| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Investigations |
Cancer Research Program, Garvan Institute of Medical Research [D. I. Q., S. M. H., D. R. H., R. L. S.], and Departments of Urology [D. G., J. D. W., P. C. B., P. D. S.], Anatomical Pathology [J. J. T.], and Medical Oncology [J. J. G.], St. Vincent’s Hospital, Darlinghurst, New South Wales 2010; Douglass Hanly Moir Pathology, North Ryde, New South Wales 2113 [W. D.]; and Sydney Diagnostic Services, North Ryde, New South Wales 2113 [J. F. F.], Australia
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2,
Kruskal-Wallis, and Mann-Whitney tests for clinicopathological
variables and the Kaplan-Meier method, log-rank test, and univariate
and multivariate Cox regression modeling for evaluation of contribution
to relapse and disease-specific survival. At mean follow-up of 55.1
months (range, 4.9–123.0 months), 39% (102 of 263) of patients had
relapsed and 2.3% (6 of 253) had died of PCa. Pretreatment serum
prostate-specific antigen concentration, pathological tumor stage,
lymph node involvement, Gleason score, and p53 nuclear accumulation, as
determined by either percentage score or cluster status, were
independent predictors of relapse in multivariate analysis. Clustering
of p53-positive cells distinguished between favorable and poor
prognosis patients within the lowest p53-positive stratum (>0 to
<2%) and was the most discriminatory threshold for predicting relapse
in the entire cohort. p53 status predicted outcome in patients with a
Gleason score of 5 and above but not those with a score of 4 and below.
In patients treated with neoadjuvant hormonal therapy, p53 cluster
positivity carried a 90% (19 of 21) risk of relapse by 36 months. All
six patients who died from PCa in the period of the study exhibited p53
nuclear accumulation in 20% or more tumor nuclei. This study
demonstrates strong relationships between p53 nuclear accumulation and
relapse and disease-specific mortality in a large series of localized
PCas. Furthermore, the presence of clusters of p53-positive nuclei
delineates a group of patients with poor prognosis not identified by
traditional scoring methods and supports the hypothesis that p53
dysfunction within PCa may exist in foci of tumor cells that are
clonally expanded in metastases. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Inactivation of the tumor suppressor gene p53 is implicated in tumorigenesis for >50% of all human cancers (1) . p53 functions as a transcriptional regulator involved in G1 phase growth arrest of cells in response to DNA damage as well as having roles in the regulation of the spindle checkpoint, centrosome homeostasis, and G2-M phase transition (2) . p53 also induces apoptosis by transcription-dependent and -independent mechanisms in many cell types (1, 2, 3) and regulates tumor angiogenesis and expression of the Kai1 metastasis suppressor gene (1 , 4, 5, 6) . Nuclear accumulation of p53 detected by IHC typically indicates the presence of p53 gene mutations (7 , 8) , although the correlation between nuclear accumulation of p53 and the presence of p53 gene mutations can vary (9) . Nuclear accumulation of p53 is a prognostic indicator in several human cancers, including breast (4 , 10 , 11) , lung (12) , and colorectal carcinoma (13) .
The value of p53 nuclear accumulation as a prognostic factor in localized PCa is controversial. A number of studies have shown that p53 nuclear accumulation detected by IHC is prognostic at a variety of dichotomizing cutoff points based on the number of p53-positive nuclei. These studies describe either a group of poor prognosis patients with 20% p53-positive nuclei (14 , 15) or a group of patients with lower percentages of positive cells in a heterogeneous, focal staining pattern where either the presence of any nuclear accumulation or the presence of clusters of cells showing nuclear accumulation is adversely prognostic (16, 17, 18) . However, other studies comparing p53 nuclear accumulation with assessment of p53 gene mutations have failed to provide conclusive evidence for the importance of p53 in localized PCa or a strong correlation between nuclear accumulation and p53 gene mutation (19, 20, 21, 22) . One study suggested that a chromosome 17p locus close to the p53 gene was an important prognostic feature when both alleles at the site were lost but found that p53 IHC is noncontributory in predicting prognosis and concluded that another gene or genes on chromosome 17p may be involved in PCa progression (19) . In studying other cancers, several authors have suggested that assessment of p53 gene mutations and p53 expression in combination may more accurately define prognostically important p53 dysfunction (9 , 23 , 24) .
Comparison of PCa metastases with primary PCas in the same patients suggest that foci with p53 mutations are clonally expanded in metastases (25 , 26) , perhaps explaining the high frequency of IHC positivity and the presence of gene mutations in hormone-refractory and metastatic PCa (20, 21, 22 , 27 , 28) . Two studies have demonstrated significant heterogeneity in the distribution of p53 mutations between and within foci of carcinoma in the same prostate (29 , 30) . Other studies have documented heterogeneity for other genes and suggested that clones responsible for metastases do not always originate from within the dominant tumor focus (31) . The possibility exists that in localized PCa, p53 overexpression and mutations as well as other genetic aberrations may be limited to subgroups of prognostically important malignant cells.
In the present study, we evaluated p53 nuclear accumulation by IHC in a series of 263 primary PCas. Nuclear accumulation was scored for overall number of positive nuclei and for the presence of clusters of 12 or more cells within a x200 magnification field showing accumulation of p53. Assessment included sections from all identified major foci of cancer within an individual prostate. These findings were correlated with clinicopathological features, including PCa relapse and death from PCa.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Of these 263 patients, 164 were treated with RP alone with no
neoadjuvant or adjuvant therapy of any kind. Thirty-nine patients
received preoperative NHT (34 with goserelin and flutamide, 4 with
flutamide alone, 1 with cyproterone alone) for between 1 and 8 months.
Sixty-nine patients had postoperative adjuvant therapy with hormonal
therapy, radiation therapy, or both (see Table 1
).
|
0.4 ng/ml rising over a
3-month period; local recurrence on DRE confirmed by biopsy or by
subsequent rise in PSA; or institution of long-term hormonal therapy or
orchidectomy. All patients fulfilling the last criterion commenced
long-term hormonal therapy or were treated with orchidectomy based on
pathological features in the prostatectomy specimen.
Pathological Examination.
RP specimens were step sectioned at 2-mm intervals and completely
embedded in paraffin after fixation in neutral-buffered 10% formalin.
Up to three (mean, 1.74; median, 2) blocks from each case were obtained
to provide representative material from the major foci of cancer within
each prostate. The mean number of cancer foci sampled in this way for
each case was 1.96 (median, 2; range, 1–4). The pathological staging
as per the TNM classification, involvement of surgical margins, Gleason
score, and WHO classification for nuclear grade and glandular
differentiation was reported contemporaneously by one of three
histopathologists (J. F. F., W. D., or J. J. T.) and confirmed at review with consensus where reported parameters
differed (32, 33, 34)
.
IHC.
IHC was performed on formalin-fixed paraffin-embedded blocks sectioned
at 5 µm, mounted on SuperFrost Plus slides (Menzel-Glaser,
Braunschweig, Germany) and processed within 10 days of
sectioning. The mouse monoclonal antibody DO-7 (DAKO Corporation,
Carpinteria, CA) and avidin-biotin-peroxidase and diaminobenzidine
kits (Vector Laboratories, Burlingham, CA) were used according to the
manufacturers’ instructions. Briefly, sections were deparaffinized in
xylene, rehydrated through graded ethanol, and then heated in a
pressure cooker in 0.01 M citrate buffer (pH 6.0) for 10
min to enhance antigen retrieval. The sections were then treated with
2% H2O2 for 10 min at room
temperature to inactivate endogenous peroxidase activity. After
a blocking step with 10% normal horse serum, the sections were
incubated with DO-7 antibody diluted to 1:200 in 2% BSA/PBS overnight
at 4°C. Subsequently, sections were sequentially incubated with a
biotinylated horse antimouse IgG, avidin-biotinylated complex, and
diaminobenzidine. Counterstaining was undertaken with Whitlock’s
hematoxylin and light green before dehydration through graded ethanol
and xylene and coverslipping. A contiguous section was stained with
H&E. Positive controls for p53 nuclear accumulation used with each run
of staining included a paraffin-embedded pellet of the PCa cell line
DU145 (35)
, which has a documented p53
mutation; a colon cancer specimen with p53 missense
mutation; and a tongue cancer specimen with p53 nuclear accumulation.
Negative controls included a paraffin-embedded pellet of the PCa cell
line PC3 (35)
, which does not express p53 protein; and the
above described positive controls processed with the substitution of a
non-immune mouse monoclonal antibody for the DO-7 antibody.
Scoring for p53 nuclear accumulation required assessment of all cancer
in selected sections from an individual patient. Counting of a minimum
of 200 cancer cells in each cancer (mean, 812; range, 210-2000 cells)
was undertaken to determine the percentage of nuclei showing
accumulation across all areas of cancer present. The target cell count
was 500 per case where possible, but cancers with fewer cells were not
excluded because of the potential for selection bias. Where the cancers
had multiple foci, or were extensive or heterogeneous, more cells were
counted and selected from areas of varying p53 nuclear accumulation to
provide a sample representative of staining across the entirety of the
cancer. Additionally, assessors scored the cancers cluster positive if
12 or more cancer cell nuclei within any x200 power microscopic field
showed p53 accumulation. In arriving at the cluster definition, we
initially relied on observations of other workers using clusters of 15
cells (18
, 25)
and tested a variety of thresholds for the
definition of a cluster within a range of 6–20 cells in the p53 score
stratum with >0 to <2% p53 nuclear positivity (see below). Below 10
cells per x200 field, there was no correlation with prognosis, and
positive cells tended to be dispersed throughout the cancer rather than
clustered within the area of a single field, whereas there were few
cases with clusters of >15 cells within the >0 to <2% p53
score stratum. This definition differentiated for outcome when between
10 and 15 cells (log-rank, P = 0.04 and
P = 0.05, respectively, within the >0 to
<2% stratum, whereas use of 9 cells produced P = 0.12, and 16 cells produced P = 0.24)
were included as thresholds; therefore 12 cells (log-rank,
P < 0.001) within the field was selected as
a midpoint in this range. Sections were scored independently for p53 by
two assessors (D. I. Q. and S. M. H.) and one
pathologist (J. F. F., W. D., or J. J. T.),
all of whom were blinded to patient outcome. The interobserver Spearman
rank coefficients for p53 score were between 0.92 and 0.96, signifying
close agreement between scorers. The 2 test
for the p53 cluster status initially assigned by different scorers
produced P values of 0.87 and 0.76. Specifically, the two
assessors identified 12 p53 cluster-positive cases not identified on
initial assessment by the pathologist, whereas the pathologist
identified 2 cases not initially identified by the assessors. All of
these cases were deemed cluster positive at consensus review.
Statistical Analysis.
Data were evaluated for relapse and disease-specific mortality
prediction using the Kaplan-Meier product limit method and log-rank
test, and by univariate and multivariate analysis in a Cox proportional
hazards model for p53 cluster status and score, and other recognized
clinical and pathological predictors of outcome (36
, 37)
.
To produce multivariate models relevant to clinical practice, variables
including factors previously described as predictive of outcome and
found to be statistically significant on univariate analysis, but
excluding p53 status, were modeled as dichotomized or continuous
variables to determine their independent prognostic value. Further
modeling with independent variables and p53 score was then undertaken.
All statistical analyses were performed using StatView 4.5 software
(Abacus Systems, Berkeley, CA). Statistical significance in this study
was set at P < 0.05. All reported
P values are two-sided.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Of 263 patients, 115 had organ-confined disease; 139 had
extraprostatic extension, 36 had SVI; 142 had no surgical margin
involvement, whereas 62 had a single margin positive and 58 had
multiple positive margins; and 5 had pelvic lymph node metastases. The
pathological stage and correlation with a series of pathological
variables are shown in Table 2
. Forty-five (17%) patients had well-differentiated, 151 (57%) had
moderately differentiated, and 67 (26%) had poorly differentiated
tumors according to the WHO classification (34)
.
|
. One hundred two (38.8%) patients relapsed during the period
of follow-up (mean, 55.1 months; range, 4.9–123.0 months). Mean time
to relapse was 19.6 months (range, 0.1–59.5 months). Ninety-two
patients who relapsed developed elevated serum PSA measurements in the
period of study either as the first sign of relapse or subsequent to
evidence of local recurrence. Twelve patients developed evidence of
local recurrence either with or without PSA elevation as the first sign
of recurrence, but all developed a simultaneous or subsequent rise in
PSA. Ten patients were deemed to have relapsed on the basis of
commencement of continuous postoperative hormonal therapy or
orchidectomy based on adverse features within the histopathology
report. Omission of these 10 patients from the analyses described below
yielded results essentially the same as for the overall cohort in each
instance. Fourteen patients in the cohort died during the study period,
6 of them from PCa.
Significant predictors of relapse on univariate analysis in this series
of clinically localized PCas treated with RP were preoperative serum
PSA, clinical stage, Gleason score, worst single Gleason grade,
surgical margin involvement, overall pathological stage, pathological
tumor stage, extraprostatic extension, SVI, and lymph node involvement
(see Tables 3
and 4
). Patients undergoing NHT or any form of adjuvant therapy had a
significantly worse prognosis compared with patients treated with RP
alone (see Tables 3
and 4
).
|
|
).
p53 Nuclear Accumulation and Relapse.
The pattern of p53 nuclear accumulation seen was consistent with that
described in previous studies where there was considerable
heterogeneity between different areas of cancer within the same
prostate and within single foci of carcinoma. Homogeneous p53 nuclear
accumulation was unusual, with heterogeneous staining containing foci
of varying size encompassing a variable number of p53-positive cells
being the commonest pattern (Fig. 1)
. The p53 percentage score and p53 cluster status were significantly
correlated with Gleason score, pathological stage, and pretreatment
serum PSA concentration (Table 2)
. Univariate analysis demonstrated
that p53 score as a continuous variable was prognostic (Table 3)
.
|
; overall log-rank, P < 0.0001).
Differences between individual strata were significant by log-rank
calculation between 0 and >0 to <2% (P = 0.01), >0 to <2% and 2 to <5% (P = 0.0002), and 2–5% and 20% (P = 0.009).
|
, and Tables 3
and 4
).
|
5 (log-rank, P < 0.0001) but not for those
with a score 
4 (log-rank, P = 0.61), and
for those cases designated moderately (log-rank, P < 0.0001) and poorly differentiated (log-rank,
P < 0.0001) on WHO criteria but not for
those designated well differentiated (log-rank, P = 0.28). p53 cluster status was statistically discriminatory for
relapse in all pathological staging groups including cases where the
cancer was organ confined
(pT2N0; log-rank,
P < 0.0001).
In comparing the separate scoring systems, all cancers with p53 nuclear
accumulation >2% were also p53 cluster positive. Those cancers with
<2% p53 nuclear accumulation exhibited the presence of clusters in a
proportion of cases with no strict relationship to the overall
percentage of tumor nuclei positive, i.e. a single cluster
could be present as the only p53-positivity in the entire cancer. Given
this observation, the stratum for a p53 nuclear accumulation score
between 0 and 2% was further subdivided according to the presence or
absence of clusters. Subsequently, Kaplan-Meier analysis demonstrated
that cluster-positive cases within this stratum relapsed at a similar
rate to those with 2 to <5% p53 nuclear accumulation (Fig. 1B
; log-rank, P = 0.73; and Fig. 1C
; log rank, P = 0.46), and this
was confirmed by univariate analysis (Tables 3
and 4)
and maintained in
multivariate models (Table 5)
. There was no significant difference in outcome between those cases
deemed cluster negative within the >0 to <2% stratum and those in
the 0 stratum (log-rank, P = 0.19), although
those patients within the 0 stratum had significantly better
relapse-free survival compared with all other patients,
i.e. those with any p53 nuclear accumulation (log-rank,
P < 0.0001).
|
. When stepwise multivariate
analyses were constructed with backward elimination in the Cox
proportional hazard regression model, the factors most predictive of
relapse were, in descending order: lymph node involvement, Gleason
score, p53 status measured by either score or cluster status,
pretreatment PSA less than or greater than 10 ng/ml, and pathological
stage pT3 or greater against
pT2.
p53 Nuclear Accumulation and Relapse in Patients Treated with NHT.
Thirty-nine patients received NHT prior to RP, 23 of whom relapsed, and
1 of whom died of PCa during the study period. p53 score (Mann-Whitney,
P = 0.0008), p53 strata
(2, P = 0.005), and
p53 cluster status (Fisher’s exact test, P = 0.0002) predicted relapse in the NHT group. Ninety percent (19 of 21)
of NHT patients who were p53 cluster positive relapsed within 36 months
of prostatectomy (Fig. 4
; log-rank, P < 0.0001) compared with 22% (4
of 18) of those who were p53 cluster negative. Univariate analysis in
this group revealed that p53 cluster status and pretreatment PSA
concentration were significant predictors of outcome (Table 6)
. Multivariate analysis extended to include Gleason score stratified at
the 4–7 and 8–10 levels and pathological T stage stratified between
pT2 and pT3 or greater
showed that all were significant predictors of outcome in the model
(Table 6)
. However, with stepwise regression analysis, the factors most
predictive of relapse were, in descending order: p53 cluster
positivity, pretreatment PSA level, Gleason score of 8–10, and
pathological T stage. In bivariate analysis using each of these
variables with p53 cluster status, pretreatment PSA level maintained
statistical significance, whereas Gleason score and pathological T
stage did not. Hence, p53 cluster status was the strongest predictor of
outcome in patients treated with NHT prior to RP.
|
|
20%. Kaplan-Meier analysis of the 20% stratum against the
<20% strata was highly significant in predicting death from
PCa (log-rank, P < 0.0001; Fig. 4
).
Evaluation of p53 status in predicting overall survival failed to reach
significance (data not shown). |
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The novel findings of this study are
as follows: that increasing p53 score carries with it an increased risk
of relapse and death; that the presence of a cluster of p53-positive
tumor cells within a x200 magnification field provides a
discriminatory point predictive of relapse; and that p53 nuclear
accumulation is the most powerful predictor of outcome in patients
given NHT prior to RP. These data support the findings of previous
studies that demonstrated relationships between outcome and p53 IHC at
thresholds of no nuclear accumulation versus any
accumulation (16
, 17)
, the presence of clusters of
p53-positive cells versus their absence (18
, 42)
, and between 20% p53-positive nuclei and <20%
(14
, 15)
. In predicting early death from PCa following RP,
p53 positivity in 20% of nuclei defines a group of patients with
highly aggressive disease that progresses much more rapidly than is the
case for most patients experiencing PSA relapse (41)
. Our
data also support previously described relationships between p53
nuclear accumulation and pathological tumor stage and Gleason score,
and define a positive correlation between p53 score and pretreatment
serum PSA concentration. These data run counter to a number of studies
that found p53 nuclear accumulation to be uncommon and/or nonprognostic
in localized PCa (19, 20, 21, 22)
.
p53 status was not predictive of outcome in patients with better
differentiated tumors, i.e. those that were well
differentiated or with a Gleason score 4. The outcome following
surgery for patients with better differentiated tumors was good, and
there was a low event rate in these groups. It may be that an effect of
p53 clustering has not been apparent within our follow-up period.
Alternatively, it may be that these cancers are intrinsically indolent
and that p53 status is not of importance within this subset because of
a lack of other genetic and epigenetic factors contributing to cancer
progression. Conversely, p53 status predicted outcome in all subgroups
based on pathological stage and pretreatment serum PSA strata. One
potential limitation of our study is that although biochemical relapse
correlates with subsequent development of clinical metastases, the rate
of development of these metastases varies greatly; further evaluation
of this and other cohorts is needed to determine whether p53 status
predicts clinical relapse.
p53 cluster status was the most predictive factor in determining
outcome in patients who received NHT prior to RP (Table 6)
. Grignon
et al. (15)
, in reporting on 129 patients with
clinically localized PCa treated with radiation therapy in the RTOG
8610 trial, found that p53 nuclear accumulation in pretreatment
diagnostic material predicted reduced time to distant metastases in
those patient given NHT but not in those treated with radiation therapy
alone. The predictive value of p53 IHC in the NHT group raises
interesting questions about the effect of hormonal therapy on PCa cells
with and without p53 nuclear accumulation. Following androgen ablation,
a small proportion of PCa cells undergo apoptosis, but the majority of
cells responding undergo cell cycle arrest mediated, at least in part,
by p53, entering G0 phase and losing cell volume
(43
, 44) . Cells with dysfunctional p53 may be resistant to
hormonal therapy and fail to undergo cell cycle arrest or apoptosis
(15
, 45) , conferring a relative growth advantage and
greater prominence when evaluated for p53 nuclear accumulation. Such a
postulated mechanism may explain the observation that p53 dysfunction
is associated with hormonal resistance in some studies of breast cancer
and PCa (15
, 45)
.
Our study reports a higher rate of p53 positivity than some previous
studies, with 79% of cases showing at least occasional p53 positivity
and 52% being p53 cluster positive (Table 2)
. The most obvious
explanation for this relates to a modified definition of p53 positivity
that integrates p53 cluster status into the p53 accumulation score.
Another potential reason is the inclusion of a greater proportion of
patients with adverse features not as prevalent in other series, such
as pathological stage (57%
pT3N0 or greater), tumor
differentiation (16% with Gleason score 8 and 26% poorly
differentiated by WHO criteria), and pretreatment serum PSA
concentration (25% greater than 20 ng/ml), that contribute to a
relapse rate of 38.8% at the mean follow-up of 55.1 months. In a study
of 175 RP cases evaluated for p53 nuclear accumulation, 65% of cases
were reported to demonstrate occasional or greater (i.e.
>0) p53 positivity (17)
with a corresponding relapse of
37.7% at a mean follow-up of 55.2 months. Within that series
(17)
, 57.1% of cases were pathological stage
pT3N0 or greater, 10.8%
had a Gleason score 8, and 5.7% were poorly differentiated;
pretreatment serum PSA concentrations were not reported. This suggests
that although the stage of tumors studied was similar in both studies,
the tumors in our study had a higher Gleason score, were more poorly
differentiated on WHO criteria, and had a slightly higher chance of
relapse. The inclusion of patients receiving neoadjuvant and adjuvant
therapy is likely to have resulted in a cohort with relatively more
aggressive or advanced cancer as indicated by pathological stage,
Gleason score, and preoperative serum PSA concentrations. Further
comparison between our cohort and a recently reported large
single-surgeon case series, where 54.4% of cases were pathological
stage pT3N0 or greater,
7.8% had a Gleason score 8, and 5.5% of patients had a pretreatment
serum PSA concentration 20 ng/ml, also suggests that our group
contains patients with more adverse features than others
(41)
. Given these differences and the trend toward
screening-detected cancers treated with RP being of lower stage,
further evaluation of p53 nuclear accumulation in cohorts with less
aggressive features is desirable. Finally, it may be that the antigen
retrieval and IHC techniques used in our laboratory are more sensitive
than those reported in some other studies.
Methodological factors may also have contributed to our higher p53-positivity rate because of the assessment of all major foci within each individual prostate and the use of a p53 antibody directed at the DO7 epitope, which has a high correlation with p53 gene mutation compared with other antibodies (46) . Other researchers have found that a cocktail of DO1 and DO7 epitope-directed antibodies is slightly more sensitive than DO7 alone (47) . We achieved equivalent results with DO7 antibody, which detects both wild-type and mutant p53 protein, having noted increased background staining in initial experiments with the described cocktail (data not shown). In determining p53 status, it appears that IHC is sensitive but may suffer from lack of specificity for detecting mutation, whereas methods of direct sequencing are highly specific in detecting p53 mutation but may lack sensitivity when such mutations are present in only a small percentage of cells within a population containing wild-type p53. This may in part explain differences in p53 status based on the technique used.
Accumulation of p53 protein may occur in response to a number of stimuli independent of p53 gene mutation, including DNA damage, hypoxia, and redox stress (48) . Elledge et al. (49 , 50) have suggested that even low levels of p53 protein accumulation are prognostic in breast cancer specimens regardless of whether p53 gene mutation could be detected concurrently, whereas Silvestrini et al. (51) have shown that an increasing percentage of p53-positive nuclei between 0 and 12% has a corresponding adverse "dose" effect on prognosis in a large series of breast cancer patients. Our study describes a similar relationship between percentage of p53-positive nuclei and relapse in PCa, but it also illustrates that clusters of p53-positive cells are prognostic when <2% of all tumor cells are positive. The prognostic significance of p53 clusters suggests that cells in such clusters have important characteristics such as p53 mutation or are affected by other factors capable of producing local p53 nuclear accumulation and are associated with a poorer outcome. In this regard, p53 clusters may represent foci of cells with p53 gene mutations that expand within the prostate to increase the p53 score with time so that more advanced and/or higher Gleason grade cancer is associated with a higher p53 score. It also is possible that micrometastases occur at a similar time in PCa progression to the development of clusters with a critical number of cells showing p53 nuclear accumulation. This hypothesis suggests a focal dose-response threshold for p53 dysfunction and genesis of metastases (see below).
PCa is usually multifocal. The clone responsible for human PCa metastases may reside in smaller tumor foci within the prostate gland rather than the largest (31 , 52 , 53) . Heterogeneity between and within different foci of PCa within the same gland is inferred by studies of DNA flow cytometry (54) and allelic loss (52 , 55) . Heterogeneity of p53 mutations and p53 protein nuclear accumulation within and between individual foci of early-stage localized cancer in the same prostate has been demonstrated previously (29 , 30) . p53 is crucial in determining metastatic potential in models of skin carcinogenesis where p53 protein dose does not contribute to initiation and/or promotion but is associated with metastasis (56) . Similarly, one study found that in ras+myc-initiated prostate carcinoma metastasis is concurrent with loss of expression of the wild-type p53 allele (57) . In that study, comparative DNA analysis between primary tumors and metastases demonstrated that metastases did not necessarily derive from the most abundant clone but seeded from small subpopulations from within the primary tumor (57) . The clonal expansion of p53 dysfunctional cells (57) has been confirmed in small series of primary human PCas and metastases within the same patients (25 , 26 , 58) . These studies add to others that demonstrate increased p53 nuclear accumulation in metastatic, recurrent, and/or androgen-insensitive PCa compared with clinically localized disease (20, 21, 22 , 27 , 28) . In a study of 50 metastatic PCa foci in 19 men with lethal PCa, p53 gene mutations demonstrated homogeneity for mutation in virtually all metastases assessed from individual patients, in contrast to PTEN/MMAC1 mutations, which demonstrated intermetastasis heterogeneity (59) . Taken together, these studies suggest that prostate tumor cells harboring p53 mutations and perhaps other genetic aberrations are clonally expanded in metastases. Our study extends this concept by suggesting that a "p53 dose" effect across the entire cancer and at a given focal threshold within clusters is important in the metastatic process.
The work presented in this study and that of others (16, 17, 18 , 60) suggests that p53 has an important prognostic role in approximately half of all patients with clinical localized PCa. Given the finding that p53 is the most important predictor of outcome in patients given NHT before RP and similar observations in patients given NHT prior to radiation therapy (15) , p53 status may have significant implications for patients treated with hormonal therapy as part of these regimens. The genetic and epigenetic basis for clustered p53 nuclear accumulation and its effect on clinical as well as biochemical relapse require further investigation.
|
FOOTNOTES |
|---|
1 This research was supported by grants from the
National Health and Medical Research Council of Australia, RT Hall
Trust, Laurence Freedman Trust, New South Wales Cancer Council, Leo and
Jenny Leukemia, and Cancer Foundation of Australia, St. Vincent’s
Clinic Foundation, and Merck Sharp and Dohme Research Foundation.
D. Q. is a National Health and Medical Research Council of
Australia Medical Postgraduate Research Scholar and recipient of the
Vincent Fairfax Family Foundation Fellowship from the Royal
Australasian College of Physicians. 
3 To whom requests for reprints should be
addressed, at Cancer Research Program, Garvan Institute of Medical
Research, 384 Victoria Street, Darlinghurst, NSW 2010, Australia.
Phone: 612-9295-8322; Fax: 612-9295-8321; E-mail: r.sutherland{at}garvan.unsw.edu.au
4 The abbreviations used are: PCa, prostate
cancer; IHC, immunohistochemistry; PSA, prostate-specific antigen; RP,
radical prostatectomy; NHT, neoadjuvant hormonal therapy; DRE, digital
rectal examination; TNM, Tumor-Node-Metastasis; SVI, seminal vesicle
involvement.
Received 9/27/99. Accepted 1/19/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L.-Y. Khor, K. Bae, R. Paulus, T. Al-Saleem, M. E. Hammond, D. J. Grignon, M. Che, V. Venkatesan, R. W. Byhardt, M. Rotman, et al. MDM2 and Ki-67 Predict for Distant Metastasis and Mortality in Men Treated With Radiotherapy and Androgen Deprivation for Prostate Cancer: RTOG 92-02 J. Clin. Oncol., July 1, 2009; 27(19): 3177 - 3184. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. K. Rasiah, J. G. Kench, M. Gardiner-Garden, A. V. Biankin, D. Golovsky, P. C. Brenner, R. Kooner, G. F. O'Neill, J. J. Turner, W. Delprado, et al. Aberrant neuropeptide y and macrophage inhibitory cytokine-1 expression are early events in prostate cancer development and are associated with poor prognosis. Cancer Epidemiol. Biomarkers Prev., April 1, 2006; 15(4): 711 - 716. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Xia, S. R. Kumar, R. Masood, S. Zhu, R. Reddy, V. Krasnoperov, D. I. Quinn, S. M. Henshall, R. L. Sutherland, J. K. Pinski, et al. EphB4 Expression and Biological Significance in Prostate Cancer Cancer Res., June 1, 2005; 65(11): 4623 - 4632. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Ray, M. Ho, J. Ma, R. K. Parkes, T. G. Mainprize, S. Ueda, J. McLaughlin, E. Bouffet, J. T. Rutka, and C. E. Hawkins A Clinicobiological Model Predicting Survival in Medulloblastoma Clin. Cancer Res., November 15, 2004; 10(22): 7613 - 7620. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Foley, D. Hollywood, and M. Lawler Molecular pathology of prostate cancer: the key to identifying new biomarkers of disease Endocr. Relat. Cancer, September 1, 2004; 11(3): 477 - 488. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Ren, L. Li, A. A. Goltsov, T. L. Timme, S. A. Tahir, J. Wang, L. Garza, A. C. Chinault, and T. C. Thompson mRTVP-1, a Novel p53 Target Gene with Proapoptotic Activities Mol. Cell. Biol., May 15, 2002; 22(10): 3345 - 3357. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. F. Pollack, S. D. Finkelstein, J. Woods, J. Burnham, E. J. Holmes, R. L. Hamilton, A. J. Yates, J. M. Boyett, J. L. Finlay, R. Sposto, et al. Expression of p53 and Prognosis in Children with Malignant Gliomas N. Engl. J. Med., February 7, 2002; 346(6): 420 - 427. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. G. Horvath, S. M. Henshall, C-S. Lee, D. R. Head, D. I. Quinn, S. Makela, W. Delprado, D. Golovsky, P. C. Brenner, G. O'Neill, et al. Frequent Loss of Estrogen Receptor-{beta} Expression in Prostate Cancer Cancer Res., July 1, 2001; 61(14): 5331 - 5335. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Henshall, D. I. Quinn, C. S. Lee, D. R. Head, D. Golovsky, P. C. Brenner, W. Delprado, P. D. Stricker, J. J. Grygiel, and R. L. Sutherland Altered Expression of Androgen Receptor in the Malignant Epithelium and Adjacent Stroma Is Associated with Early Relapse in Prostate Cancer Cancer Res., January 1, 2001; 61(2): 423 - 427. [Abstract] [Full Text]
|
|
|
|
 |
|
 J. R. Thornbury, D. K. Ornstein, P. L. Choyke, C. P. Langlotz, and J. C. Weinreb Prostate Cancer: What Is the Future Role for Imaging? Am. J. Roentgenol., January 1, 2001; 176(1): 17 - 22. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
); 
); 
); 
) of immunoreactive nuclei. Survival curves were generated
according to the Kaplan-Meier method, and statistical comparisons were
made by use of the log-rank method (n = 263; overall log-rank, P < 0.0001;
intergroup log-rank: 0 versus >0 to <2%,
P = 0.01; >0 to <2%
versus 
); >0 to <2% and
cluster positive (
); 
) of immunoreactive nuclei (n = 263, overall log-rank, P < 0.0001; intergroup log-rank: 0 versus >0 to <2%,
cluster negative, P = 0.19; >0 to <2%,
cluster negative versus >0 to <2%, cluster positive,
P = 0.0004; >0 to <2%, cluster
positive versus 
