Context.—Gastrointestinal stromal tumors (GISTs) are Kit/CD117-expressing mesenchymal neoplasms of uncertain malignant potential. The lack of a reliable method of prognostication hampers the selection of patients eligible for STI571 therapy. 10q22-q23 is a region involved in chromosomal losses found in a fraction of malignant primary and metastatic GISTs harboring PTEN (phosphatase and tensin homologue deleted on chromosome 10), a tumor suppressor gene often altered in human neoplasms.

Objective.—To investigate the role of PTEN in GISTs, an issue that to our knowledge has not been addressed previously.

Design.—PTEN status was determined in a series of 21 GISTs, with follow-up ranging between 6 and 198 months, using immunohistochemistry correlated with clinical data.

Results.—A greater than 25% fraction of cells with low or absent PTEN immunostaining was detected in 9 GISTs, including all those showing malignancy. By the log-rank test, a fraction of PTEN-deficient cells greater than 25% was associated with malignancy (P < .001). Percentage of cells underexpressing PTEN, size, cellularity, MIB-1 immunoreactivity, and coagulative necrosis proved to be associated with malignancy by Cox proportional hazards univariate analysis; low or absent expression of PTEN was the only factor selected by multivariate analysis (P = .03).

Conclusions.—PTEN downregulation is implied in GIST progression. The immunohistochemical assessment of PTEN status appears to be a promising method of GIST prognostication.

The relatively recent characterization of the KIT pathway and the exclusion of Kit protein (CD117)–immunonegative neoplasms from the gastrointestinal stromal tumor (GIST) category shed new light on the classification of this formerly confused group of neoplasms. While these events had an immediate therapeutic impact, with the use of the tyrosine kinase inhibitor STI571 (imatinib), so far the prognostic approach remains rather elusive. In a recent workshop organized by the National Institutes of Health to establish a consensus approach to diagnosis and morphologic prognostication of GISTs, a scheme based on mitotic rate and tumor size as features predictive of outcome was proposed, recognizing that none among the several additionally considered parameters offers consistent advantages. Moreover, because it is probably impossible at present to reliably separate benign from malignant cases, the resulting prediction was given only in terms of relative risk.1 The lack of a reliable method of prognostication hampers the selection of patients eligible for STI571 therapy, a critical step for avoiding waste of economic resources and possible escapes of responsiveness, particularly with the progress of adjuvant and neoadjuvant clinical trials.2–4 

Among the parameters investigated in GISTs, many genetic alterations have been found, with diverse relevance in terms of either diffusion or association with aggressiveness. Among these alterations, losses in chromosome 10q, often involving the whole long arm of that chromosome, have been found in a fraction of malignant primary and metastatic GISTs.5,6 10q22-q23 is the site of PTEN (phosphatase and tensin homologue deleted on chromosome 10),7 a tumor suppressor gene that produces a protein able to dephosphorylate both protein and lipid substrates. A pivotal role of PTEN is exerted through interference in the signal transduction pathways of phosphoinositide second messengers. In fact, by dephosphorylating phosphoinositol 3,4,5-triphosphate (PIP3), PTEN decreases the cellular membrane translocation of AKT, a serine-threonine kinase involved in proliferative metabolic and antiapoptotic pathways, thereby reducing its phosphorylation and resulting in increased apoptosis. Moreover, by upregulating p27, PTEN induces a downregulation of cyclin D1, leading to G1 arrest. PTEN may interfere with the mitogen-activated kinase pathway also. On top of that, the PTEN-induced decrease of tyrosine phosphorylation levels of focal adhesion kinase inhibits cell spreading, suggesting that PTEN may be important in the regulation of cellular interactions.7 

PTEN has been referred to as “the most highly mutated tumor suppressor gene in the post-p53 era”; its inactivation plays a role in several human neoplasms, including ovarian, endometrial, and prostate carcinoma, and glioblastoma.8 However, with regard to soft tissue sarcomas, the role played by PTEN seems rather secondary.9–11 To our knowledge, no data regarding the role of PTEN in GISTs have been published to date. Recently, a role of p16/INK4a (p16) derangement in the progression of these tumors has been shown.12,13 Dual inactivation of PTEN and p16 has been commonly observed in a broad spectrum of human cancer types.14 

We therefore investigated PTEN status in a series of GISTs with diverse levels of biological aggressiveness, an issue that to our knowledge has not been addressed previously. Our investigation focused on protein expression; in particular, a possible prognostic approach in GISTs using PTEN immunohistochemical analysis (an easily affordable technique) was tested. The results point to a role for PTEN underexpression in GIST malignant progression; moreover, PTEN immunohistochemical assessment emerges as a promising and easily affordable method of prognostication in GISTs.

Tumors

Case files for 21 GISTs (17 gastric, 3 small intestinal, and 1 large intestinal) from different patients (10 men, 11 women; age range at diagnosis, 51–87 years) were retrieved from the surgical pathology files of the Catholic University of Rome. Gastrointestinal stromal tumors were defined as CD117-immunoreactive spindle cell or epithelioid mesenchymal neoplasms of the gastrointestinal tract. Although most cases were diffusely immunoreactive for CD117, a few showed more limited immunoreactivity (see “Results”); we adopted a threshold at 10% of stained cells for considering a diagnosis of GIST, as reported previously.15 Our immunohistochemical approach to CD117 was conservative, in that we did not use antigen retrieval for this molecule. The routine hematoxylin-eosin–stained sections were reviewed for verifying the morphologic diagnosis; sections from the selected representative blocks were cut at a constant thickness of 5 μm for further investigations. Smooth muscle tumors (desmin-positive, CD117-negative), schwannian tumors (S100-positive, CD117-negative), and tumors with positive resection margins were excluded from the study. Morphologic features noted were tumor size; spindle and/or epithelioid cellular morphology; presence of coagulative necrosis; grade of cytologic pleomorphism; cellular density; and immunoreactivity for CD117, desmin, S100, CD34, and MIB-1. For calculation of cellular density, a Leitz Diaplan microscope (Leitz Messtechnik GmbH, Wetzlar, Germany) with a ×40 objective, which gives a field area of 0.312 mm2, was used. In tumors with heterogeneous cellularity, fields were selected to reflect the whole tumor composition. Necrotic, fibrous, and other regressive areas were excluded. Results were expressed as the number of cells per square millimeter. Follow-up data for 6 to 198 months (mean, 58.6 months) were available for all patients included in the study.

Immunohistochemical Analysis

Immunohistochemistry was performed using antibodies to CD117 (rabbit polyclonal), desmin, and MIB-1 (Dako, Glostrup, Denmark), CD34 and S100 (YLEM, Avezzano, Italy), and human PTEN (Santa Cruz Biotech, Santa Cruz, Calif). For desmin, S100, and MIB-1, antigen retrieval was accomplished by microwave irradiation at 750 W for 15 minutes in 10mM citrate buffer (pH 6). Specific preimmune sera or isotype-specific unrelated primary antibodies were used for the control stainings. Immunoreactivity for PTEN was exclusively nuclear. For a semiquantitative immunostaining evaluation, the slides were screened independently by 3 pathologists (R.R., L.M.L., and F.C.) who were unaware of the clinicopathologic data, counting at least 1000 neoplastic cells in the most representative areas. Nuclear staining was scored using 3 classes of intensity, relative to the immunoreactivity of vascular endothelium (which acted as internal positive control), as previously reported.16,17 Intensity scores were classified as intensely positive (staining intensity equal to that of the vascular endothelium), weakly positive (staining intensity diminished relative to the endothelium), and negative (Figure 1). The first category was considered normal, being detected in the nontumoral endothelial and inflammatory cells of the samples; the last 2 categories were considered to be the result of abnormally low (reduced or lacking, respectively) PTEN expression, and their cumulated percentage was adopted for scoring the samples. Differences between the extreme counts of the 3 pathologists never exceeded 6%. For the log-rank test, a threshold at 25% of cells with low or absent PTEN immunostaining was adopted for dividing the GISTs into high- and low-PTEN–expressing cases. Interobserver agreement was reached in the first analysis in 90% of cases; for the remaining cases, a consensus was reached by a joint review of the samples. For the continuous scores used in the Cox proportional hazards regression analysis, the global percentages from cumulated counts of the 3 pathologists were used. MIB-1 score was assessed in the areas with the highest nuclear-labeling density, counting at least 1000 cells.

Figure 1.

Degrees of nuclear immunoreactivity for PTEN in gastrointestinal stromal tumors (GISTs). In the context of GISTs heterogeneously expressing PTEN at the nuclear level, light spots evidence examples of intense (A, bottom cell in the left spot; B, right cell in the spot), weak (A, top cell in the left spot; B, left cell in the spot), and absent (C, spot) immunostaining with anti-PTEN antibodies; the spot with the asterisk in A shows a capillary lumen, filled with erythrocytes and lined by intensely immunoreactive endothelial cells (avidin-biotin complex immunoperoxidase, original magnification ×400).Figure 2. Immunoreactivity for PTEN in indolent and aggressive gastrointestinal stromal tumors (GISTs). Intense immunoreactivity for PTEN, detectable in the vast majority of nuclei in GISTs with unremarkable follow-up (A), was restricted to a fraction of cells (<75%) in GISTs with evidence of biological aggressiveness (B) (avidin-biotin complex immunoperoxidase, original magnification ×200)

Figure 1.

Degrees of nuclear immunoreactivity for PTEN in gastrointestinal stromal tumors (GISTs). In the context of GISTs heterogeneously expressing PTEN at the nuclear level, light spots evidence examples of intense (A, bottom cell in the left spot; B, right cell in the spot), weak (A, top cell in the left spot; B, left cell in the spot), and absent (C, spot) immunostaining with anti-PTEN antibodies; the spot with the asterisk in A shows a capillary lumen, filled with erythrocytes and lined by intensely immunoreactive endothelial cells (avidin-biotin complex immunoperoxidase, original magnification ×400).Figure 2. Immunoreactivity for PTEN in indolent and aggressive gastrointestinal stromal tumors (GISTs). Intense immunoreactivity for PTEN, detectable in the vast majority of nuclei in GISTs with unremarkable follow-up (A), was restricted to a fraction of cells (<75%) in GISTs with evidence of biological aggressiveness (B) (avidin-biotin complex immunoperoxidase, original magnification ×200)

Close modal

Statistical Analysis

The Statistica statistical software package (release 5.5, StatSoft, Tulsa, Okla) was used for all calculations. Malignant behavior criteria adopted for uncensored events in survival curves were local recurrence, invasiveness of adjacent organs, peritoneal dissemination, distant metastases, and disease-specific death. Low PTEN immunoreactivity (cutpoint at 25% of cells with low or absent PTEN immunoreactivity) was tested by calculating cumulative survival rates by the Kaplan-Meier method, followed by comparison with the log-rank test with the Yates correction for continuity. Using Cox proportional hazards regression analysis, age at diagnosis, size, MIB-1 reactivity, cellular density, and percentage of cells with low to absent immunoreactivity for PTEN were considered continuous variables; sex, site (stomach vs others), morphology (spindle cell vs epithelioid or mixed spindle cell and epithelioid), cytologic pleomorphism (low vs high grade), presence of coagulative necrosis, and CD34 positivity were categorical 2-step variables. For multivariate analysis, the prognostic parameters with P < .05 using Cox univariate analysis were selected. Any P < .05 was considered statistically significant.

Characterization of GISTs

Data concerning the 21 analyzed GISTs are summarized in Table 1. Tumors ranged from 1 to 23 cm (mean, 9.5 cm). Percentages of MIB-1–immunoreactive cells ranged from 0.1% to 29.5%. Sixteen tumors were spindle cell type, 3 were epithelioid, and 2 were mixed spindle-epithelioid. Distant metastases were present in 7 cases, infiltration of adjacent organs was present in 1 case, and local relapse was present in 1 case. Three patients died of disease. All the tumors were immunoreactive for CD117; in 17 cases, the staining was diffuse; in 4, it was limited to a fraction ranging between 12% and 30% of cells. Eighteen cases (86%) stained for CD34. Desmin and S100 staining was never detected.

Table 1.

Clinicopathologic Data and PTEN Expression in the Gastrointestinal Stromal Tumors*

Clinicopathologic Data and PTEN Expression in the Gastrointestinal Stromal Tumors*
Clinicopathologic Data and PTEN Expression in the Gastrointestinal Stromal Tumors*

PTEN Immunoreactivity

Percentages of tumor cells with low or absent PTEN nuclear immunoreactivity are summarized in Table 1. Cytoplasmic positivity was never detected. By adopting a threshold at 25% of cells with low or absent PTEN immunostaining, an abnormally low expression (ie, fraction of cells underexpressing PTEN >25%) of PTEN was detected in 9 samples (42.8% of total); these comprised 100% of the cases with evidence of malignancy (n = 8) and 7.7% of the cases with unremarkable follow-up (n = 1) (Table 1 and Figure 2).

Correlations

By the Kaplan-Meier method followed by comparison using the log-rank test, the presence of a fraction of tumor cells with low or absent immunoreactivity for PTEN greater than 25% was associated with malignancy (P < .001). By Cox proportional hazards univariate analysis, tumor size, cellular density, presence of coagulative necrosis, MIB-1 immunoreactivity, and percentage of cells with low or absent PTEN immunoreactivity were significantly associated with malignancy (Table 2). The only prognostic factor selected by multivariate analysis was percentage of cells with low or absent PTEN immunoreactivity, with a relative risk of 0.080 (95% confidence interval, 0.008–0.152; P = .03) (Table 2).

Table 2.

Statistical Analysis of Potential Prognostic Factors in Gastrointestinal Stromal Tumors*

Statistical Analysis of Potential Prognostic Factors in Gastrointestinal Stromal Tumors*
Statistical Analysis of Potential Prognostic Factors in Gastrointestinal Stromal Tumors*

We investigated the role of PTEN in GISTs, an issue that to our knowledge has not been addressed to date, in a selected series of 21 GISTs with adequate clinical follow-up. We used a 3-tiered PTEN immunohistochemical scoring method and tested its biological value against the follow-up data.

We found low or absent PTEN immunostaining exceeding 25% of cells in 9 of the 21 GISTs we investigated, all except 1 of which showed evidence of malignancy in follow-up. This low immunohistochemical reactivity for PTEN was associated with evidence of malignancy in the survival curve (P < .001). As far as we know, no data concerning the immunohistochemical evaluation of PTEN in soft tissue tumors have been published previously. We adopted a 3-tiered method, distinguishing among complete negativity, weak positivity, and intense positivity (Figure 1), using intensely stained endothelial cells as positive controls, as previously reported.16,17 The threshold adopted for considering a case as underexpressing PTEN (ie, >25% tumoral cells with low or absent PTEN immunostaining) seems a valid cutoff, since it produced data that were significantly associated with the biological behavior of the investigated tumors (Tables 1 and 2).

The common detection of PTEN immunoreactivity in the cytoplasm, together with the lack of a nuclear locator sequence in this molecule and the mainly extranuclear localization of the PIP3 pathway, which is considered the main target of PTEN,16,18 have led investigators to overlook data indicating nuclear localization of PTEN expression. Recent reports have called attention to nuclear PTEN immunoreactivity, found to flank its cytoplasmic counterpart, with various relevance in various tumors, including esophageal, endometrial, breast, colorectal, melanocytic, and endocrine pancreatic neoplasms. Many of these investigations noted a progressive loss of nuclear PTEN accompanying tumor progression.17,19–21 Studies detecting exclusively nuclear PTEN immunoreactivity, as was the case for the GISTs in this study, are rare; a possible reason for this discrepancy could be the complete lack of immunohistochemical investigations of PTEN in soft tissue tumors. Interestingly, exclusively nuclear PTEN immunoreactivity has been documented in normal endometrial stromal cells.22 In general terms, an inverse correlation between nuclear PTEN immunoreactivity and prognosis seems to exist17,19,21; with this perspective, it is noteworthy that negative nuclear PTEN immunostaining, unlike cytoplasmic PTEN expression, is an independent prognostic indicator for survival in esophageal squamous cell carcinoma.19 Thus, nuclear PTEN could be the ideal target of investigations dealing with the prognostic significance of this molecule.

With regard to the possible functions of nuclear PTEN, the presence, although at low levels, of phosphoinositol in the nuclear membrane together with phosphoinositol-3-kinase, led to hypothesizing a nuclear activation of proapoptotic pathways, analogous to that carried on by PTEN at the plasma membrane; moreover, since many transcription factors are regulated via phosphorylation, PTEN could directly dephosphorylate these factors at the nuclear level7; in both cases, a decrease of nuclear PTEN could facilitate cell proliferation. Accordingly, nuclear PTEN has been shown to reach maximal expression in G0-G1 phase, and to decrease in S phase in the MCF-7 breast cancer cell line.21 However, the observed biological effect of PTEN in GISTs seems to rely on additional factors also, since MIB-1 immunoreactivity, unlike PTEN staining, is not an independent predictive marker after multivariate analysis. Further studies are warranted for defining the function of PTEN in the nucleus; with this perspective, the apparent exclusively nuclear localization of PTEN in GISTs could render these tumors a useful model.

Our finding of reduced PTEN expression in aggressive GISTs is consistent with the recent demonstration of a similar behavior of p16,12,13 since dual inactivation of p16 and PTEN does occur commonly in human neoplasms.14 

Immunohistochemical underexpression of PTEN was the only prognostic factor selected by multivariate analysis (P = .03), whereas other parameters proposed for and/or commonly used in GIST prognostication, such as presence of coagulative necrosis, size, cellularity, and MIB-1 immunoreactivity, lost the significance evidenced by univariate analysis (Table 2). This result warrants further confirmatory studies in larger series aimed at investigating the possible use of PTEN immunohistochemical assessment as a prognostic test for GISTs; such testing would be affordable for every basic pathology laboratory. Our results showing an association between PTEN underexpression and clinical malignancy favor a late role for the inactivation of this molecule in GIST progression. A similar role of PTEN has emerged in various tumors, including melanoma, glioblastoma, and prostate carcinoma.8,17 

Most GISTs express constitutively activated Kit proteins bearing structural changes that allow receptor oligomerization and cross-phosphorylation in the absence of ligand binding; the activating mechanism in most GISTs is mutation of KIT gene itself, as demonstrated by studies in which systematic sequencing of the juxtamembrane coding region was coupled with evaluation of the entire KIT coding sequence in cases lacking juxtamembrane region mutation.23,24 Evidence of involvement of type D3 cyclin in the cascade triggered by KIT activation has been produced by demonstrating the upregulation of that type of cyclin by stem cell factor, the ligand of CD117,23 leading to cell cycle progression through G1/S transition in mouse spermatogonia.25 Transfection of PTEN into PTEN-deficient endometrial carcinoma cell lines has been shown to induce a specific reduction of cyclin D3 levels and an associated increase in the amount of the inhibitor p27KIP1 complexed with CDK2; enforced expression of cyclin D3 abrogated this PTEN-induced cell cycle arrest.26 A possible resulting scheme of GIST progression would thus be the following: a first mutation occurring in KIT locus would cause the onset of the tumor, resulting in a cyclin D3–driven mitogenic boost, whose effect is limited by PTEN; the inactivation of the latter could be the event that ultimately leaves the way open to the development of a malignant phenotype.

In conclusion, our data strongly support a role for PTEN downregulation in GIST progression. Furthermore, they warrant further studies aimed at a possible use of the immunohistochemical assessment of PTEN status as a method of GIST prognostication.

Supported in part by Fondi d'Ateneo, Università Cattolica, Rome, Italy.

Fletcher
,
C. D.
,
J. J.
Berman
, and
C.
Corless
.
et al
.
Diagnosis of gastrointestinal stromal tumors: a consensus approach.
Hum Pathol
2002
.
33
:
459
465
.
Berman
,
J.
and
T. J.
O'Leary
.
Gastrointestinal stromal tumor workshop.
Hum Pathol
2001
.
32
:
578
582
.
Dematteo
,
R. P.
,
M. C.
Heinrich
,
W. M.
El-Rifai
, and
G.
Demetri
.
Clinical management of gastrointestinal stromal tumors: before and after STI-571.
Hum Pathol
2002
.
33
:
466
477
.
Eisenberg
,
B. L.
and
M.
von Mehren
.
Pharmacotherapy of gastrointestinal stromal tumours.
Expert Opin Pharmacother
2003
.
4
:
869
874
.
El-Rifai
,
W.
,
M.
Sarlomo-Rikala
,
L. C.
Andersson
,
S.
Knuutila
, and
M.
Miettinen
.
DNA sequence copy number changes in gastrointestinal stromal tumors: tumor progression and prognostic significance.
Cancer Res
2000
.
60
:
3899
3903
.
Kim
,
N. G.
,
J. J.
Kim
, and
J. Y.
Ahn
.
et al
.
Putative chromosomal deletions on 9P, 9Q and 22Q occur preferentially in malignant gastrointestinal stromal tumors.
Int J Cancer
2000
.
85
:
633
638
.
Waite
,
K. A.
and
C.
Eng
.
Protean PTEN: form and function.
Am J Hum Genet
2002
.
70
:
829
844
.
Di Cristofano
,
A.
and
P. P.
Pandolfi
.
The multiple roles of PTEN in tumor suppression.
Cell
2000
.
100
:
387
390
.
Saito
,
T.
,
Y.
Oda
, and
K.
Kawaguchi
.
et al
.
PTEN/MMAC1 gene mutation is a rare event in soft tissue sarcomas without specific balanced translocations.
Int J Cancer
2003
.
104
:
175
178
.
Amant
,
F.
,
M.
de la Rey
, and
C. M.
Dorfling
.
et al
.
PTEN mutations in uterine sarcomas.
Gynecol Oncol
2002
.
85
:
165
169
.
Lin
,
C.
,
P. A.
Meitner
, and
R. M.
Terek
.
PTEN mutation is rare in chondrosarcoma.
Diagn Mol Pathol
2002
.
11
:
22
26
.
Schneider-Stock
,
R.
,
C.
Boltze
, and
J.
Lasota
.
et al
.
High prognostic value of p16INK4 alterations in gastrointestinal stromal tumors.
J Clin Oncol
2003
.
21
:
1688
1697
.
Ricci
,
R.
,
V.
Arena
, and
N.
Maggiano
.
et al
.
Role of p16INK4a in GISTs: abstracts of the 19th European Congress of Pathology, Ljubljana, Slovenia, September 6–11, 2003.
Virchows Arch
2003
.
443
:
403
.
You
,
M. J.
,
D. H.
Castrillon
, and
B. C.
Bastian
.
et al
.
Genetic analysis of Pten and Ink4a/Arf interactions in the suppression of tumorigenesis in mice.
Proc Natl Acad Sci U S A
2002
.
99
:
1455
1460
.
Miettinen
,
M.
,
L. H.
Sobin
, and
M.
Sarlomo-Rikala
.
et al
.
Immunohistochemical spectrum of GISTs at different sites and their differential diagnosis with a reference to CD 117 (KIT).
Mod Pathol
2000
.
13
:
1134
1142
.
Choe
,
G.
,
S.
Horvath
, and
T. F.
Cloughesy
.
et al
.
Analysis of the phosphatidylinositol 3′-kinase signaling pathway in glioblastoma patients in vivo.
Cancer Res
2003
.
63
:
2742
2746
.
Whiteman
,
D. C.
,
X. P.
Zhou
,
M. C.
Cummings
,
S.
Pavey
,
N. K.
Hayward
, and
C.
Eng
.
Nuclear PTEN expression and clinicopathologic features in a population-based series of primary cutaneous melanoma.
Int J Cancer
2002
.
99
:
63
67
.
Shi
,
W.
,
X.
Zhang
, and
M.
Pintilie
.
et al
.
Dysregulated PTEN-PKB and negative receptor status in human breast cancer.
Int J Cancer
2003
.
104
:
195
203
.
Tachibana
,
M.
,
M.
Shibakita
, and
S.
Ohno
.
et al
.
Expression and prognostic significance of PTEN product protein in patients with esophageal squamous cell carcinoma.
Cancer
2002
.
94
:
1955
1960
.
Martini
,
M.
,
M.
Ciccarone
, and
G.
Garganese
.
et al
.
Possible involvement of hMLH1, p16INK4a and PTEN in the malignant transformation of endometriosis.
Int J Cancer
2002
.
102
:
398
406
.
Ginn-Pease
,
M. E.
and
C.
Eng
.
Increased nuclear phosphatase and tensin homologue deleted on chromosome 10 is associated with G0-G1 in MCF-7 cells.
Cancer Res
2003
.
63
:
282
286
.
Mutter
,
G. L.
,
M. C.
Lin
,
J. T.
Fitzgerald
,
J. B.
Kum
, and
C.
Eng
.
Changes in endometrial PTEN expression throughout the human menstrual cycle.
J Clin Endocrinol Metab
2000
.
85
:
2334
2338
.
Heinrich
,
M. C.
,
B. P.
Rubin
,
B. J.
Longley
, and
J. A.
Fletcher
.
Biology and genetic aspects of gastrointestinal stromal tumors: KIT activation and cytogenetic alterations.
Hum Pathol
2002
.
33
:
484
495
.
Rubin
,
B. P.
,
S.
Singer
, and
C.
Tsao
.
et al
.
KIT activation is a ubiquitous feature of gastrointestinal stromal tumors.
Cancer Res
2001
.
61
:
8118
8121
.
Feng
,
L. X.
,
N.
Ravindranath
, and
M.
Dym
.
Stem cell factor/c-kit up-regulates cyclin D3 and promotes cell cycle progression via the phosphoinositide 3-kinase/p70 S6 kinase pathway in spermatogonia.
J Biol Chem
2000
.
275
:
25572
25576
.
Zhu
,
X.
,
C. H.
Kwon
,
P. W.
Schlosshauer
,
L. H.
Ellenson
, and
S. J.
Baker
.
PTEN induces G(1) cell cycle arrest and decreases cyclin D3 levels in endometrial carcinoma cells.
Cancer Res
2001
.
61
:
4569
4575
.

The authors have no relevant financial interest in the products or companies described in this article.

Author notes

Reprints: Riccardo Ricci, MD, PhD, Department of Pathology, Università Cattolica del Sacro Cuore, Largo Francesco Vito, 1, I-00168 Rome, Italy ([email protected])