| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Article |
From the Clinical Cancer Genetics and Human Cancer Genetics
Programs,*
Comprehensive Cancer Center, and the Division of
Human Genetics, Department of Internal Medicine, The Ohio State
University, Columbus, Ohio; the Department of Obstetrics and
Gynecology,
Nippon Medical School, Tokyo,
Japan; the Division of Hematology/Oncology,
Beth Israel Deaconess Medical Center, Harvard Medical School, Boston,
Massachusetts; the Department of Paediatrics and Child
Health,
Section of Medical and Molecular
Genetics, University of Birmingham Medical School, Birmingham, United
Kingdom; and the Cancer Research Campaign Human Cancer Genetics
Research Group,¶
University of Cambridge,
Cambridge, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
30 to
50%, and the somatic intragenic PTEN mutation
frequency is <10%. In this study, we screened primary
adenocarcinomas of the ovary for LOH of polymorphic markers within and
flanking the PTEN gene and for intragenic mutations of
the PTEN gene and compared them to PTEN expression using
immunohistochemistry. Furthermore, we sought to detect the
expression of the presumed downstream targets of PTEN, such as
P-Akt, p27, and cyclin D1 by immunohistochemistry. LOH
at 10q23 was observed in 29 of 64 (45%) cases. Of the 117
samples, 6 somatic intragenic PTEN
mutations, 1 germline mutation, and 1 novel
polymorphism were found in 7 (6%) patients. Immunostaining of 49
ovarian cancer samples revealed that 13 (27%) were PTEN
immunostain-negative, 25 (51%) had reduced staining,
and the rest (22%) were PTEN expression-positive. Among the 44
informative tumors assessed for 10q23 LOH and PTEN
immunostaining, there was an association between 10q23 LOH and
decreased or absent staining (P = 0.0317). Of
note, there were five (11%) tumors with neither mutation nor
deletion that exhibited no PTEN expression and 10 (25%) others without
mutation or deletion but had decreased PTEN expression. Among the 49
tumors available for immunohistochemistry, 28 (57%) showed
P-Akt-positive staining, 24 (49%) had decreased p27
staining, and cyclin D1 was overexpressed in 35 (79%) cases.
In general, P-Akt expression was inversely correlated with PTEN
expression (P = 0.0083). These data suggest that
disruption of PTEN by several mechanisms,
allelic loss, intragenic mutation, or epigenetic
silencing, all contribute to epithelial ovarian
carcinogenesis, and that epigenetic silencing is a significant
mechanism. The Akt pathway is prominently involved, but clearly
not in all cases. Surprisingly, despite in vitro
demonstration that p27 and cyclin D1 lies downstream of PTEN and
Akt, there was no correlation between p27 and cyclin D1
expression and PTEN or P-Akt status. Thus, in
vivo, although PTEN and Akt play a prominent role in
ovarian carcinogenesis, p27 and cyclin D1 might not be the
primary downstream targets. Alternatively, these observations
could also suggest that pathways involving other than Akt, p27
and cyclin D1 that lie downstream of PTEN play roles in ovarian
carcinogenesis.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The tumor suppressor gene PTEN/MMAC1/TEP1, encoding a dual-specificity phosphatase, has been cloned and mapped to chromosome subband 10q23.3.11-13 Germline PTEN mutations have been found in the inherited autosomal-dominant Cowden and Bannayan-Riley-Ruvalcaba syndromes, which are characterized by multiple hamartomas and by an increased risk of malignant and benign breast and thyroid tumors.14-19 Recently, a Proteus-like syndrome was found to result from germline and germline mosaic PTEN mutations.20 Ovarian cancer, in general, is not considered part of these syndromes. Somatic mutation and/or deletion of PTEN occurs to a greater or lesser extent in a wide variety of human cancers that show LOH in this region, including glioblastoma, endometrial cancer, prostate cancer, and breast cancer.12,13,21-24
Genetic, functional, and animal modeling studies have substantiated the tumor suppressor function of PTEN. PTEN is a lipid phosphatase whose major substrate is phosphatidylinositol-3,4,5-triphosphate [PtdIns(3,4,5)P3], downstream of which lies the Akt (PKB) pathway.25-29 The serine-threonine kinase Akt, when phosphorylated, protects cells from apoptosis.30,31 PTEN may also be involved in cell migration, spreading, and focal adhesion formation through dephosphorylating focal adhesion kinase, presumably through its protein phosphatase activity.32,33 Ectopic expression of PTEN results in cell-cycle arrest at G1 and/or apoptosis at least in the glioma and breast cancer cell line models.27,28,34 It has been demonstrated that this function of PTEN is Akt-dependent, and can be rescued by expression of phosphorylated Akt.27,35 One of the targets of PTEN in its ability to block cell-cycle progression at the G1 phase has been suggested to be the cyclin-dependent kinase inhibitor p27.28,36-38 It has been shown that up-regulation of p27 by PTEN has occurred in different cell lines.36,37 However, it remains to be established whether the increase of p27 expression is truly a direct effect of PTEN's action in vivo and which intermediate steps are involved in synthesis of this cell-cycle inhibition signal. Another potential mechanism of cell-cycle control by PTEN may be through inhibition of cyclin D1 accumulation. AKT phosphorylates GSK3 (glycogen synthase kinase 3), leading to its inactivation.39 Active GSK3 phosphorylates cyclin D1, targeting it for degradation.40 Therefore, Akt seems to promote cyclin D1 accumulation.41,42
Although only rare mutations of the PTEN gene were reported in ovarian cancer,8-10,43-46 because of PTEN's role in the cell cycle and cell death as well as the gene’s localization to 10q23, PTEN is an excellent candidate to play an important role in ovarian carcinogenesis. Therefore, we sought to determine whether structural alterations in PTEN occurred with any frequency in ovarian cancer, if loss of PTEN expression, detected by immunohistochemistry, is a major mechanism of loss of function, and if there is a correlation between structural alterations of PTEN gene and PTEN protein expression. Further, we looked for any alteration in the expression of the presumed downstream targets of PTEN, such as P-Akt, p27, and cyclin D1 by immunohistochemistry, and investigated the correlation among all these variables.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One hundred seventeen epithelial ovarian tumors were obtained from patients undergoing surgery for primary epithelial ovarian cancer. Thirty-eight tumors were obtained from The Ohio State University, Columbus, OH (OSU), 25 were from the Beth Israel-Deaconess Medical Center, Boston, MA (BOS), 31 were from the University of Birmingham, UK (UK), and 23 were from Nippon Medical School, Tokyo, Japan (NMS). Apart from the 31 ovarian carcinomas from the UK that are of unknown histological sub-type, 32 were serous tumors, 28 were endometrioid cancers, 7 were clear cell cancers, 6 were mucinous tumors, 6 were mixed epithelial tumors, 6 were undifferentiated tumors, and 1 was a squamous cell carcinoma. Corresponding noncancerous tissues were available only from OSU, BOS, and NMS samples. Corresponding paraffin blocks were available only from OSU and BOS samples. Genomic DNA was extracted from tumor and matched normal tissue with a QIAamp DNA Mini Kit (Qiagen, Valencia, CA) following the manufacturer’s instructions, or by microdissection of normal and tumor areas of paraffin-embedded archival tissue using standard protocols.47
LOH Analysis
Sixty-eight ovarian cancers (35 from OSU, 11 from BOS, and 22 from NMS) in which both tumor and corresponding normal DNA were available were analyzed for LOH at five polymorphic markers flanking and within the PTEN gene. The markers are ordered from centromere to telomere: D10S579-D10S1765-IVS4 + 109ins/delTCTTA-IVS8 + 32T/G-D10S541. PTEN lies between D10S1765 and D10S541, a genetic distance of 1 cM but a physical distance of only several hundred kilobase pairs. D10S1765 is within 500 kb upstream of the transcriptional start site of PTEN and D10S541 is within 300 kb of the translational stop site. IVS4 + 109ins/delTCTTA and IVS8 + 32T/G lie within PTEN.19,48,49 The IVS4 + 109ins/delTCTTA and IVS8 + 32T/G polymorphisms were screened for by differential digestion with AflII and HincII, respectively, according to the manufacturer’s guidelines (New England Biolabs, Beverly, MA) as described previously.16,48 D10S579, D10S1765, and D10S541 were screened as documented previously.50 Polymerase chain reaction (PCR) conditions for these markers are described elsewhere.51,52
PTEN Mutation Analysis
All samples were scanned for mutations by denaturing gradient gel electrophoresis (DGGE) as previously described.24 The entire PTEN-coding region, exon-intron boundaries, and flanking intronic sequences were PCR amplified and fractionated through denaturing gradient gels according to the conditions described previously.24,51 Samples showing DGGE variation were re-amplified with another set of primers, specifically for sequence analysis, gel- and column-purified and subjected to semi-automated sequence analysis as previously published.48
RNA Extraction and Reverse Transcriptase (RT)-PCR
From 100 mg of ovarian cancer tissue, total RNA was extracted
by the guanidine thiocyanate method,53
using Trizol
reagent (Life Technologies, Inc., Gaithersburg, MD) according to the
manufacturer’s guidelines. Total RNA was treated with DNase I
(Boehringer Mannheim, Mannheim, Germany) to remove any contaminating
genomic DNA before reverse transcription. This RNA was
reverse-transcribed with a first strand cDNA synthesis kit for RT-PCR
(AMV; Boehringer Mannheim) according to the manufacturer’s
recommendations. RT-PCR using the PTEN exonic primers [1F,
5'-TCAAGAGGATGGATTCGACTT-3' (PTEN exon 1), and 5.1R,
5'-TCATTACACCAGTTCGTCC-3' (PTEN exon 5)] was
performed to determine whether splicing abnormalities would result from
intronic mutation detected in the sequencing analysis.
Immunohistochemistry
The monoclonal antibody 6H2.1 raised against the last 100 C-terminal amino acids of PTEN54 was used. Specificity of this antibody for PTEN has been documented elsewhere.50,54 Polyclonal antibody against P-Akt was obtained from Upstate Biotechnology (Lake Placid, NY). Monoclonal antibodies against p27 and cyclin D1 were obtained from Transduction Laboratories (San Diego, CA) and Santa Cruz Biotechnology (Santa Cruz, CA), respectively.
The tissue samples were fixed by immersion in buffered formalin and embedded in paraffin according to standard procedures.47 Sections (4 µm) were cut and mounted on Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA). Subsequently, the sections were deparaffinized and hydrated by passing through xylene and a graded series of ethanol. Antigen retrieval was performed for 20 minutes at 98°C in 0.01 mol/L sodium citrate buffer, pH 6.4, in a microwave oven. To block endogenous peroxidase activity, the sections were incubated with 0.3% hydrogen peroxide for 30 minutes. After blocking for 30 minutes in 0.75% normal serum, the sections were incubated with each antibody overnight at 4°C. The sections were washed in phosphate-buffered saline and then incubated with biotinylated second antibody followed by avidin peroxidase using the Vectastain ABC elite kit (Vector Laboratories, Burlingame, CA). The chromogenic reaction was performed with 3–3' diaminobenzidine using nickel cobalt amplification.55 After counterstaining with Nuclear Fast Red (Rowley Biochemical, Danvers, MA) and mounting, the slides were evaluated under a light microscope. The immunostaining patterns and intensities were independently evaluated by two investigators (KK and X-PZ) and randomly spot evaluated by a third investigator (CE). The multiple independent reading of the slides was performed to ensure quality control and consistency of results. Further, all sections, especially those that are immunostain-negative, were checked to contain internal positive and negative staining controls. Such consistency and accuracy have been previously documented over a broad range of tissue types throughout time 24,50,54,56,57 .
Immunohistochemical Analyses
Immunostaining was evaluated without knowledge of the clinical and pathological parameters. For PTEN immunostaining, according to the amount of staining, the tumors were divided into three groups: the group assigned ++ showed increased or equal staining intensity compared to the corresponding normal tissue; the group assigned + had decreased staining intensity; and the group assigned - had no trace of staining. For P-Akt, the samples were divided into two groups according to the amount of staining. The group assigned + showed increased staining intensity compared to the corresponding normal tissue and the group assigned - had no staining. For p27, positive samples were scored according to the frequency of immunopositive cells as <5%, 5 to 50%, >50% cells immunopositive. Samples from patients with <50% p27-positive tumor cells were considered low expressors, whereas those with >50% p27-positive tumor cells were considered high expressors according to the published conventions.58,59 We also examined cyclin D1 immunostaining. We scored cyclin D1 immunostaining by the percentage of total cells that were positive in both the nucleus and cytoplasm. Slides were graded as: negative (0 to 10% cells stained), + (10 to 50% cells stained), ++ (>50% cells stained). A cutoff value of 10% staining was used to separate normal staining (<10%) from cyclin D1-overexpressing cells (>10%).
The Fisher exact test and Wilcoxon rank sum test were used for statistical analysis of the results. The nonparametric Wilcoxon test was also chosen because immunostaining intensities (especially for cyclin D1 and p27) may be considered a continuous variable. A P < 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We analyzed 68 ovarian cancers (35 from OSU, 11 from BOS, and 22 from NMS), for which tumor and corresponding normal DNA were available, for LOH at five polymorphic markers flanking PTEN (D10S579, D10S1765, and D10S541) and within PTEN (IVS4 + 109ins/delTCTTA and IVS8 + 32G/T). Among the 68 samples examined, 4 were excluded from final analysis because they were not informative at all markers. LOH at 10q23 was scored when one or more of the panel of five polymorphic loci showed LOH, as is standard. Of the 64 informative samples that were informative for at least one marker, 29 (45%) had LOH at one or more loci.
PTEN Mutation Analysis
To determine whether PTEN is genetically altered during
ovarian carcinogenesis, we screened primary ovarian cancers for
mutations in the PTEN gene by DGGE analysis. Samples that
showed variant bands on DGGE analysis were used as template for direct
sequence analysis. Of the 117 total samples, DGGE analysis revealed 8
variants in 7 samples (6%), and sequence analysis of these samples
confirmed the presence of sequence variation. The identified variants
included three missense, one nonsense, two frameshift mutations, and
two intronic variants (Table 1)
. The UK
samples showed the most frequent mutations (4 of 31, 13%), and
Japanese and United States samples harbored mutations in 2 of 23 (9%)
and 1 of 63 (2%), respectively. There was a statistical significance
in mutation frequency between UK and US samples
(P = 0.022, chi-square test). The one missense
mutation in tumor NMS26 was somatic, and the corresponding germline was
wild type. Tumors NMS2 and BOS11 had intronic variants in both tumor
and corresponding germline DNA. Germline and somatic DNA from tumor
NMS2 showed a C-to-G change in intron 1, 41 bases from the exon-intron
boundary (IVS1 + 41C>G). Germline and somatic DNA from BOS11 revealed
a G-to-T change in intron 3, one base from exon-intron boundary
(IVS3–1G>T). Corresponding germline DNA of the remaining samples that
showed variants were not available.
|
. PTEN Immunohistochemistry in Ovarian Cancer
Samples from 49 sporadic ovarian epithelial carcinomas (38 from
OSU and 11 from BOS), which had paraffin-embedded sections available,
were examined for PTEN expression using the monoclonal antibody 6H2.1,
raised against the terminal 100 amino acids of human PTEN. Vascular
endothelial cells showed strong PTEN expression with a nuclear
predominance and served as an internal positive control for this study
(Figure 1, A and C)
.50,54
In
contrast, nuclear- and cytoplasmic-staining intensity of fibrocytes was
very heterogeneous and varied from strong to weak. Of 49 ovarian cancer
samples, 13 (27%) lost all PTEN immunoreactivity and was graded
- (Figure 1, A and B)
. Twenty-five of the 49 (51%) ovarian cancer
specimens stained weakly, graded +, in comparison to the normal tissue
(Figure 1C)
. The remaining 11 (22%) tumors stained positively, graded
++ (increased or equal staining intensity compared to the corresponding
normal tissue) (Figure 1D)
. In general, the quality and intensity
of PTEN immunostaining in the nucleus and cytoplasm was relatively
uniform throughout each specimen.
|
There were a total of 46 samples in which both immunohistochemical
data and LOH data are available. Among these 46 samples, LOH data from
44 tumors were informative. Ten (23%) of the 44 samples were graded ++
for PTEN expression by immunohistochemistry, and 8 of these 10 did not
show evidence of 10q23 LOH (Table 2
and
Figure 2
). The remaining two samples
(OSU50 and BOS19) might be viewed as apparently discordant when LOH
data are compared to those obtained from immunohistochemistry (Figure 2)
. Sample OSU50 showed LOH at D10S1765 (5' of PTEN) and
retention of both alleles at the remaining four loci, notably those
within PTEN. Sample BOS19 showed LOH at D10S579 (5' of
PTEN), retention of both alleles at D10S1765, and
noninformative at the remaining three loci (Figure 2)
. Because these
samples showed LOH only 5' of PTEN, it is doubtful whether
it is deleted within PTEN, thus explaining the ++
immunohistochemistry.
|
|
and Figure 2
). Four of
these five samples (OSU14, OSU28, OSU58, and OSU68) showed retention of
both alleles at the flanking markers, and noninformative at the
intragenic PTEN polymorphic markers. Therefore, there is a
possibility that a part or entire PTEN might be deleted
(Figure 2)
.
Twenty-one samples (48%) were immunostain-positive. Of these 21, 11
might be classified as having LOH representing hemizygous
PTEN allelic loss that could correspond to the diminished
immunostaining. The remaining 10 with weak immunoreactivity retained
heterozygosity at their respective informative loci (Table 2
and Figure 2
).
We investigated potential relationships between the presence of LOH at
10q23 and the intensity of PTEN immunostaining. Among the 44
informative tumors assessed for LOH and PTEN immunostaining, there
seemed to be an association between decreased or absent staining and
10q23 LOH. The frequency of tumors that showed LOH at 10q23 steadily
increased from tumors with ++ immunostaining (20% LOH), + staining
(52% LOH), to negative (-) staining (62% LOH)
(P = 0.0317 Wilcoxon rank-sum test) (Table 2)
.
However, it is also clear that one-third (15 of 44) of tumors did not
show structural deletion or mutation but showed diminished or no PTEN
expression.
P-Akt, p27, Cyclin D1 Immunohistochemistry in Ovarian Cancer
Of 49 ovarian cancer samples, 28 (57%) stained positive, graded
+, in comparison to the normal tissue (Figure 1E)
. Twenty-one of the 49
(43%) ovarian cancer specimens lost all P-Akt immunoreactivity and
showed negative immunostaining, graded - (Figure 1F)
. P-Akt
immunostaining showed a nuclear predominance, however, weak cytoplasmic
staining was also observed. Among these same 49 ovarian cancers, 4
(8%) showed negative immunostaining with p27 antibody, graded -
(Figure 3A)
. Twenty of the 49 (41%)
ovarian cancer specimens stained weakly, graded +, in comparison to the
normal tissue. Thus, 24 samples (49%) were classified as low
expressors. The remaining 25 (51%) tumors stained positively, graded
++ (Figure 3B)
and were classified as high expressors. p27
immunostaining was mostly nuclear; however, weak cytoplasmic staining
was also observed. Overexpression of cyclin D1 was detected in 35/49
(71%) of the ovarian cancers examined. The level of overexpression was
variable between the slides and was graded accordingly. Cyclin D1 was
overexpressed predominantly in the nuclei of tumor cells in 6 of 49
(12%) of the samples, and simultaneously in the nucleus and cytoplasm
of tumor cells in 10 of 49 (20%) (Figure 3C)
. Therefore, 16 of 49
(33%) of the tumors examined showed nuclear overexpression of cyclin
D1 (Table 3)
. In 19 of 49 (39%) of the
tumors examined, overexpression was detected only in the cytoplasm of
the tumor cells (Figure 3D)
. Therefore, 29 of 49 (59%) of the tumors
examined showed cytoplasmic overexpression of cyclin D1 (Table 3)
.
Representative examples of simultaneous nuclear and cytoplasmic, and
exclusively cytoplasmic staining are shown (Figure 3, C and D)
.
|
|
. The proportion of tumors that had
positive P-Akt immunostaining steadily increased from tumors with ++
PTEN expression (36%), to those with no PTEN expression (85%)
(P = 0.0083 Wilcoxon rank sum test; Figure 4
).
Of the 11 tumors that showed ++ PTEN expression, 7 (64%) of these
tumors stained p27 ++, 4 (36%) of these showed p27 + staining, and no
tumors showed negative p27 staining. (Figure 4)
. However, this
correlation between PTEN expression and p27 expression did not achieve
statistical significance (P > 0.05, chi-square
test; P = 0.4, Fisher two-tailed exact test). No
significant association was found between PTEN or P-Akt expression and
cyclin D1.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Among the samples that had intragenic PTEN mutations, only
two samples had both tumor and normal DNA available for investigation.
Interestingly, both these samples (NMS26 and BOS11) showed biallelic
structural inactivation by intragenic PTEN mutation plus
loss of the remaining wild-type allele. Although normal DNA was not
available, the exon 5 and exon 7 double-intragenic mutations in UK33
almost certainly represents mutations involving both alleles (Table 1)
.
Previous studies have shown that biallelic structural inactivation
occurs in ovarian cancer, either by somatic intragenic PTEN
mutation plus loss of the remaining wild-type allele, by double-somatic
intragenic PTEN mutations, or by homozygous deletion at
10q23.9,10
Our study and these data suggest that a subset
of ovarian cancer, albeit small, might be in the same category as
glioblastoma multiforme, a subset of endometrial cancers and cervical
cancers, where biallelic structural PTEN inactivation is
also an important mechanism.21,23,60,61
In case BOS11, mutation within intron 3, one base from the exon-intron boundary (IVS3–1G>T), has been shown to cause aberrant splicing and almost certainly also results in a transcript with decreased stability. This mutation was also identified in the corresponding germline DNA; thus, among 107 apparently sporadic ovarian cancer cases, we detected at least one (1%) carrying an occult germline PTEN mutation. It is, therefore, vital that this patient be followed closely for development of Cowden-related cancers, especially those of the breast, thyroid, and endometrium, because the presence of a germline PTEN mutation is a sensitive molecular diagnostic marker for Cowden syndrome.18,62 If our observations can be replicated, clinicians might wish to inquire about other features of Cowden syndrome and to take a good family history when faced with apparently sporadic ovarian cancer cases. Hereditary ovarian cancer occurs as a part of three clinically distinct syndromes, site-specific ovarian cancer and breast-ovarian cancer, both of which are because of germline mutations in BRCA1 and BRCA2 in >90% of such cases,63 and hereditary nonpolyposis colorectal cancer. It is possible that Cowden syndrome might be included as a minor differential diagnosis of genetic ovarian cancer cases.
Among the samples that were available for both PTEN
immunohistochemistry data and LOH analysis, we found that 77% (34 of
44) of these tumors had either partial or complete expressional loss of
PTEN at the protein level. Of these 34 samples, only 3% (1 of 34) had
structural biallelic inactivation, 53% (18 of 34) had structural
monoallelic deletion of PTEN, and the remaining 15 samples
(44%) had no evidence of PTEN structural anomalies and
thus, lost of PTEN expression might be because of epigenetic silencing
(Table 2)
. Especially worthy of note are the 13 tumors with no PTEN
expression (-): only 1 had structural biallelic inactivation by
intragenic PTEN mutation and loss of the remaining wild-type
allele; 7 had evidence of loss of one allele only and so, silencing of
the remaining wild-type allele might be by epigenetic means; and 5 had
complete silencing of both PTEN alleles via mechanisms
beyond structural alteration. Similarly, among the 21 tumors with weak
(+) PTEN immunostaining, 11 had monoallelic deletion that would account
for the decreased PTEN expression. However, the remaining 10 had
neither PTEN mutation nor deletion. These 10 tumors might
have monoallelic silencing via epigenetic mechanisms (Table 2
and
Figure 2
). In contrast, all tumors with strong PTEN expression had both
intact PTEN alleles. The two seemingly discordant samples
with strong expression but LOH only had LOH 5' of PTEN. Recent
accumulating knowledge has suggested that PTEN may be inactivated by
several mechanisms other than mutations and/or deletions in a
tissue-specific manner 24,29,50,54,56,57,64
. In general,
the data to date would suggest that more than one mechanism of PTEN
inactivation can occur in a given tumor type. It would seem, however,
that certain mechanisms of PTEN inactivation predominate in
certain tissues. For example, in primary cervical carcinomas and
glioblastoma multiforme, the predominant mechanism of PTEN silencing is
biallelic structural alteration (mutation; deletion).61
In
metastatic malignant melanoma, biallelic epigenetic mechanisms of PTEN
silencing predominate.57
In nonmedullary thyroid cancers
and endocrine pancreatic tumors, PTEN inactivation seems to be mediated
by differential subcellular compartmentalization.50,56
In
the case of primary epithelial ovarian cancers (this study), biallelic
structural alterations are rare whereas a mixed genetic/epigenetic and
biallelic epigenetic silencing are evident in the great majority of
tumors with decreased or absent PTEN expression. In contrast to thyroid
tumors and endocrine pancreatic tumors, there is no evidence of
subcellular compartmentalization as a mechanism of PTEN inactivation in
ovarian carcinomas. Nonetheless, the precise epigenetic or other
mechanisms, such as methylation or degradation, involved in modulating
PTEN expression are yet to be elucidated. Further, why
subcellular compartmentalization involving the nuclear compartment for
a molecule such as PTEN lacking a nuclear localization signal is
actively being investigated.32,50,56,65,66
It has been well established that PTEN signals down the PI3K-Akt pathway and that PTEN inversely correlates with P-Akt.25-27,29 Of the 49 tumors examined, 28 (57%) showed positive (+) P-Akt staining and 21 (43%) showed negative (-) staining, a general trend showing this inverse correlation. However, it should be noted that the inverse correlation is not observed among all tumors: six (12%) did not show this inverse correlation. These observations might suggest that in the subset of PTEN null-P-Akt null ovarian cancers, at least one other non-PI3K-Akt pathway downstream of PTEN is the major pathway involved. In contrast, in the subset of PTEN and P-Akt co-expressors, PTEN is likely not involved and P-Akt is anti-apoptotic via other upstream mechanisms. These in vivo observations in noncultured tumors corroborate our in vitro functional analyses that demonstrate that PTEN can induce cell-cycle arrest and apoptosis via PI3K-Akt-dependent and -independent pathways.67
Because of several reports of finding a relationship between PTEN and p27 and between PTEN and cyclin D1, albeit in vitro, we sought to examine differential expression levels among these proteins. Several reports have suggested that PTEN induced p27 expression downstream of Akt36-38 and that this was a mechanism of cell-cycle arrest mediated by PTEN. If this were true, we would have observed a direct correlation between PTEN expression and p27 expression. Although we saw an associative trend, this was not statistically significant nor was it convincing. Similarly, the literature has suggested that PTEN down-regulates cyclin D1.68 If this were true, then we should have observed a trend of increased cyclin D1 expression with decreased PTEN expression. In our series, cyclin D1 expression levels and subcellular location were random irrespective of PTEN or P-Akt status. Our observations might suggest that in vivo, at least in primary epithelial ovarian carcinomas, p27 and cyclin D1 are not prominent or direct targets of PTEN action.
In summary, despite the low frequency of intragenic somatic PTEN mutations in primary ovarian cancer, we have found a high frequency of decreased or absent PTEN protein expression, associated with increased P-Akt expression. Our data argues for the prominent role of PTEN inactivation in ovarian carcinomas via multiple mechanisms, ranging from biallelic genetic alteration to biallelic epigenetic silencing. Further, although the Akt pathway is involved in PTEN's role in a proportion of ovarian tumorigenesis, p27 and cyclin D1, previously described to be downstream of PTEN, do not seem to have this relationship with PTEN, at least not in epithelial ovarian carcinogenesis. Therefore, it would seem that not only are mechanisms of PTEN inactivation tissue-specific but downstream pathways of PTEN as well.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported in part by grants from the American Cancer Society, Atlanta, GA (RPG-98-211-01-CCE to C. E.); the United States Army Breast Cancer Research Program, Ft. Detrick, MD (DAMD17-98-1-8058 to C. E.); and the National Cancer Institute, Bethesda, MD (P30CA16058 to The Ohio State University Comprehensive Cancer Center).
Accepted for publication March 12, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G. W. Montgomery, D. R. Nyholt, Z. Z. Zhao, S. A. Treloar, J. N. Painter, S. A. Missmer, S. H. Kennedy, and K. T. Zondervan The search for genes contributing to endometriosis risk Hum. Reprod. Update, September 1, 2008; 14(5): 447 - 457. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A Dennis Targeting Akt in Cancer: Promise, Progress, and Potential Pitfalls Am. Assoc. Cancer Res. Educ. Book, April 12, 2008; 2008(1): 25 - 35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Yang, W. Kong, L. He, J.-J. Zhao, J. D. O'Donnell, J. Wang, R. M. Wenham, D. Coppola, P. A. Kruk, S. V. Nicosia, et al. MicroRNA Expression Profiling in Human Ovarian Cancer: miR-214 Induces Cell Survival and Cisplatin Resistance by Targeting PTEN Cancer Res., January 15, 2008; 68(2): 425 - 433. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Selvendiran, L. Tong, S. Vishwanath, A. Bratasz, N. J. Trigg, V. K. Kutala, K. Hideg, and P. Kuppusamy EF24 Induces G2/M Arrest and Apoptosis in Cisplatin-resistant Human Ovarian Cancer Cells by Increasing PTEN Expression J. Biol. Chem., September 28, 2007; 282(39): 28609 - 28618. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Treloar, Z. Z. Zhao, L. Le, K. T. Zondervan, N. G. Martin, S. Kennedy, D. R. Nyholt, and G. W. Montgomery Variants in EMX2 and PTEN do not contribute to risk of endometriosis Mol. Hum. Reprod., August 1, 2007; 13(8): 587 - 594. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. C.K. Leung and J.-H. Choi Endocrine signaling in ovarian surface epithelium and cancer Hum. Reprod. Update, March 1, 2007; 13(2): 143 - 162. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Stanford, J. W. Barrett, S. H. Nazarian, S. Werden, and G. McFadden Oncolytic Virotherapy Synergism with Signaling Inhibitors: Rapamycin Increases Myxoma Virus Tropism for Human Tumor Cells J. Virol., February 1, 2007; 81(3): 1251 - 1260. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Chaudhuri, S. Orsulic, and B. T. Ashok Antiproliferative activity of sulforaphane in Akt-overexpressing ovarian cancer cells Mol. Cancer Ther., January 1, 2007; 6(1): 334 - 345. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Hecht and G. L. Mutter Molecular and Pathologic Aspects of Endometrial Carcinogenesis J. Clin. Oncol., October 10, 2006; 24(29): 4783 - 4791. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Jiang, X. Chen, J. Shen, Y. Wei, T. Wu, Y. Yang, H. Wang, H. Zong, J. Yang, S. Zhang, et al. beta1,4-Galactosyltransferase V Functions as a Positive Growth Regulator in Glioma J. Biol. Chem., April 7, 2006; 281(14): 9482 - 9489. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
