| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Department of Pathology,*
Kitasato University
School of Medicine, Sagamihara, Japan; the Department of
Pathology,
Cancer Institute, Tokyo,
Japan; the Department of Pathology,
Free
University, Amsterdam, the Netherlands; and the Department of
Urology,§
Yokohama City University School of
Medicine, Yokohama, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
mRNA transcription of CCNDs and protein synthesis is highest in middle-late G1 and lowest during S phase.2 Cyclin D1 (CCND1), D2 (CCND2), and D3 (CCND3) map to different chromosomes but have an identical cyclin box and PEST sequence.1,4 CCND1 has been proposed as a protooncogene because its derangement contributes to the uncontrolled cell growth characteristic of tumors.5 Overexpression and rearrangement of the CCND1 have been reported to be associated with tumor progression and/or poor prognosis in many different tumor types, such as the carcinomas of breast,6 esophagus,7 and pancreas,8 as well as mantle cell lymphomas.9 The commonest genetic abnormality affecting CCND1 is considered to be DNA amplification, resulting in increased transcription.1,2 There has been no report concerning any correlation between overexpression of CCND2, normally expressed in B lymphocytes, and tumor progression, although colorectal tumors may occasionally demonstrate amplification.10 It is possible that CCND2 resides in a chromosomal region that does not readily undergo amplification.11 However, it is considered to be a protooncogene because the immortalization of primary B lymphocytes by the Epstein-Barr virus is accompanied by transcriptional activation of CCND212.
Overexpression of CCNE has been reported to be a poor prognostic factor in cancers of various organs,13,14 and in this respect it is similar to CCND1. Over 10% of transgenic mice overexpressing human CCNE spontaneously developed mammary carcinoma.15 Akama et al stated that CCNE amplification and overexpression correlate with gastric cancer progression without any linkage to amplification of CCND1.16
CDK4, one of the catalytic partners for CCNDs, also shows DNA amplification, presumably leading to increased activity of CCND-dependent kinase complexes.17 It has been described that an association with CDK4 is required for CCND1 function in G1 progression.18 In addition, CDK2 is a catalytic partner of CCNE and is thought to be a poor prognostic factor in combination with CCNE in various organs.19,20
The p27 protein (p27), identified from the observation that transforming growth factor-ß and cell-cell contact prevent activation of the CCNE-CDK2 complex during the G1 phase,21 is a 27-kd heat-stable protein present in extracts from cells made quiescent by these signals. It has been implicated in the negative regulation of G1 progression in response to a number of antiproliferative signals by binding tightly to CCNE-CDK2 and CCND1/CCND2-CDK4 complexes and inhibiting their CDK activity in a stoichiometric manner.22 p27 activity peaks in G1, and it is thought that protein abundance rather than modification is the primary mode of its regulation. The amount of p27 varies, whereas the abundance of p27 mRNA remains unchanged. It is believed that the posttranslational alteration of p27 levels is achieved in part by a mechanism of translational control.3 In addition, a large amount of p27 is found in quiescent cells, suggesting that p27 maintains cells in G0.23 p27 levels decrease when cells reenter the cell cycle, mostly because of ubiquitin-proteasome-dependent degradation.24
Recent research using p27 knock-out mice indicated that p27 is a tumor suppressor gene, because -/- homozygotes often develop pituitary tumors spontaneously, a characteristic similar to that of mice with the Rb mutation.25,26 Moreover, it has been reported that lack or low expression of p27 is associated with aggressive progression and a poor prognosis in human breast,27 non-small lung,28 colorectum,29 stomach,30,31 and prostate cancers32,33 and lymphoma.34
In the present study a comparison of these cell-cycle-related factors was made, using a large series of gastric cancers to shed light on their importance for malignancy.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Two hundred and sixty cases of surgically resected gastric cancers were selected from the patients charts at Kitasato University Hospital from 1989 to 1991. All patients underwent total or subtotal gastrectomy, and in 256 cases regional lymph nodes were removed. A total of 253 cases could be followed up until November 1997. The patient series comprised 167 males and 93 females, with a mean age of 60.9 (from 28 to 87) years. The median follow-up time for the surviving patients was 65.3 months. Histopathological diagnoses were made, using the modified classification described elsewhere,35 of well (78 cases), moderately (42), and poorly (140) differentiated adenocarcinomas. Each of the three types were classified into four groups, in accordance with the tumor grading, as follows: limited to the mucosa (m.; 50), invasion within the submucosa (s.m.; 58), proper muscle (m.p.; 32), and beyond (s.; 120). They were subclassified as superficial (m.; 50) or truly invasive (s.m., m.p. and s.; 210).
Immunohistochemistry
Four-micron-thick sections of formalin-fixed and paraffin-embedded tissue were used for the study, with a combination of the standard labeled streptavidin-biotin-peroxidase (LSAB kit; Dako, Glostrup, Denmark) method and microwave oven heating (500 W, 5 minutes for trhee times in 10 mmol/L citric acid at pH 6.0). DCS-6 (x10; YLEM, Rome, Italy), cyclin E (x40; Novocastra Laboratories, Newcastle, UK), Cdk2 (x100; Transduction Laboratories, Lexington, MA), and Kip1 (x1000 dilution; Transduction Laboratories) mouse monoclonal antibodies for the CCND1, CCNE, CDK2, and p27 proteins and cyclin D2 (c-17) and cdk4 (c-22) rabbit polyclonal antibodies (x1000 dilutions; Santa Cruz Biotechnology, Santa Cruz, CA) for CCND2 and CDK4 proteins were applied. Counterstaining was achieved with 0.3% methyl green solution.
While cut-off values were basically determined on the basis of our own large scale of studies, those for CCND1 and CCNE positivity were defined as 5% of the cells, in accordance with an earlier report.36 Cases were defined as positive for CCND2 immunostaining when over 15% of the cells were stained in each section as described previously.37 The same criterion was used for CDK2 and CDK4. Gastric cancer cases positive for CCND1, CCND2, CCNE, CDK2, and CDK4 proved by Western blotting were included in each run as positive controls. Samples of gastric normal epithelium similarly established to be negative for all of these proteins were used as negative controls.
In areas of cancer tissue, percentages of p27-positive cells were calculated after counting randomly at least 1000 nuclei in the mucosal site in a high-power field (x400). The percentage was used as p27 labeling indices. In 197 cases, p27 indices were also calculated from examination of more than five normal crypts adjacent to cancerous lesions. We used a cut-off of 30% for plus or minus, in accordance with an earlier report.38 Lymphocytes in the same specimen were used as internal controls. Nonimmune rabbit serum was used for negative control sections.
Statistical analyses were made of data of three gastroenteropathologists concerned with inter- and intraobserver reproducibility. A close correlation was noted in the cases of CCND1, CCND2, CCNE, CDK2, CDK4, and p27 LI (P < 0.0001).
Western Blotting
Twenty cases of fresh gastric cancers, all invasive, and matched normal epithelium samples frozen at -80°C were examined.
Tissue samples were homogenized in 0.01 mol/L phosphate-buffered saline (PBS) solution and centrifuged (12,000 rpm, 30 minutes). The supernatants were mixed with 62.5 mmol/L Tris-HCl buffer (pH 6.8) containing 2% sodium dodecyl sulfate (SDS), 5% 2-mercaptoethanol, 7% glycerol, and 0.01% bromophenol blue and boiled for 10 minutes. Proteins (40 µg protein) were electrophoresed on 8% SDS-polyacrylamide gels at 30 mA for 3 hours and transferred onto 0.45-mm polyvinylidene fluoride membranes (Immobilon-P; Millipore, Bedford, MA), using a semidry system (Biocraft, Tokyo, Japan) at 200 mA for 30 minutes. Membranes were blocked with 15% skimmed milk in phosphate-buffered saline and then incubated with antibodies (double the dilutions used for immunohistochemistry) at 4°C overnight, followed by exposure to horseradish peroxidase-conjugated rabbit anti-mouse IgG and swine anti-rabbit IgG (Dako). Specific binding of antibody was determined using enhanced chemiluminescence of X-ray films (Fuji RX-U). All Western blots were repeated in triplicate for confirmation.
To evaluate overexpression on X-ray film, data for both cancer and matched normal tissues were read into a computer (Power Macintosh 7600/120) with a scanner (Epson GT-9500), and densities of bands were calculated with NIH Image software to obtain cancer-to-normal ratios. The cut-off ratio was 2.0 for a positive.
Reverse Transcription-Polymerase Chain Reaction Southern Blotting
mRNA was isolated by the phenol and guanidinium thiocyanate
method, and 1 µg of aliquots was dissolved in 20 µl of reaction
buffer containing reverse transcriptase (RAV-2; Takara, Ohtu, Japan),
random primers (Takara), and ribonuclease inhibitor (Takara). After
incubation at 37°C for 1 hour, c-DNAs were obtained and amplified in
a final volume of 100 µl of reaction mixture containing 2 U of
Taq polymerase (Takara), using 1 µmol/L primers. The
polymerase chain reaction (PCR) conditions were 94°C for 1 minute,
55°C for 1 minute, and 72°C for 2 minutes for 25 cycles. Primers,
oligonucleotide probes, hybrid temperatures, and lengths of PCR
products are listed in Table 1
. All
primers and oligonucleotide probes except the CCND1 primers
were originally designed by ourselves. The PCR products were
electrophoresed through 2% agarose gels containing ethidium
bromide and photographed under UV light. After gels were denatured,
neutralized, and Southern blot transferred onto nylon membranes (Hybon
N+; Amersham, Buckinghamshire, UK), hybridization was performed
using 10 pmol/ml of digoxygenin-labeled oligonucleotide probes against
the respective samples at 48°C to 64°C, adjusted for each probe
(Table 1)
, for 16 hours. After the membrane was washed with 2x SSC,
0.1% SDS and 0.1x SSC, 0.1% SDS at the same temperature as the
hybrid temperature, luminescence was detected with X-ray films (Fuji
RX-U), employing a Dig luminescent detection kit for nucleic acids
(Boehringer Mannheim).
|
Statistical analysis was performed using the Mann-Whitney
U test and the 
2
test for
demonstration of an interrelation among CCND1, CCND2, CCNE, CDK2, CDK4,
and p27 and other parameters. The survival period was defined as the
time from the operation to cancer death. Multivariate survival analysis
was performed with the Cox regression model to assess the additional
prognostic value of the different variables, using Statview J4.5 and
BMDP software.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Figure 1
shows the results of
Western blotting with CCND2 and CCNE for 20 cases of matched gastric
cancer and normal tissues. Single or doublet CCND2 bands
(phosphorylated and unphosphorylated) were found around 34 kd. Two
major bands of 49 and 43 kd, corresponding to the two alternatively
spliced forms, were obtained for CCNE. Figure 2
shows the results of reverse
transcriptase-polymerase chain reaction (RT-PCR) Southern blotting for
G1 cyclins, CDKs, and p27 in cancer
tissues, compared with ß-actin at the bottom. CCND1
protein overexpression was noted in 15% of cases and mRNA
overexpression in 35%. The respective figures were 30% and 50% for
CCND2, 50% and 55% for CCNE, 55% and 75% for
CDK2, and 35% and 45% for CDK4. There was a
clear tendency for mRNA positivity to be more frequent than protein
synthesis. There was no case demonstrating protein overexpression
without mRNA overexpression, except for that of CCND2. In
normal tissues, slight protein or mRNA expression was sometimes found,
but this was generally very weak. However, there were two discrepant
cases (2 and 15) between RNA and protein expression for
CCND2. Also, even in normal tissue, overexpression was
observed in cases 2, 7, and 11. There were no discrepancies between the
results of Western blotting and those of
immunohistochemistry.
|
|
shows Western blotting results
for p27 in fresh cancer (T) and matched normal tissues (N). Single or
doublet bands around 27 kd were obtained in all cases except for normal
tissue No6. T/N density ratios calculated with NIH Image software are
listed under the case numbers. There were nine cases with values under
1.0. The results of immunohistochemistry were not always the same as
with Western blotting. RT-PCR Southern blotting revealed a single band
of weak to strong density, proving intact production of p27 mRNA in all
cases. However, there was no significant linkage between protein and
mRNA production.
|
Immunohistochemistry of CCND1 and CCNE demonstrated 56 (21.5%)
and 79 (30.4%) of 260 gastric cancer cases to show nuclear staining
(Figure 4,a and c)
. CCND2 was positive in
nuclei only in 20 (7.8%), in cytoplasm only in 68 (26.2%), and both
in one (0.4%), making a total of 89 cases (34.2%). Characteristic
cytoplasmic staining at the perimembranes was often recognized (Figure 4b)
. CDK2 and CDK4 demonstrated nuclear or both nuclear and cytoplasmic
staining in 115 (44.2%) and 125 (48.0%) cases (Figure 4, d and e)
.
|
Table 2
summarizes data for
correlations among G1 cyclin and CDK positivity and clinicopathological
data. Although the male and female positive ratios did not differ for
either CCND2 or CCNE, the mean age of positive cases was significantly
higher in both cases (P = 0.0251 and
P = 0.0296). Statistically significant differences were
noted with CCND1 positive versus negative for less lymph
node metastasis (P = 0.0009), with CCND2 for
well differentiated (P < 0.0001) and vessel
invasion of cancer cells (P = 0.0008), with CCNE
for well differentiated (P = 0.0058) and less
depth of cancer invasion (P = 0.0256), with CDK2
for well differentiated (0.0234), and with CDK4 for vessel invasion
(P = 0.0053). With regard to CCND2, nuclear
stained lesions (21 cases) tended to be well differentiated
(P = 0.0006), with a lower depth of cancer
invasion (P = 0.0274), fewer lymph node
metastases (P = 0.002), and less vessel invasion
(P = 0.0659). In contrast, cytoplasmic staining
was associated with greater depth of invasion (P
= 0.0020), more lymph node metastasis (P =
0.0007), and vessel invasion (P < 0.0001),
although it was linked to good differentiation (Table 3)
. Thus CCND2 cytoplasmic staining
closely correlated with cancer progression.
|
|
p27 immunohistochemistry discovered three types of distribution of
p27-positive cells in normal gastric crypts of 197 cases: 1) positive
cells in the upper half and bottom of the crypt (89 cases; 66.4 LI); 2)
upper half positive but not bottom (57; 54.7); and 3) only upper
one-fourth positive (51; 27.0). Figure 5
demonstrates the close correlation of p27-positive cells between
cancers and normal crypts for all cases (correlation coefficient,
0.698; P < 0.0001).
|
Table 4
summarizes data for
correlations between p27 LIs and clinicopathological data.
Statistically significant differences were noted for histological
differentiation (P < 0.0001), depth of cancer
invasion (P < 0.0001), lymph node metastasis
(P = 0.0004), and vessel invasion
(P = 0.0013).
|
Correlations among G1 Cyclins, CDKs, and p27
Correlations were noted for CCND1 versus CCND2
(P = 0.0006), CCND1 versus CDK2
(0.0124), CCND1 versus CDK4 (P <
0.0001), CCND2 versus CCNE (P =
0.0237), CCNE versus CDK2 (P <
0.0001), CDK2 versus CDK4 (P =
0.0295), and CDK2 versus p27 (P =
0.0413) (Table 5)
. Pairs of CCND1 and
CDK4, and CCNE and CDK2, were strictly positively correlated. CCND1 and
CCND2, and CCND2 and CCNE, were inversely correlated.
|
Kaplan-Meier survival curves with comparisons of (+)
versus (-) for CCND2, CCNE, CDK4, and p27 are shown in
lymph node metastatic cases in Figure 6, A–D
. CCND2 (+) and CDK4 (+) were associated with a poor prognosis and
CCNE (+) and p27 (+) with a favorable prognosis. Table 6
summarizes values for
statistical significance. The CCND2 nuclear staining group showed
favorable prognosis without statistics, but the CCND2 cytoplasmic
staining group revealed an entirely poor prognosis. In Cox regression
analysis in the whole group of patients, depth of cancer invasion,
angioinvasion, lymph node status, age, sex, CCNE, p27, CDK4, CCND2, and
cytoplasmic CCND2 staining appeared to be independent prognosticators.
In the subgroups of invasive cancers, CCND2, CCND2 cytoplasmic
staining, p27, CDK4, and CCNE retained their additional prognostic
values, and in the subgroup of cases with lymph node metastases, this
was the case for CDK4, CCND2, CCND2 cytoplasmic staining, and p27.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Immunohistochemistry in the present study indicated that 1) overexpression of CCND2 and CDK4 is closely related to a poor prognosis of gastric cancers, 2) overexpression of CCNE correlates with favorable prognosis and not with tumor progression, 3) CCND1 and CDK2 are not linked with prognosis, and 4) lack of p27 is apparently associated with progression and a poor prognosis.
In an earlier large series of gastric cancers, CCND2 positivity was similarly correlated with cancer progression,37 this being expected to induce CCNE overexpres- sion, resulting in unrestricted cell proliferation.16 However, no such correlation was observed here, and, indeed, CCNE overexpression was rather linked to a favorable outcome. However, the difference in CCND2 localization might be important in this respect. It has been reported that CCND2 is resistant to extraction and is localized predominantly in nuclei in G1 but becomes more soluble and becomes distributed in both nuclei and cytoplasm from the G1/S transition onward.43 In addition, our Western blotting and immunohistochemistry data for fresh cases suggest that it is actually present in the cytoplasm. In breast cancers, one-third of cases show CCND2 immunopositivity in Western blots and only cytoplasmic staining in immunohistochemistry (our own unpublished data). CCND2 cytoplasmic localization found here might reflect an important physiological role in tumor progression and prognosis, although it is hard to explain the reason for the discrepancies between RNA and protein expression and overexpression observed in some normal gastric tissues.
CDK4 has been considered to play a role in tumorigenesis from the finding that erythroleukemia cells terminally differentiate when CDK4 is suppressed, whereas its overexpression causes uncontrolled cell growth and eventual malignant transformation.44 Recently it was suggested that gene amplification and overexpression of CDK4 are associated with high proliferative activity of breast cancers, with CDK4 amplification also appearing to be of importance in the pathogenesis of the subset of sporadic mammary tumors.45 Zang et al reported that CDK4 overexpression is associated with a significant increase in the proliferating cell number in colon adenomas on the basis of bromodeoxyuridine incorporation and immunohistochemistry for proliferating cell nuclear antigen.46 In the present study, although CDK4 overexpression did not relate to cancer progression except in terms of vessel invasion, it was associated with a poor prognosis, indicating some physiological role in the development of gastric cancer.
In vitro studies have revealed that CCNE contributes to the process of neuronal or osteoblastic cell line differentiation,47,48 and CCNE-CDK2 complexes accumulate in well-differentiated cells without catalyzing significant kinase activity.49 While CCNE overexpression may be related to progression and a poor prognosis in breast cancers50 and colorectal carcinomas,51 the present results suggest that gastric cancers are exceptional in this respect, but similar to lung cancers.52
Whereas CCND1 and CDK2 overexpression was not linked with progression or prognosis in the present study, this may not be the case for CCND1 in cancers of other organs,8,9,53 with some leading to the relation to abnormal cell proliferation, resulting in carcinogenesis.54 However, an association with limited progression and a better prognosis was found for bladder cancers55 and breast cancers intensively investigated for CCND1.6,56 It was concluded by van Diest et al that CCND1 overexpression correlates with tumor differentiation rather than progression.13
In our present study, a negative link between p27 overexpression and prognosis was noted, as in earlier reports.30,31 With regard to the relation of p27 expression to normal and cancer tissues, in prostatic cancers, p27 LI of normal epithelium was usually high, whereas that of cancer cells was extremely low.32,33 The significance of the correlation between values for normal and cancer tissues found here remains to be determined. It may be possible that p27 expression of cancer cells is regulated by homeostatic cell-cycle factors from normal epithelium. Further studies are needed to clarify this point. It has been demonstrated that p27 abundance is mostly regulated at a posttranscriptional level by ubiquitin-proteasome-mediated proteolysis.24 Esposito et al showed that p27 degradation was enhanced in lung non-small cell carcinomas with little or no p27.28 Loda et al described enhanced proteolytic activity specific for p27 in colorectal carcinomas, suggesting that low p27 levels can result from increased proteasome-mediated degradation rather than altered gene expression.29 With regard to the potential mechanism whereby p27 could exert an influence, the fact that loss of cell adhesion up-regulates p27 levels,22 because of an increase in protein stability,57 may be of interest. Cell adhesion plays an important role in regulative proliferation in the G1 phase, and its lack results in transformation.58 Correlations among G1 cyclins, CDKs, and p27 were found to be partly logical but partly not. Cell cycle control is a jigsaw puzzle, and because only some elements were examined here, a clear picture cannot be expected. More comprehensive studies will be needed in the future.
In the present study, down-regulation of p27 was proved to be a major prognostic factor and biomarker of gastric cancer, but without any apparent linkage to G1 cyclins or CDKs. Therefore the results suggested independent actions as oncoproteins.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by a grant from the Princess Takamatsu Cancer Research Fund.
Accepted for publication October 2, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  H. Kashima, T. Shiozawa, T. Miyamoto, A. Suzuki, J. Uchikawa, M. Kurai, and I. Konishi Autocrine stimulation of IGF1 in estrogen-induced growth of endometrial carcinoma cells: involvement of the mitogen-activated protein kinase pathway followed by up-regulation of cyclin D1 and cyclin E Endocr. Relat. Cancer, March 1, 2009; 16(1): 113 - 122. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A N A Milne, R Carvalho, M Jansen, E K Kranenbarg, C J H van de Velde, F M Morsink, A R Musler, M A J Weterman, and G J A Offerhaus Cyclin E low molecular weight isoforms occur commonly in early-onset gastric cancer and independently predict survival J. Clin. Pathol., March 1, 2008; 61(3): 311 - 316. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Mao, A. K. Stewart, R. Hurren, A. Datti, X. Zhu, Y. Zhu, C. Shi, K. Lee, R. Tiedemann, Y. Eberhard, et al. A chemical biology screen identifies glucocorticoids that regulate c-maf expression by increasing its proteasomal degradation through up-regulation of ubiquitin Blood, December 1, 2007; 110(12): 4047 - 4054. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Li, H. Wen, C. Brayton, P. Das, L. A. Smithson, A. Fauq, X. Fan, B. J. Crain, D. L. Price, T. E. Golde, et al. Epidermal Growth Factor Receptor and Notch Pathways Participate in the Tumor Suppressor Function of {gamma}-Secretase J. Biol. Chem., November 2, 2007; 282(44): 32264 - 32273. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Susaki, K. Nakayama, and K. I. Nakayama Cyclin D2 Translocates p27 out of the Nucleus and Promotes Its Degradation at the G0-G1 Transition Mol. Cell. Biol., July 1, 2007; 27(13): 4626 - 4640. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Gamboa-Dominguez, S. Seidl, E. Reyes-Gutierrez, C. Hermannstadter, L. Quintanilla-Martinez, R. Busch, H. Hofler, F. Fend, and B. Luber Prognostic significance of p21WAF1/CIP1, p27Kip1, p53 and E-cadherin expression in gastric cancer J. Clin. Pathol., July 1, 2007; 60(7): 756 - 761. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Weber, L. Shen, M. A. Aldred, C. D. Morrison, A. Frilling, M. Saji, F. Schuppert, C. E. Broelsch, M. D. Ringel, and C. Eng Genetic Classification of Benign and Malignant Thyroid Follicular Neoplasia Based on a Three-Gene Combination J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2512 - 2521. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Furuya, H. Kato, N. Nishimura, I. Ishiwata, H. Ikeda, R. Ito, T. Yoshiki, and H. Ishikura Down-regulation of CD9 in Human Ovarian Carcinoma Cell Might Contribute to Peritoneal Dissemination: Morphologic Alteration and Reduced Expression of {beta}1 Integrin Subsets Cancer Res., April 1, 2005; 65(7): 2617 - 2625. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. E. Bani-Hani, N. M. Almasri, Y. S. Khader, F. M. Sheyab, and H. N. Karam Combined Evaluation of Expressions of Cyclin E and p53 Proteins as Prognostic Factors for Patients with Gastric Cancer Clin. Cancer Res., February 15, 2005; 11(4): 1447 - 1453. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W K Leung, A H C Bai, V Y W Chan, J Yu, M W Y Chan, K-F To, J-R Wu, K-K Chan, Y-G Fu, F K L Chan, et al. Effect of peroxisome proliferator activated receptor {gamma} ligands on growth and gene expression profiles of gastric cancer cells Gut, March 1, 2004; 53(3): 331 - 338. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Padar, U. G. Sathyanarayana, M. Suzuki, R. Maruyama, J.-T. Hsieh, E. P. Frenkel, J. D. Minna, and A. F. Gazdar Inactivation of Cyclin D2 Gene in Prostate Cancers by Aberrant Promoter Methylation Clin. Cancer Res., October 15, 2003; 9(13): 4730 - 4734. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. GOUKASSIAN, A. DIEZ-JUAN, T. ASAHARA, P. SCHRATZBERGER, M. SILVER, T. MURAYAMA, J. M. ISNER, and V. ANDRES Overexpression of p27Kip1 by doxycycline-regulated adenoviral vectors inhibits endothelial cell proliferation and migration and impairs angiogenesis FASEB J, September 1, 2001; 15(11): 1877 - 1885. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Andres, J. Urena, E. Poch, D. Chen, and D. Goukassian Role of Sp1 in the Induction of p27 Gene Expression in Vascular Smooth Muscle Cells In Vitro and After Balloon Angioplasty Arterioscler Thromb Vasc Biol, March 1, 2001; 21(3): 342 - 347. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Evron, C. B. Umbricht, D. Korz, V. Raman, D. M. Loeb, B. Niranjan, L. Buluwela, S. A. Weitzman, J. Marks, and S. Sukumar Loss of Cyclin D2 Expression in the Majority of Breast Cancers Is Associated with Promoter Hypermethylation Cancer Res., March 1, 2001; 61(6): 2782 - 2787. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
