F., Jessup, J. M., & Pagano, M. 1997, "Increased proteasome-dependent degradation of the cyclin-dependent kinase inhibitor p27 in aggressive colorectal carcinomas", Nat.Med pp. 231-234.

Losa, A., Nava, L. D., Di Muzio, N., Mangili, P., Longobardi, B., Rigatti, P., & Guazzoni, G. 2003, "Salvage brachytherapy for local recurrence after radical prostatectomy and subsequent external beam radiotherapy", Urology, vol. 62, no. 6, pp. 1068-1072.

Lu, Y. J., Dong, X. Y., Guo, S. P., Zhang, R. G., Ke, Y., Zhang, L. Z., Xu, L. H., & Cheng, S. J. 1996, "2q-, a non-random chromosomal abnormality in

259 human non-small-cell lung cancer", Carcinogenesis, vol. 17, no. 8, pp. 1589- 1593.

Lu, Y. J., Hing, S., Williams, R., Pinkerton, R., Shipley, J., & Pritchard- Jones, K. 2002, "Chromosome lq expression profiling and relapse in Wilms' tumour", Lancet , vol. 360, no. 9330, pp. 385-386.

Lu-Yao, G. L., Potosky, A. L., Albertsen, P. C., Wasson, J. H., Barry, M. J., & Wennberg, J. E. 1996, "Follow-up prostate cancer treatments after radical prostatectomy: a population-based study", JNatLCancer Inst., vol. 88, no. 3- 4, pp. 166-173.

Lundberg, S. & Berge, T. 1970, "Prostatic carcinoma. An autopsy study", ScandJ Urol NephroL, vol. 4, no. 2, pp. 93-97.

Lynch, H. T., Shaw, M. W., Magnuson, C. W., Larsen, A. L., & Krush, A. J. 1966, "Hereditary factors in cancer. Study of two large midwestern kindreds", Arch.Intern.Med., vol. 117, no. 2, pp. 206-212.

MacGrogan, D., Levy, A., Bostwick, D., Wagner, M., Wells, D., & Bookstein, R. 1994, "Loss of chromosome arm 8p loci in prostate cancer: mapping by quantitative allelic imbalance", Genes Chromosomes.Cancer, vol. 10, no. 3, pp. 151-159.

Mackey, T. J., Borkowski, A., Amin, P., Jacobs, S. C., & Kyprianou, N. 1998, "bcl-2/bax ratio as a predictive marker for therapeutic response to radiotherapy in patients with prostate cancer", Urology, vol. 52, no. 6, pp. 1085-1090.

Macoska, J. A., Micale, M. A., Sakr, W. A., Benson, P. D., & Wolman, S. R. 1993, "Extensive genetic alterations in prostate cancer revealed by dual PCR and FISH analysis", Genes Chromosomes.Cancer, vol. 8, no. 2, pp. 88-97.

Macoska, J. A., Trybus, T. M., Benson, P. D., Sakr, W. A., Grignon, D. J., Wojno, K. D., Pietruk, T., & Powell, I. J. 1995, "Evidence for three tumor suppressor gene loci on chromosome 8p in human prostate cancer", Cancer Res., vol. 55, no. 22, pp. 5390-5395.

Madersbacher, S., Haidinger, G., Temml, C., & Schmidbauer, C. P. 1998, "Prevalence of lower urinary tract symptoms in Austria as assessed by an open survey of 2,096 men", Eur.Urol., vol. 34, no. 2, pp. 136-141.

Mannervik, B., Alin, P., Guthenberg, C., Jensson, H., Tahir, M. K., Warholm, M., & Jornvall, H. 1985, "Identification of three classes of cytosolic glutathione transferase common to several mammalian species: correlation between structural data and enzymatic properties", Proc.Natl.Acad.ScL U.S.A , vol. 82, no. 21, pp. 7202-7206.

Manyak, M. J. & Javitt, M. C. 1998, "The role of computerized tomography, magnetic resonance imaging, bone scan, and monoclonal antibody nuclear scan for prognosis prediction in prostate cancer", Semin.UroLOncol., vol. 16, no. 3, pp. 145-152.

260 Mao, L., Schoenberg, M. P., Scicchitano, M., Erozan, Y. S., Merlo, A., Schwab, D., & Sidransky, D. 1996, "Molecular detection of primary bladder cancer by microsatellite analysis", Science, vol. 271, no. 5249, pp. 659-662.

Marsh, D. J., Dahia, P. L., Zheng, Z., Liaw, D., Parsons, R., Gorlin, R. J., & Eng, C. 1997, "Germline mutations in PTEN are present in Bannayan- Zonana syndrome", NatGenet ., vol. 16, no. 4, pp. 333-334.

Matsushima, H., Sasaki, T., Goto, T., Hosaka, Y., Homma, Y., Kitamura, T., Kawabe, K., Sakamoto, A., Murakami, T., & Machinami, R. 1998, "Immunohistochemical study of p21WAFl and p53 proteins in prostatic cancer and their prognostic significance", HunuPathoi, vol. 29, no. 8, pp. 778-783.

McDonnell, T. J., Navone, N. M., Troncoso, P., Pisters, L. L., Conti, C., von Eschenbach, A. C., Brisbay, S., & Logothetis, C. J. 1997, "Expression of bcl-2 oncoprotein and p53 protein accumulation in bone marrow metastases of androgen independent prostate cancer", J.Urol., vol. 157, no. 2, pp. 569-574.

Mclndoe, R. A., Stanford, J. L., Gibbs, M., Jarvik, G. P., Brandzel, S., Neal, C. L., Li, S., Gammack, J. T., Gay, A. A., Goode, E. L., Hood, L., & Ostrander, E. A. 1997, "Linkage analysis of 49 high-risk families does not support a common familial prostate cancer-susceptibility gene at lq24-25", AnuJ.HunuGenel., vol. 61, no. 2, pp. 347-353.

McNeal, J. E. 1972, "The prostate and prostatic urethra: a morphologic synthesis", J Urol, vol. 107, no. 6, pp. 1008-1016.

McNeal, J. E., Redwine, E. A., Freiha, F. S., & Stamey, T. A. 1988, "Zonal distribution of prostatic adenocarcinoma. Correlation with histologic pattern and direction of spread", AnuJ.Surg.PalhoL, vol. 12, no. 12, pp. 897-906.

McNeal, J. E. & Yemoto, C. E. 1996, "Significance of demonstrable vascular space invasion for the progression of prostatic adenocarcinoma", AnuJ.Surg.PathoL, vol. 20, no. 11, pp. 1351-1360.

Medical Research Council Prostate Cancer Working Party Investigators Group 1997, "Immediate versus deferred treatment for advanced prostatic cancer: initial results of the Medical Research Council Trial.", Br.J Urol, vol. 79, no. 2, pp. 235-246.

Merrill, R. M. & Wiggins, C. L. 2002, "Incidental detection of population- based prostate cancer incidence rates through transurethral resection of the prostate", UroLOncol., vol. 7, no. 5, pp. 213-219.

Messing, E. M., Manola, J., Sarosdy, M., Wilding, G., Crawford, E. D., & Trump, D. 1999, "Immediate hormonal therapy compared with observation after radical prostatectomy and pelvic lymphadenectomy in men with node­ positive prostate cancer", N.EngLJ.Medvol. 341, no. 24, pp. 1781-1788.

Miyake, H., Hara, I., Yamanaka, K., Gohji, K., Arakawa, S., & Kamidono, S. 1999, "Overexpression of Bcl-2 enhances metastatic potential of human bladder cancer cells", Br.J.Cancer , vol. 79, no. 11-12, pp. 1651-1656.

261 Montironi, R., Diamanti, L., Santinelli, A., Galetti-Prayer, T., Zattoni, F., Selvaggi, F. P., Pagano, F., & Bono, A. V. 1999, ’’Effect of total androgen ablation on pathologic stage and resection limit status of prostate cancer. Initial results of the Italian PROSIT study”, PathoLRes.Pract., vol. 195, no. 4, pp. 201-208.

Morita, R., Saito, S., Ishikawa, J., Ogawa, O., Yoshida, O., Yamakawa, K., & Nakamura, Y. 1991, ’’Common regions of deletion on chromosomes 5q, 6q, and lOq in renal cell carcinoma”, Cancer Res., vol. 51, no. 21, pp. 5817-5820.

Morris, J. F., Rauscher, F. J., Ill, Davis, B., Klemsz, M., Xu, D., Tenen, D., & Hromas, R. ’’The myeloid zinc finger gene, MZF-1, regulates the CD34 promoter in vitro”.

Moul, J. W., Connelly, R. R., Lubeck, D. P., Bauer, J. J., Sun, L., Flanders, S. C., Grossfeld, G. D., & Carroll, P. R. 2001, ’’Predicting risk of prostate specific antigen recurrence after radical prostatectomy with the Center for Prostate Disease Research and Cancer of the Prostate Strategic Urologic Research Endeavor databases”, J Urol, vol. 166, no. 4, pp. 1322-1327.

Narayan, P., Patel, M., Rice, L., & Furman, J. 2000, ’’Molecular biology, endocrinology and physiology of the prostate and male accessory sex glands,” in Benign Prostatic hyperplasia, P. Narayan, ed., Churchill Livingstone, pp. 3- 43.

Narla, G., Difeo, A., Reeves, H. L., Schaid, D. J., Hirshfeld, J., Hod, E., Katz, A., Isaacs, W. B., Hebbring, S., Komiya, A., McDonnell, S. K., Wiley, K. E., Jacobsen, S. J., Isaacs, S. D., Walsh, P. C., Zheng, S. L., Chang, B. L., Friedrichsen, D. M., Stanford, J. L., Ostrander, E. A., Chinnaiyan, A. M., Rubin, M. A., Xu, J., Thibodeau, S. N., Friedman, S. L., & Martignetti, J. A. 2005, ” A germline DNA polymorphism enhances alternative splicing of the KLF6 tumor suppressor gene and is associated with increased prostate cancer risk”, Cancer Res., vol. 65, no. 4, pp. 1213-1222.

Narla, G., Heath, K. E., Reeves, H. L., Li, D., Giono, L. E., Kimmelman, A. C., Glucksman, M. J., Narla, J., Eng, F. J., Chan, A. M., Ferrari, A. C., Martignetti, J. A., & Friedman, S. L. 2001, "KLF6, a candidate tumor suppressor gene mutated in prostate cancer”, Science, vol. 294, no. 5551, pp. 2563-2566.

National Institute for Clinical Excellence 2002, Guidance on Cancer Services: Improving outcomes in urological cancers.

Navone, N. M., Troncoso, P., Pisters, L. L., Goodrow, T. L., Palmer, J. L., Nichols, W. W., von Eschenbach, A. C., & Conti, C. J. 1993, ”p53 protein accumulation and gene mutation in the progression of human prostate carcinoma”, J.NatLCancer Inst., vol. 85, no. 20, pp. 1657-1669.

Nelson, W. G. & Wilding, G. 2001, ’’Prostate cancer prevention agent development: Criteria and pipeline for candidate chemoprevention agents”, Urology, vol. 57, no. 4 Suppl 1, pp. 56-63.

262 Neshat, M. S., Mellinghoff, I. K., Tran, C., Stiles, B., Thomas, G., Petersen, R., Frost, P., Gibbons, J. J., Wu, H., & Sawyers, C. L. "Enhanced sensitivity of PTEN-deficient tumors to inhibition of FRAP/mTOR".

Noda, H., Maehara, Y., Irie, K., Kakeji, Y., Yonemura, T., & Sugimachi, K. 2001, "Growth pattern and expressions of cell cycle regulator proteins p53 and p21WAFl/CIPl in early gastric carcinoma", Cancer , vol. 92, no. 7, pp. 1828-1835.

Noguchi, M., Stamey, T. A., Neal, J. E., & Yemoto, C. E. 2000, "An analysis of 148 consecutive transition zone cancers: clinical and histological characteristics", J Urol, vol. 163, no. 6, pp. 1751-1755.

Noordzij, M. A., van Steenbrugge, G. J., Verkaik, N. S., Schroder, F. H., & van der Kwast, T. H. 1997, "The prognostic value of CD44 isoforms in prostate cancer patients treated by radical prostatectomy", Clin.Cancer Res., vol. 3, no. 5, pp. 805-815.

Nudell, D. M., Grossfeld, G. D., Weinberg, V. K., Roach, M., Ill, & Carroll, P. R. 1999, "Radiotherapy after radical prostatectomy: treatment outcomes and failure patterns", Urology, vol. 54, no. 6, pp. 1049-1057.

Nupponen, N. & Visakorpi, T. 1999, "Molecular biology of progression of prostate cancer", Eur.UroL, vol. 35, no. 5-6, pp. 351-354.

Nupponen, N. N., Kakkola, L., Koivisto, P., & Visakorpi, T. 1998, "Genetic alterations in hormone-refractory recurrent prostate carcinomas", AnuJ.PathoL, vol. 153, no. 1, pp. 141-148.

Oba, K., Matsuyama, H., Yoshihiro, S., Kishi, F., Takahashi, M., Tsukamoto, M., Kinjo, M., Sagiyama, K., & Naito, K. 2001, "Two putative tumor suppressor genes on chromosome arm 8p may play different roles in prostate cancer", Cancer GenetCytogenet, vol. 124, no. 1, pp. 20-26.

Obek, C., Sadek, S., Lai, S., Civantos, F., Rubinowicz, D., & Soloway, M. S. 1999, "Positive surgical margins with radical retropubic prostatectomy: anatomic site-specific pathologic analysis and impact on prognosis", Urology, vol. 54, no. 4, pp. 682-688.

Oesterling, J. E., Chan, D. W., Epstein, J. I., Kimball, A. W., Jr., Bruzek, D. J., Rock, R. C., Brendler, C. B., & Walsh, P. C. 1988, "Prostate specific antigen in the preoperative and postoperative evaluation of localized prostatic cancer treated with radical prostatectomy", J Urol, vol. 139, no. 4, pp. 766-772.

Ohori, M., Wheeler, T. M., Kattan, M. W., Goto, Y., & Scardino, P. T. 1995, "Prognostic significance of positive surgical margins in radical prostatectomy specimens", J .U r o lvol. 154, no. 5, pp. 1818-1824.

Olsson, C. A., de Vries, G. M., Benson, M. C., Raffo, A., Buttyan, R., Cama, C., O’Toole, K., & Katz, A. E. 1996, "The use of RT-PCR for prostate- specific antigen assay to predict potential surgical failures before radical prostatectomy: molecular staging of prostate cancer", Br.J.UroL, vol. 77, no. 3, pp. 411-417.

263 Ornstein, D. K., Colberg, J. W., Virgo, K. S., Chan, D., Johnson, E. T., Oh, J., & Johnson, F. E. 1998, "Evaluation and management of men whose radical prostatectomies failed: results of an international survey”, Urology, vol. 52, no. 6, pp. 1047-1054.

Orozco, R., O’Dowd, G., Kunnel, B., Miller, M. C., & Veltri, R. W. 1998, "Observations on pathology trends in 62,537 prostate biopsies obtained from urology private practices in the United States", Urology, vol. 51, no. 2, pp. 186-195.

Ostrander, E. A. & Stanford, J. L. 2000, "Genetics of prostate cancer: too many loci, too few genes", Am. J.HunuGenet., vol. 67, no. 6, pp. 1367-1375.

Pannek, J., Rittenhouse, H. G., Chan, D. W., Epstein, J. 1., Walsh, P. C., & Partin, A. W. 1998, "The use of percent free prostate specific antigen for staging clinically localized prostate cancer", J.Urol., vol. 159, no. 4, pp. 1238- 1242.

Parker, C. 2004, "Active surveillance: towards a new paradigm in the management of early prostate cancer", Lancet Oncol., vol. 5, no. 2, pp. 101- 106.

Partin, A. W., Kattan, M. W., Subong, E. N., Walsh, P. C., Wojno, K. J., Oesterling, J. E., Scardino, P. T., & Pearson, J. D. 1997, "Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update", JAMA, vol. 277, no. 18, pp. 1445-1451.

Partin, A. W., Mangold, L. A., Lamm, D. M., Walsh, P. C., Epstein, J. I., & Pearson, J. D. 2001, "Contemporary update of prostate cancer staging nomograms (Partin Tables) for the new millennium", Urology, vol. 58, no. 6, pp. 843-848.

Partin, A. W., Pearson, J. D., Landis, P. K., Carter, H. B., Pound, C. R., Clemens, J. Q., Epstein, J. I., & Walsh, P. C. 1994, "Evaluation of serum prostate-specific antigen velocity after radical prostatectomy to distinguish local recurrence from distant metastases", Urology, vol. 43, no. 5, pp. 649- 659.

Partin, A. W., Piantadosi, S., Sanda, M. G., Epstein, J. I., Marshall, F. F., Mohler, J. L., Brendler, C. B., Walsh, P. C., & Simons, J. W. 1995, "Selection of men at high risk for disease recurrence for experimental adjuvant therapy following radical prostatectomy", Urology, vol. 45, no. 5, pp. 831-838.

Partin, A. W., Pound, C. R., Clemens, J. Q., Epstein, J. I., & Walsh, P. C. 1993, "Serum PSA after anatomic radical prostatectomy. The Johns Hopkins experience after 10 years", Urol.Clin.North Am., vol. 20, no. 4, pp. 713-725.

Patel, A., Dorey, F., Franklin, J., & deKernion, J. B. 1997, "Recurrence patterns after radical retropubic prostatectomy: clinical usefulness of prostate specific antigen doubling times and log slope prostate specific antigen", J Urol, vol. 158, no. 4, pp. 1441-1445.

264 Paulson, D. F., Moul, J. W., & Walther, P. J. 1990, "Radical prostatectomy for clinical stage T1-2N0M0 prostatic adenocarcinoma: long-term results", J.Urol., vol. 144, no. 5, pp. 1180-1184.

Pavelic, J., Zeljko, Z., & Bosnar, M. H. 2003, "Molecular genetic aspects of prostate transition zone lesions", Urology, vol. 62, no. 4, pp. 607-613.

Paweletz, C. P., Liotta, L. A., & Petricoin, E. F., Ill 2001, "New technologies for biomarker analysis of prostate cancer progression: Laser capture microdissection and tissue proteomics", Urology, vol. 57, no. 4 Suppl 1, pp. 160-163.

Peehl, D. M., Skowronski, R. J., Leung, G. K., Wong, S. T., Stamey, T. A., & Feldman, D. 1994, "Antiproliferative effects of 1,25-dihydroxyvitamin D3 on primary cultures of human prostatic cells", Cancer Res., vol. 54, no. 3, pp. 805-810.

Pegram, M. D. & Slamon, D. J. 1999, "Combination therapy with trastuzumab (Herceptin) and cisplatin for chemoresistant metastatic breast cancer: evidence for receptor-enhanced chemosensitivity", Semin.Oncol., vol. 26, no. 4 Suppl 12, pp. 89-95.

Peiffer, S. L., Herzog, T. J., Tribune, D. J., Mutch, D. G., Gersell, D. J., & Goodfellow, P. J. 1995, "Allelic loss of sequences from the long arm of chromosome 10 and replication errors in endometrial cancers", Cancer Res., vol. 55, no. 9, pp. 1922-1926.

Perinchery, G., Bukurov, N., Nakajima, K., Chang, J., Hooda, M., Oh, B. R., & Dahiya, R. 1999, "Loss of two new loci on chromosome 8 (8p23 and 8ql2- 13) in human prostate cancer", IntJ.Oncol., vol. 14, no. 3, pp. 495-500.

Perou, C. M., Sorlie, T., Eisen, M. B., van de, R. M., Jeffrey, S. S., Rees, C. A., Pollack, J. R., Ross, D. T., Johnsen, H., Akslen, L. A., Fluge, O., Pergamenschikov, A., Williams, C., Zhu, S. X., Lonning, P. E., Borresen- Dale, A. L., Brown, P. O., & Botstein, D. 2000, "Molecular portraits of human breast tumours", Nature, vol. 406, no. 6797, pp. 747-752.

Pesche, S., Latil, A., Muzeau, F., Cussenot, O., Fournier, G., Longy, M., Eng, C., & Lidereau, R. 1998, "PTEN/MMAC1/TEP1 involvement in primary prostate cancers", Oncogene, vol. 16, no. 22, pp. 2879-2883.

Peschel, R. E., Robnett, T. J., Hesse, D., King, C. R., Ennis, R. D., Schiff, P. B., & Wilson, L. D. 2000, "PSA based review of adjuvant and salvage radiation therapy vs. observation in postoperative prostate cancer patients", Int. J.Cancer, vol. 90, no. 1, pp. 29-36.

Petricoin, E. F., Ardekani, A. M., Hitt, B. A., Levine, P. J., Fusaro, V. A., Steinberg, S. M., Mills, G. B., Simone, C., Fishman, D. A., Kohn, E. C., & Liotta, L. A. 2002, "Use of proteomic patterns in serum to identify ovarian cancer", Lancet, vol. 359, no. 9306, pp. 572-577.

Peyromaure, M., Allouch, M., Eschwege, F., Verpillat, P., Debre, B., & Zerbib, M. 2003, "Salvage radiotherapy for biochemical recurrence after

265 radical prostatectomy: a study of 62 patients", Urology, vol. 62, no. 3, pp. 503-507.

Phillips, S. M., Morton, D. G., Lee, S. J., Wallace, D. M., & Neoptolemos, J. P. 1994, "Loss of heterozygosity of the retinoblastoma and adenomatous polyposis susceptibility gene loci and in chromosomes lOp, lOq and 16q in human prostate cancer", Br.J Urol, vol. 73, no. 4, pp. 390-395.

Pound, C. R., Partin, A. W., Eisenberger, M. A., Chan, D. W., Pearson, J. D., & Walsh, P. C. 1999, "Natural history of progression after PSA elevation following radical prostatectomy", JAMA, vol. 281, no. 17, pp. 1591-1597.

Pound, C. R., Partin, A. W., Epstein, J. I., & Walsh, P. C. 1997, "Prostate- specific antigen after anatomic radical retropubic prostatectomy. Patterns of recurrence and cancer control", Urol.Clin.North Am., vol. 24, no. 2, pp. 395- 406.

Prasad, M. A., Trybus, T. M., Wojno, K. J., & Macoska, J. A. 1998, "Homozygous and frequent deletion of proximal 8p sequences in human prostate cancers: identification of a potential tumor suppressor gene site", Genes Chromosomes.Cancer, vol. 23, no. 3, pp. 255-262.

Prendergast, N. J., Atkins, M. R., Schatte, E. C., Paulson, D. F., & Walther, P. J. 1996, "p53 immunohistochemical and genetic alterations are associated at high incidence with post-irradiated locally persistent prostate carcinoma", J.UroL, vol. 155, no. 5, pp. 1685-1692.

Prochownik, E. V., Eagle, G. L., Deubler, D., Zhu, X. L., Stephenson, R. A., Rohr, L. R., Yin, X., & Brothman, A. R. 1998, "Commonly occurring loss and mutation of the MXI1 gene in prostate cancer", Genes Chromosomes.Cancer, vol. 22, no. 4, pp. 295-304.

Qiu, Y., Ravi, L., & Kung, H. J. 1998, "Requirement of ErbB2 for signalling by interleukin-6 in prostate carcinoma cells", Nature , vol. 393, no. 6680, pp. 83-85.

Quinn, D. I., Henshall, S. M., Haynes, A. M., Brenner, P. C., Kooner, R., Golovsky, D., Mathews, J., O’Neill, G. F., Turner, J. J., Delprado, W., Finlayson, J. F., Sutherland, R. L., Grygiel, J. J., & Strieker, P. D. 2001, "Prognostic significance of pathologic features in localized prostate cancer treated with radical prostatectomy: implications for staging systems and predictive models", J Clin.Oncol., vol. 19, no. 16, pp. 3692-3705.

Raaijmakers, R., Blijenberg, B. G., Finlay, J. A., Rittenhouse, H. G., Wildhagen, M. F., Roobol, M. J., & Schroder, F. H. 2004, "Prostate cancer detection in the prostate specific antigen range of 2.0 to 3.9 ng/ml: value of percent free prostate specific antigen on tumor detection and tumor aggressiveness", J.Urol., vol. 171, no. 6 Pt 1, pp. 2245-2249.

Rashid, M., Wojno, K. J., & Marcovich, R. 1999, "Maximum tumour dimension provides a clinically useful and independently specific measure for predicting PSA-free survival following radical prostatectomy", J.Urol., vol. 161(suppl), p. 241.

266 Ravery, V., Chastang, C., Toublanc, M., Boccon-Gibod, L., Delmas, V., & Boccon-Gibod, L. 2000, "Percentage of cancer on biopsy cores accurately predicts extracapsular extension and biochemical relapse after radical prostatectomy for T1-T2 prostate cancer", Eur.UroL, vol. 37, no. 4, pp. 449- 455.

Rebbeck, T. R., Jaffe, J. M., Walker, A. H., Wein, A. J., & Malkowicz, S. B. 1998, "Modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4", J.NatLCancer Inst., vol. 90, no. 16, pp. 1225- 1229.

Rebbeck, T. R., Walker, A. H., Zeigler-Johnson, C., Weisburg, S., Martin, A. M., Nathanson, K. L., Wein, A. J., & Malkowicz, S. B. 2000, "Association of HPC2/ELAC2 genotypes and prostate cancer", Am. J.Hum. Genet., vol. 67, no. 4, pp. 1014-1019.

Reiter, R. E., Sato, I., Thomas, G., Qian, J., Gu, Z., Watabe, T., Loda, M., & Jenkins, R. B. 2000, "Coamplification of prostate stem cell antigen (PSCA) and MYC in locally advanced prostate cancer", Genes Chromosomes.Cancer, vol. 27, no. 1, pp. 95-103.

Rempel, S. A., Schwechheimer, K., Davis, R. L., Cavenee, W. K., & Rosenblum, M. L. 1993, "Loss of heterozygosity for loci on chromosome 10 is associated with morphologically malignant meningioma progression", Cancer Res., vol. 53, no. 10 Suppl, pp. 2386-2392.

Renshaw, A. A., Richie, J. P., Loughlin, K. R., Jiroutek, M., Chung, A., & D’Amico, A. V. 1999, "Maximum diameter of prostatic carcinoma is a simple, inexpensive, and independent predictor of prostate-specific antigen failure in radical prostatectomy specimens. Validation in a cohort of 434 patients", Am.J Clin.PathoL, vol. Ill, no. 5, pp. 641-644.

Rhodes, D. R., Sanda, M. G., Otte, A. P., Chinnaiyan, A. M., & Rubin, M. A. 2003, "Multiplex biomarker approach for determining risk of prostate- specific antigen-defined recurrence of prostate cancer", J.NatLCancer Inst., vol. 95, no. 9, pp. 661-668.

Richmond, P. J., Karayiannakis, A. J., Nagafuchi, A., Kaisary, A. V., & Pignatelli, M. 1997, "Aberrant E-cadherin and alpha-catenin expression in prostate cancer: correlation with patient survival", Cancer Res., vol. 57, no. 15, pp. 3189-3193.

Rifkin, M. D., Zerhouni, E. A., Gatsonis, C. A., Quint, L. E., Paushter, D. M., Epstein, J. I., Hamper, U., Walsh, P. C., & McNeil, B. J. 1990, "Comparison of magnetic resonance imaging and ultrasonography in staging early prostate cancer. Results of a multi-institutional cooperative trial", N.EngLJMed., vol. 323, no. 10, pp. 621-626.

Roberts, W. W., Bergstralh, E. J., Blute, M. L., Slezak, J. M., Carducci, M., Han, M., Epstein, J. I., Eisenberger, M. A., Walsh, P. C., & Partin, A. W. 2001, "Contemporary identification of patients at high risk of early prostate cancer recurrence after radical retropubic prostatectomy", Urology, vol. 57, no. 6, pp. 1033-1037.

267 Roman-Gomez, J., Castillejo, J. A., Jimenez, A., Gonzalez, M. G., Moreno, F., Rodriguez, M. C., Barrios, M., Maldonado, J., & Torres, A. 2002, ”5’ CpG island hypermethylation is associated with transcriptional silencing of the p21(CIPl/WAFl/SDIl) gene and confers poor prognosis in acute lymphoblastic ", Blood, vol. 99, no. 7, pp. 2291-2296.

Ross, J. S., Sheehan, C. E., Ambros, R. A., Nazeer, T., Jennings, T. A., Kaufman, R. P., Jr., Fisher, H. A., Rifkin, M. D., & Kallakury, B. V. 1999, "Needle biopsy DNA ploidy status predicts grade shifting in prostate cancer", Am. J.Surg.Pathol., vol. 23, no. 3, pp. 296-301.

Ruijter, E., van de, K. C., Miller, G., Ruiter, D., Debruyne, F., & Schalken, J. 1999, "Molecular genetics and epidemiology of prostate carcinoma", Endocr.Rev., vol. 20, no. 1, pp. 22-45.

Sakr, W. A., Macoska, J. A., Benson, P., Grignon, D. J., Wolman, S. R., Pontes, J. E., & Crissman, J. D. 1994, "Allelic loss in locally metastatic, multisampled prostate cancer", Cancer Res., vol. 54, no. 12, pp. 3273-3277.

Saleem, M. D., Sanders, H., Abu, E. N., & El Galley, R. 1998, "Factors predicting cancer detection in biopsy of the prostatic fossa after radical prostatectomy", Urology, vol. 51, no. 2, pp. 283-286.

Salomon, L., Levrel, O., Anastasiadis, A. G., Irani, J., De La, T. A., Saint, F., Vordos, D., Cicco, A., Hoznek, A., Chopin, D., & Abbou, C. C. 2003, "Prognostic significance of tumor volume after radical prostatectomy: a multivariate analysis of pathological prognostic factors", Eur.UroL, vol. 43, no. 1, pp. 39-44.

Salonga, D., Danenberg, K. D., Johnson, M., Metzger, R., Groshen, S., Tsao- Wei, D. D., Lenz, H. J., Leichman, C. G., Leichman, L., Diasio, R. B., & Danenberg, P. V. 2000, "Colorectal tumors responding to 5-fluorouracil have low gene expression levels of dihydropyrimidine dehydrogenase, thymidylate synthase, and thymidine phosphorylase", Clin.Cancer Res., vol. 6, no. 4, pp. 1322-1327.

Saric, T., Brkanac, Z., Troyer, D. A., Padalecki, S. S., Sarosdy, M., Williams, K., Abadesco, L., Leach, R. J., & O’Connell, P. 1999, "Genetic pattern of prostate cancer progression", IntJ.Cancer , vol. 81, no. 2, pp. 219-224.

Sarkar, F. H., Sakr, W., Li, Y. W., Macoska, J., Ball, D. E., & Crissman, J. D. 1992, "Analysis of retinoblastoma (RB) gene deletion in human prostatic carcinomas", Prostate, vol. 21, no. 2, pp. 145-152.

Sato, K., Qian, J., Slezak, J. M., Lieber, M. M., Bostwick, D. G., Bergstralh, E. J., & Jenkins, R. B. 1999, "Clinical significance of alterations of chromosome 8 in high-grade, advanced, nonmetastatic prostate carcinoma", J.NatLCancer Inst., vol. 91, no. 18, pp. 1574-1580.

Scardino, P. T., Weaver, R., & Hudson, M. A. 1992, "Early detection of prostate cancer", HunuPathoL, vol. 23, no. 3, pp. 211-222.

Scattoni, V., Roscigno, M., Raber, M., Montorsi, F., Da Pozzo, L., Guazzoni, G., Freschi, M., & Rigatti, P. 2003, "Multiple vesico-urethral biopsies

268 following radical prostatectomy: the predictive roles of TRUS, DRE, PSA and the pathological stage", Eur.Urol, vol. 44, no. 4, pp. 407-414.

Schaid, D. J. & Rowland, C. 1998, "Use of parents, sibs, and unrelated controls for detection of associations between genetic markers and disease", Am.J.Hum.Genet., vol. 63, no. 5, pp. 1492-1506.

Scher, H. I. 2000, "HER2 in prostate cancer~a viable target or innocent bystander?", J.Natl.Cancer Inst., vol. 92, no. 23, pp. 1866-1868.

Schild, S. E., Wong, W. W., Grado, G. L., Buskirk, S. J., Robinow, J. S., Frick, L. M., & Ferrigni, R. G. 1994, "Radiotherapy for isolated increases in serum prostate-specific antigen levels after radical prostatectomy", Mayo Clin.Proc., vol. 69, no. 7, pp. 613-619.

Schreiber-Agus, N., Meng, Y., Hoang, T., Hou, H., Jr., Chen, K., Greenberg, R., Cordon-Cardo, C., Lee, H. W., & DePinho, R. A. 1998, "Role of Mxil in ageing organ systems and the regulation of normal and neoplastic growth", Nature , vol. 393, no. 6684, pp. 483-487.

See, W. A., Wirth, M. P., McLeod, D. G., Iversen, P., Klimberg, I., Gleason, D., Chodak, G., Montie, J., Tyrrell, C., Wallace, D. M., Delaere, K. P., Vaage, S., Tammela, T. L., Lukkarinen, O., Persson, B. E., Carroll, K., & Kolvenbag, G. J. 2002, "Bicalutamide as immediate therapy either alone or as adjuvant to standard care of patients with localized or locally advanced prostate cancer: first analysis of the early prostate cancer program", J.Urol., vol. 168, no. 2, pp. 429-435.

Sellers, T. A., Potter, J. D., Rich, S. S., Drinkard, C. R., Bostick, R. M., Kushi, L. H., Zheng, W., & Folsom, A. R. 1994, "Familial clustering of breast and prostate cancers and risk of postmenopausal breast cancer", J.NatLCancer Inst., vol. 86, no. 24, pp. 1860-1865.

Sellers, W. & Sawyers, C. L. 2001, in Prostate Cancer: Principles and Practice, P. W. Kantoff, P. R. Carroll, & A. V. D'Amico, eds., Lippincott Williams and Wilkins, Philadelphia, pp. 53-76.

Serrano, M., Hannon, G. J., & Beach, D. 1993, "A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4", Nature, vol. 366, no. 6456, pp. 704-707.

Shankey, T. V., Kallioniemi, O. P., Koslowski, J. M., Lieber, M. L., Mayall, B. H., Miller, G., & Smith, G. J. 1993, "Consensus review of the clinical utility of DNA content cytometry in prostate cancer", Cytometry, vol. 14, no. 5, pp. 497-500.

Shariat, S. F., Tokunaga, H., Zhou, J., Kim, J., Ayala, G. E., Benedict, W. F., & Lerner, S. P. 2004, "p53, p21, pRB, and pl6 expression predict clinical outcome in cystectomy with bladder cancer", J Clin.Oncol, vol. 22, no. 6, pp. 1014-1024.

Shekarriz, B., Upadhyay, J., Wood, D. P., Jr., Hinman, J., Raasch, J., Cummings, G. D., Grignon, D., & Littrup, P. J. 1999, "Vesicourethral

269 anastomosis biopsy after radical prostatectomy: predictive value of prostate- specific antigen and pathologic stage", Urology, vol. §4, no. 6, pp. 1044-1048.

Sherr, C. J. 1996, "Cancer cell cycles", Science, vol. 274, no. 5293, pp. 1672- 1677.

Sherr, C. J. & Roberts, J. M. 1999, "CDK inhibitors: positive and negative regulators of Gl-phase progression", Genes Dev., vol. 13, no. 12, pp. 1501- 1512.

Shi, X. B., Bodner, S. M., deVere White, R. W., & Gumerlock, P. H. 1996, "Identification of p53 mutations in archival prostate tumors. Sensitivity of an optimized single-strand conformational polymorphism (SSCP) assay", Diagn.MoLPathoL, vol. 5, no. 4, pp. 271-278.

Shibata, A. & Whittemore, A. S. 2001, "Re: Prostate cancer incidence and mortality in the United States and the United Kingdom", J.NatLCancer Inst., vol. 93, no. 14, pp. 1109-1110.

Shiohara, M., el Deiry, W. S., Wada, M., Nakamaki, T., Takeuchi, S., Yang, R., Chen, D. L., Vogelstein, B., & Koeffler, H. P. 1994, "Absence of WAF1 mutations in a variety of human malignancies", Blood, vol. 84, no. 11, pp. 3781-3784.

Shoji, T., Tanaka, F., Takata, T., Yanagihara, K., Otake, Y., Hanaoka, N., Miyahara, R., Nakagawa, T., Kawano, Y., Ishikawa, S., Katakura, H., & Wada, H. 2002, "Clinical significance of p21 expression in non-small-cell lung cancer", J Clin.Oncol, vol. 20, no. 18, pp. 3865-3871.

Sidransky, D. 2002, "Emerging molecular markers of cancer", NatRev.Cancer, vol. 2, no. 3, pp. 210-219.

Singh, D., Febbo, P. G., Ross, K., Jackson, D. G., Manola, J., Ladd, C., Tamayo, P., Renshaw, A. A., D'Amico, A. V., Richie, J. P., Lander, E. S., Loda, M., Kantoff, P. W., Golub, T. R., & Sellers, W. R. 2002, "Gene expression correlates of clinical prostate cancer behavior", Cancer Cell, vol. 1, no. 2, pp. 203-209.

Singh, S. P., Lipman, J., Goldman, H., Ellis, F. H., Jr., Aizenman, L., Cangi, M. G., Signoretti, S., Chiaur, D. S., Pagano, M., & Loda, M. "Loss or altered subcellular localization of p27 in Barrett's associated adenocarcinoma".

Skowronski, R. J., Peehl, D. M., & Feldman, D. 1993, "Vitamin D and prostate cancer: 1,25 dihydroxy vitamin D3 receptors and actions in human prostate cancer cell lines", Endocrinology, vol. 132, no. 5, pp. 1952-1960.

Slesak, B., Harlozinska-Szmyrka, A., Knast, W., Sedlaczek, P., van Dalen, A., & Einarsson, R. 2000, "Tissue polypeptide specific antigen (TPS), a marker for differentiation between pancreatic carcinoma and chronic pancreatitis. A comparative study with CA 19-9", Cancer, vol. 89, no. 1, pp. 83-88.

Smith, J. R., Freije, D., Carpten, J. D., Gronberg, H., Xu, J., Isaacs, S. D., Brownstein, M. J., Bova, G. S., Guo, H., Bujnovszky, P., Nusskern, D. R., Damber, J. E., Bergh, A., Emanuelsson, M., Kallioniemi, O. P., Walker-

270 Daniels, J., Bailey-Wilson, J. E., Beaty, T. H., Meyers, D. A., Walsh, P. C., Collins, F. S., Trent, J. M., & Isaacs, W. B. 1996, ’’Major susceptibility locus for prostate cancer on chromosome 1 suggested by a genome-wide search", Science, vol. 274, no. 5291, pp. 1371-1374.

Sofer, M., Savoie, M., Kim, S. S., Civantos, F., & Soloway, M. S. 2003, "Biochemical and pathological predictors of the recurrence of prostatic adenocarcinoma with seminal vesicle invasion", J Urol, vol. 169, no. 1, pp. 153-156.

Soh, S., Kattan, M. W., Berkman, S., Wheeler, T. M., & Scardino, P. T. 1997, "Has there been a recent shift in the pathological features and prognosis of patients treated with radical prostatectomy?", J.U rol, vol. 157, no. 6, pp. 2212-2218.

Soloway, M. S. & et al 1997, "Radical prostatectomy alone versus radical prostatectomy preceded by androgen blockade in cT2b prostate cancer - 24 month results", J Urol, vol. 157, p. 160.

Soloway, M. S. & Neulander, E. 2000, "Bladder-neck preservation during radical retropubic prostatectomy", Semin.Urol.Oncol, vol. 18, no. 1, pp. 51- 56.

Soloway, M. S., Sharifi, R., Wajsman, Z., McLeod, D., Wood, D. P., Jr., & Puras-Baez, A. 1995, "Randomized prospective study comparing radical prostatectomy alone versus radical prostatectomy preceded by androgen blockade in clinical stage B2 (T2bNxM0) prostate cancer. The Lupron Depot Neoadjuvant Prostate Cancer Study Group", J.U rol, vol. 154, no. 2 Pt 1, pp. 424-428.

Sommerfeld, H. J., Meeker, A. K., Piatyszek, M. A., Bova, G. S., Shay, J. W., & Coffey, D. S. 1996, "Telomerase activity: a prevalent marker of malignant human prostate tissue", Cancer Res., vol. 56, no. 1, pp. 218-222.

Stackhouse, G. B., Sesterhenn, I. A., Bauer, J. J., Mostofi, F. K., Connelly, R. R., Srivastava, S. K., & Moul, J. W. 1999, "p53 and bcl-2 immunohistochemistry in pretreatment prostate needle biopsies to predict recurrence of prostate cancer after radical prostatectomy", J Urol, vol. 162, no. 6, pp. 2040-2045.

Stamey, T. A., Dietrick, D. D., & Issa, M. M. 1993, "Large, organ confined, impalpable transition zone prostate cancer: association with metastatic levels of prostate specific antigen", J Urol, vol. 149, no. 3, pp. 510-515.

Stamey, T. A., Freiha, F. S., McNeal, J. E., Redwine, E. A., Whittemore, A. S., & Schmid, H. P. 1993, "Localized prostate cancer. Relationship of tumor volume to clinical significance for treatment of prostate cancer", Cancer, vol. 71, no. 3 Suppl, pp. 933-938.

Stamey, T. A., Kabalin, J. N., McNeal, J. E., Johnstone, I. M., Freiha, F., Redwine, E. A., & Yang, N. 1989, "Prostate specific antigen in the diagnosis and treatment of adenocarcinoma of the prostate. II. Radical prostatectomy treated patients", J.U rol, vol. 141, no. 5, pp. 1076-1083.

271 Stamey, T. A., McNeal, J. E., Yemoto, C. M., Sigal, B. M., & Johnstone, I. M. 1999, "Biological determinants of cancer progression in men with prostate cancer", JAMA, vol. 281, no. 15, pp. 1395-1400.

Stamey, T. A., Villers, A. A., McNeal, J. E., Link, P. C., & Freiha, F. S. 1990, "Positive surgical margins at radical prostatectomy: importance of the apical dissection", J.Urol., vol. 143, no. 6, pp. 1166-1172.

Stamey, T. A., Yang, N., Hay, A. R., McNeal, J. E., Freiha, F. S., & Redwine, E. 1987, "Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate", N.Engl. J.Med., vol. 317, no. 15, pp. 909-916.

Stamey, T. A., Yemoto, C. M., McNeal, J. E., Sigal, B. M., & Johnstone, I. M. 2000, "Prostate cancer is highly predictable: a prognostic equation based on all morphological variables in radical prostatectomy specimens", J.Urol., vol. 163, no. 4, pp. 1155-1160.

Stanford, J. L., Just, J. J., Gibbs, M., Wicklund, K. G., Neal, C. L., Blumenstein, B. A., & Ostrander, E. A. 1997, "Polymorphic repeats in the androgen receptor gene: molecular markers of prostate cancer risk", Cancer Res., vol. 57, no. 6, pp. 1194-1198.

Stattin, P. 1997, "Prognostic factors in prostate cancer", Scand.J.UroLNephroLSuppl,xo\. 185, pp. 1-46.

Steinberg, D. M., Sauvageot, J., Piantadosi, S., & Epstein, J. 1 .1997, "Correlation of prostate needle biopsy and radical prostatectomy Gleason grade in academic and community settings", AnuJ.Surg.Pathol, vol. 21, no. 5, pp. 566-576.

Steinberg, G. D., Carter, B. S., Beaty, T. H., Childs, B., & Walsh, P. C. 1990, "Family history and the risk of prostate cancer", Prostate, vol. 17, no. 4, pp. 337-347.

Strathdee, G. & Brown, R. 2002, "Aberrant DNA methylation in cancer: potential clinical interventions.", Exp Rev M ol Med, vol. http://www- ermm.cbcu.cam.ac.uk/02004222h.htm. no. 4.

Straub, B., Muller, M., Krause, H., Goessl, C., Schrader, M., Heicappell, R., & Miller, K. 2001, "Molecular staging of surgical margins after radical prostatectomy by detection of telomerase activity", Prostate, vol. 49, no. 2, pp. 140-144.

Suarez, B. K., Lin, J., Burmester, J. K., Broman, K. W., Weber, J. L., Banerjee, T. K., Goddard, K. A., Witte, J. S., Elston, R. C., & Catalona, W. J. 2000, "A genome screen of multiplex sibships with prostate cancer", Am.J.Hum.Genet., vol. 66, no. 3, pp. 933-944.

Suzuki, H., Freije, D., Nusskern, D. R., Okami, K., Cairns, P., Sidransky, D., Isaacs, W. B., & Bova, G. S. 1998, "Interfocal heterogeneity of PTEN/MMAC1 gene alterations in multiple metastatic prostate cancer tissues", Cancer Res., vol. 58, no. 2, pp. 204-209.

272 Suzuki, H., Komiya, A., Emi, M., Kuramochi, H., Shiraishi, T., Yatani, R., & Shimazaki, J. 1996, "Three distinct commonly deleted regions of chromosome arm 16q in human primary and metastatic prostate cancers", Genes Chromosomes.Cancer, vol. 17, no. 4, pp. 225-233.

Takeichi, M. 1991, "Cadherin cell adhesion receptors as a morphogenetic regulator", Science, vol. 251, no. 5000, pp. 1451-1455.

Tan, P., Cady, B., Wanner, M., Worland, P., Cukor, B., Magi-Galluzzi, C., Lavin, P., Draetta, G., Pagano, M., & Loda, M. "The cell cycle inhibitor p27 is an independent prognostic marker in small (Tla,b) invasive breast carcinomas".

Taplin, M. E., Bubley, G. J., Shuster, T. D., Frantz, M. E., Spooner, A. E., Ogata, G. K., Keer, H. N., & Balk, S. P. 1995, "Mutation of the androgen- receptor gene in metastatic androgen- independent prostate cancer", N.Engl. J.Med., vol. 332, no. 21, pp. 1393-1398.

Tavtigian, S. V., Simard, J., Teng, D. H., Abtin, V., Baumgard, M., Beck, A., Camp, N. J., Carillo, A. R., Chen, Y., Dayananth, P., Desrochers, M., Dumont, M., Farnham, J. M., Frank, D., Frye, C., Ghaffari, S., Gupte, J. S., Hu, R., Iliev, D., Janecki, T., Kort, E. N., Laity, K. E., Leavitt, A., Leblanc, G., McArthur-Morrison, J., Pederson, A., Penn, B., Peterson, K. T., Reid, J. E., Richards, S., Schroeder, M., Smith, R., Snyder, S. C., Swedlund, B., Swensen, J., Thomas, A., Tranchant, M., Woodland, A. M., Labrie, F., Skolnick, M. H., Neuhausen, S., Rommens, J., & Cannon-Albright, L. A. 2001, "A candidate prostate cancer susceptibility gene at chromosome 17p", N at Genet, vol. 27, no. 2, pp. 172-180.

Taylor, J. A., Hirvonen, A., Watson, M., Pittman, G., Mohler, J. L., & Bell, D. A. 1996, "Association of prostate cancer with vitamin D receptor gene polymorphism", Cancer Res., vol. 56, no. 18, pp. 4108-4110.

Tempany, C. M., Zhou, X., Zerhouni, E. A., Rifkin, M. D., Quint, L. E., Piccoli, C. W., Ellis, J. H., & McNeil, B. J. 1994, "Staging of prostate cancer: results of Radiology Diagnostic Oncology Group project comparison of three MR imaging techniques", Radiology, vol. 192, no. 1, pp. 47-54.

TFSEARCH: Searching Transcription Factor Binding Sites (ver 1.3) 2004, htty://molsun I. cbrc. aist.so.ip/research/db/TFSEARCH. html.

Thiessen, E. U. 1974, "Concerning a familial association between breast cancer and both prostatic and uterine malignancies", Cancer, vol. 34, no. 4, pp. 1102-1107.

Thomas K, Bott SRJ, Besarani D, & Kirby RS 2002, "What is the quality of life for men without a prostate?", British Prostate Group (Spring meeting).

Thomas, H. J., Whitelaw, S. C., Cottrell, S. E., Murday, V. A., Tomlinson, I. P., Markie, D., Jones, T., Bishop, D. T., Hodgson, S. V., Sheer, D., Northover, J. M., Talbot, I. C., Solomon, E., & Bodmer, W. F. 1996, "Genetic mapping of hereditary mixed polyposis syndrome to chromosome 6q", Am. J HunuGenet., vol. 58, no. 4, pp. 770-776.

273 Thompson, I. M., Goodman, P. J., Tangen, C. M., Lucia, M. S., Miller, G. J., Ford, L. G., Lieber, M. M., Cespedes, R. D., Atkins, J. N., Lippman, S. M., Carlin, S. M., Ryan, A., Szczepanek, C. M., Crowley, J. J., & Coltman, C. A., Jr. 2003, "The influence of finasteride on the development of prostate cancer", N.EngLJ.Med'.9 vol. 349, no. 3, pp. 215-224.

Thompson, I. M., Pauler, D. K., Goodman, P. J., Tangen, C. M., Lucia, M. S., Parnes, H. L., Minasian, L. M., Ford, L. G., Lippman, S. M., Crawford, E. D., Crowley, J. J., & Coltman, C. A., Jr. 2004, "Prevalence of prostate cancer among men with a prostate-specific antigen level < or =4.0 ng per milliliter", N.EngLJ.Med., vol. 350, no. 22, pp. 2239-2246.

Tiguert, R., Kabbani, W., Sakr, W., Gheiler, E. L., & Pontes, J. E. 1999, "Origin and racial distribution of glandular tissue in the anterior compartment of the prostate: an autopsy study", Prostate, vol. 39, no. 4, pp. 310-315.

Toyoshima, H. & Hunter, T. 1994, "p27, a novel inhibitor of G1 cyclin-Cdk protein kinase activity, is related to p21", Cell, vol. 78, no. 1, pp. 67-74.

Trapasso, J. G., deKernion, J. B., Smith, R. B., & Dorey, F. 1994, "The incidence and significance of detectable levels of serum prostate specific antigen after radical prostatectomy", J Urol, vol. 152, no. 5 Pt 2, pp. 1821- 1825.

Trapman, J., Sleddens, H. F., van der Weiden, M. M., Dinjens, W. N., Konig, J. J., Schroder, F. H., Faber, P. W., & Bosman, F. T. 1994, "Loss of heterozygosity of chromosome 8 microsatellite loci implicates a candidate tumor suppressor gene between the loci D8S87 and D8S133 in human prostate cancer", Cancer Res.) vol. 54, no. 23, pp. 6061-6064.

Trybus, T. M., Burgess, A. C., Wojno, K. J., Glover, T. W., & Macoska, J. A. 1996, "Distinct areas of allelic loss on chromosomal regions lOp and lOq in human prostate cancer", Cancer Res.) vol. 56, no. 10, pp. 2263-2267.

Tsien, C., Griffith, K. A., Sandler, H. M., McLaughlin, P., Sanda, M. G., Montie, J., Reddy, S., & Hayman, J. A. 2003, "Long-term results of three- dimensional conformal adjuvant and salvage radiotherapy after radical prostatectomy", Urology) vol. 62, no. 1, pp. 93-98.

Tu, S. M., Hossan, E., Amato, R., Kilbourn, R., & Logothetis, C. J. 1995, "Paclitaxel, cisplatin and methotrexate combination chemotherapy is active in the treatment of refractory urothelial malignancies", J.Urol.) vol. 154, no. 5, pp. 1719-1722.

Tulinius, H., Egilsson, V., Olafsdottir, G. H., & Sigvaldason, H. 1992, "Risk of prostate, ovarian, and endometrial cancer among relatives of women with breast cancer", BMJ, vol. 305, no. 6858, pp. 855-857.

Tuzel, E., Sevinc, M., Obuz, F., Sade, M., & Kirkali, Z. 1998, "Is magnetic resonance imaging necessary in the staging of prostate cancer?", Urol Int9 vol. 61, no. 4, pp. 227-231.

274 Uchida, T., Wada, C., Wang, C., Ishida, H., Egawa, S., Yokoyama, E., Ohtani, H., & Koshiba, K. 1995, "Microsatellite instability in prostate cancer", Oncogene, vol. 10, no. 5, pp. 1019-1022.

Uchida, T., Wang, C., Sato, T., Gao, J., Takashima, R., Irie, A., Ohori, M., & Koshiba, K. 1999, "BRCA1 gene mutation and loss of heterozygosity on chromosome 17q21 in primary prostate cancer", Int. J Cancer, vol. 84, no. 1, pp. 19-23.

Ueda, T., Komiya, A., Emi, M., Suzuki, H., Shiraishi, T., Yatani, R., Masai, M., Yasuda, K., & Ito, H. 1997, "Allelic losses on 18q21 are associated with progression and metastasis in human prostate cancer", Genes Chromosomes.Cancer, vol. 20, no. 2, pp. 140-147.

Umbas, R., Isaacs, W. B., Bringuier, P. P., Schaafsma, H. E., Karthaus, H. F., Oosterhof, G. O., Debruyne, F. M., & Schalken, J. A. 1994, "Decreased E- cadherin expression is associated with poor prognosis in patients with prostate cancer", Cancer Res., vol. 54, no. 14, pp. 3929-3933.

Valicenti, R. K., Gomella, L. G., Ismail, M., Strup, S. E., Mulholland, S. G., Dicker, A. P., Petersen, R. O., & Newschaffer, C. J. 1999, "The efficacy of early adjuvant radiation therapy for pT3N0 prostate cancer: a matched-pair analysis", In tJRadiat.OncoLBiol.Phys., vol. 45, no. 1, pp. 53-58. van den Ouden, D., Bentvelsen, F. M., Boeve, E. R., & Schroder, F. H. 1993, "Positive margins after radical prostatectomy: correlation with local recurrence and distant progression", Br.J.UroL, vol. 72, no. 4, pp. 489-494. van den Ouden, D., Hop, W. C., Kranse, R., & Schroder, F. H. 1997, "Tumour control according to pathological variables in patients treated by radical prostatectomy for clinically localized carcinoma of the prostate", Br J UroL, vol. 79, no. 2, pp. 203-211.

Van Poppel, H., De Ridder, D., Elgamal, A. A., Van, d., V, Werbrouck, P., Ackaert, K., Oyen, R., Pittomvils, G., & Baert, L. 1995, "Neoadjuvant hormonal therapy before radical prostatectomy decreases the number of positive surgical margins in stage T2 prostate cancer: interim results of a prospective randomized trial. The Belgian Uro- Oncological Study Group", J.UroL, vol. 154, no. 2 Pt 1, pp. 429-434.

Varambally, S., Dhanasekaran, S. M., Zhou, M., Barrette, T. R., Kumar- Sinha, C., Sanda, M. G., Ghosh, D., Pienta, K. J., Sewalt, R. G., Otte, A. P., Rubin, M. A., & Chinnaiyan, A. M. 2002, "The polycomb group protein EZH2 is involved in progression of prostate cancer", Nature, vol. 419, no. 6907, pp. 624-629.

Veer, L. J., Dai, H., van de Vijver, M. J., He, Y. D., Hart, A. A., Mao, M., Peterse, H. L., van der, K. K., Marton, M. J., Witteveen, A. T., Schreiber, G. J., Kerkhoven, R. M., Roberts, C., Linsley, P. S., Bernards, R., & Friend, S. H. 2002, "Gene expression profiling predicts clinical outcome of breast cancer", Nature, vol. 415, no. 6871, pp. 530-536.

Velasco, A., Hewitt, S. M., Albert, P. S., Hossein, M., Rosenberg, H., Martinez, C., Sagalowsky, A. I., McConnell, J. D., Marston, W., & Leach, F. 275 S. 2002, ’’Differential expression of the mismatch repair gene hMSH2 in malignant prostate tissue is associated with cancer recurrence", Cancer , vol. 94, no. 3, pp. 690-699.

Villers, A., McNeal, J. E., Freiha, F. S., & Stamey, T. A. 1992, "Multiple cancers in the prostate. Morphologic features of clinically recognized versus incidental tumors", Cancer , vol. 70, no. 9, pp. 2313-2318.

Visakorpi, T., Hyytinen, E., Koivisto, P., Tanner, M., Keinanen, R., Palmberg, C., Palotie, A., Tammela, T., Isola, J., & Kallioniemi, O. P. 1995a, "In vivo amplification of the androgen receptor gene and progression of human prostate cancer", N atG en et , vol. 9, no. 4, pp. 401-406.

Visakorpi, T., Kallioniemi, A. H., Syvanen, A. C., Hyytinen, E. R., Karhu, R., Tammela, T., Isola, J. J., & Kallioniemi, O. P. 1995b, "Genetic changes in primary and recurrent prostate cancer by comparative genomic hybridization", Cancer Res., vol. 55, no. 2, pp. 342-347.

Voeller, H. J., Augustus, M., Madike, V., Bova, G. S., Carter, K. C., & Gelmann, E. P. 1997, "Coding region of NKX3.1, a prostate-specific homeobox gene on 8p21, is not mutated in human prostate cancers", Cancer Res., vol. 57, no. 20, pp. 4455-4459.

Voeller, H. J., Truica, C. I., & Gelmann, E. P. 1998, "Beta-catenin mutations in human prostate cancer", Cancer Res., vol. 58, no. 12, pp. 2520-2523.

Wall, D. B., Kachman, M. T., Gong, S., Hinderer, R., Parus, S., Misek, D. E., Hanash, S. M., & Lubman, D. M. 2000, "Isoelectric focusing nonporous RP HPLC: a two-dimensional liquid-phase separation method for mapping of cellular proteins with identification using MALDI-TOF mass spectrometry", AnaLChem ., vol. 72, no. 6, pp. 1099-1111.

Wallen, M. J., Linja, M., Kaartinen, K., Schleutker, J., & Visakorpi, T. 1999, "Androgen receptor gene mutations in hormone-refractory prostate cancer", J.Pathol., vol. 189, no. 4, pp. 559-563.

Wang, S. I., Parsons, R., & Ittmann, M. 1998, "Homozygous deletion of the PTEN tumor suppressor gene in a subset of prostate adenocarcinomas", Clin.Cancer Res., vol. 4, no. 3, pp. 811-815.

Ward, J. F., Zincke, H., Bergstralh, E. J., Slezak, J. M., Myers, R. P., & Blute, M. L. 2004, "The impact of surgical approach (nerve bundle preservation versus wide local excision) on surgical margins and biochemical recurrence following radical prostatectomy", J.Urol., vol. 172, no. 4 Pt 1, pp. 1328-1332.

Wartin, A. 1913, "Heredity with reference to carcinoma.", Arch IntMed, vol. 12, pp. 546-555.

Watson, R. B., Civantos, F., & Soloway, M. S. 1996, "Positive surgical margins with radical prostatectomy: detailed pathological analysis and prognosis", Urology, vol. 48, no. 1, pp. 80-90.

276 Weldon, V. E., Tavel, F. R., Neuwirth, H., & Cohen, R. 1995, ’’Patterns of positive specimen margins and detectable prostate specific antigen after radical perineal prostatectomy”, J.UroL, vol. 153, no. 5, pp. 1565-1569.

Welsh, J. B., Sapinoso, L. M., Su, A. I., Kern, S. G., Wang-Rodriguez, J., Moskaluk, C. A., Frierson, H. F., Jr., & Hampton, G. M. 2001, ’’Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer”, Cancer Res. pp. 5974-5978.

Wick, W., Petersen, I., Schmutzler, R. K., Wolfarth, B., Lenartz, D., Bierhoff, E., Hummerich, J., Muller, D. J., Stangl, A. P., Schramm, J., Wiestler, O. D., & von Deimling, A. 1996, "Evidence for a novel tumor suppressor gene on chromosome 15 associated with progression to a metastatic stage in breast cancer”, Oncogene, vol. 12, no. 5, pp. 973-978.

Wieder, J. A. & Soloway, M. S. 1998, "Incidence, etiology, location, prevention and treatment of positive surgical margins after radical prostatectomy for prostate cancer”, J.Urol., vol. 160, no. 2, pp. 299-315.

Wilkinson, J. & Clapper, M. L. 1997, "Detoxication enzymes and chemoprevention”, Proc.Soc.Exp.Biol.Med., vol. 216, no. 2, pp. 192-200.

Wise, A. M., Stamey, T. A., McNeal, J. E., & Clayton, J. L. 2002, "Morphologic and clinical significance of multifocal prostate cancers in radical prostatectomy specimens”, Urology, vol. 60, no. 2, pp. 264-269.

Witjes, W. P., Schulman, C. C., & Debruyne, F. M. 1997, "Preliminary results of a prospective randomized study comparing radical prostatectomy versus radical prostatectomy associated with neoadjuvant hormonal combination therapy in T2-3 NO M0 prostatic carcinoma. The European Study Group on Neoadjuvant Treatment of Prostate Cancer”, Urology, vol. 49, no. 3A Suppl, pp. 65-69.

Wu, Y. Q., Chen, H., Rubin, M. A., Wojno, K. J., & Cooney, K. A. 2001, "Loss of heterozygosity of the putative prostate cancer susceptibility gene HPC2/ELAC2 is uncommon in sporadic and familial prostate cancer”, Cancer Res., vol. 61, no. 24, pp. 8651-8653.

Xu, J., Meyers, D., Freije, D., Isaacs, S., Wiley, K., Nusskern, D., Ewing, C., Wilkens, E., Bujnovszky, P., Bova, G. S., Walsh, P., Isaacs, W., Schleutker, J., Matikainen, M., Tammela, T., Visakorpi, T., Kallioniemi, O. P., Berry, R., Schaid, D., French, A., McDonnell, S., Schroeder, J., Blute, M., Thibodeau, S., Trent, J., & . 1998, "Evidence for a prostate cancer susceptibility locus on the X chromosome”, Nat.Genet., vol. 20, no. 2, pp. 175-179.

Xu, L. L., Srikantan, V., Sesterhenn, I. A., Augustus, M., Dean, R., Moul, J. W., Carter, K. C., & Srivastava, S. 2000, "Expression profile of an androgen regulated prostate specific homeobox gene NKX3.1 in primary prostate cancer”, J Urol, vol. 163, no. 3, pp. 972-979.

Xu, X., Fu, X. Y., Plate, J., & Chong, A. S. "IFN-gamma induces cell growth inhibition by Fas-mediated apoptosis: requirement of ST ATI protein for up- regulation of Fas and FasL expression”.

277 Yang, X. J., Tan, M. H., Kim, H. L., Ditlev, J. A., Betten, M. W., Png, C. E., Kort, E. J., Futami, K., Furge, K. A., Takahashi, M., Kanayama, H. O., Tan, P. H., Teh, B. S., Luan, C., Wang, K., Pins, M., Tretiakova, M., Anema, J., Kahnoski, R., Nicol, T., Stadler, W., Vogelzang, N. G., Amato, R., Seligson, D., Figlin, R., Belldegrun, A., Rogers, C. G., & Teh, B. T. 2005, ”A molecular classification of papillary renal cell carcinoma", Cancer Res. , vol. 65, no. 13, pp. 5628-5637.

Yeh, S., Miyamoto, H., Nishimura, K., Kang, H., Ludlow, J., Hsiao, P., Wang, C., Su, C., & Chang, C. "Retinoblastoma, a tumor suppressor, is a coactivator for the androgen receptor in human prostate cancer DU145 cells".

Yin, Z., Babaian, R. J., Troncoso, P., Strom, S. S., Spitz, M. R., Caudell, J. J., Stein, J. D., & Kagan, J. 2001, "Limiting the location of putative human prostate cancer tumor suppressor genes on chromosome 18q", Oncogene, vol. 20, no. 18, pp. 2273-2280.

Yoshida, K., Sugino, T., Tahara, H., Woodman, A., Bolodeoku, J., Nargund, V., Fellows, G., Goodison, S., Tahara, E., & Tarin, D. 1997, "Telomerase activity in bladder carcinoma and its implication for noninvasive diagnosis by detection of exfoliated cancer cells in urine", Cancer , vol. 79, no. 2, pp. 362-369.

Yu, K. K. & Hricak, H. 2000, "Imaging prostate cancer", Radiol.Clin.North Am., vol. 38, no. 1, pp. 59-85, viii.

Yu, Y., Yang, J. L., Markovic, B., Jackson, P., Yardley, G., Barrett, J., & Russell, P. J. 1997, "Loss of KAI1 messenger RNA expression in both high- grade and invasive human bladder cancers", Clin.Cancer Res., vol. 3, no. 7, pp. 1045-1049.

Zhang, W., Kapusta, L. R., Slingerland, J. M., & Klotz, L. H. 1998, "Telomerase activity in prostate cancer, prostatic intraepithelial neoplasia, and benign prostatic epithelium", Cancer Res., vol. 58, no. 4, pp. 619-621.

Zhong, H., Chiles, K., Feldser, D., Laughner, E., Hanrahan, C., Georgescu, M. M., Simons, J. W., & Semenza, G. L. "Modulation of hypoxia-inducible factor 1 alpha expression by the epidermal growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells: implications for tumor angiogenesis and therapeutics".

Zincke, H., Oesterling, J. E., Blute, M. L., Bergstralh, E. J., Myers, R. P., & Barrett, D. M. 1994, "Long-term (15 years) results after radical prostatectomy for clinically localized (stage T2c or lower) prostate cancer", J.UroL, vol. 152, no. 5 Pt 2, pp. 1850-1857.

278 Abbreviations bp base pairs Cl Confidence Interval CpG Cytosine phosphodiester Guanine cT clinical tumour stage CT Computerised Tomography CZ Central Zone DNA Deoxyribose Nucleic Acid DRE Digital Rectal Examination MRI Magnetic Resonance Imaging NICE National Institute o f Clinical Excellence PET Positron Emission Tomography PSA Prostate Specific Antigen PT postoperative/pathological tumour stage PZ Peripheral Zone RR Relative Risk RT-PCR Reverse Transcriptase - Polymerase Chain Reaction SD Standard Deviation TNM Tumour Nodes Metastases TRUS Trans Rectal Ultrasound Scan TURP Trans Urethral Resection of the Prostate TZ Transition Zone