| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Ovarian Cancer and Tumor Biology Programs, Department of Medical Oncology, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111 [C. D. D., I. H. R., L. V., R. B., T. C. H., A. K. G., X-X. X.]; Department of Pathology, Emory University School of Medicine, Atlanta, Georgia 30322 [C. C.]; and Third Department of Internal Medicine, University of Tokyo Hospital, Bunkyo-ku, Tokyo, Japan [T. Y., H. H.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The GATA transcription factors are conserved in insects and vertebrates from fly to humans and function in cell lineage specification during embryonic development and organ formation (10) . In mammals, there are six GATA family members: GATA-1, -2, and -3 are involved mainly in the development of the hematopoietic systems (11 , 12) ; GATA-4, -5, and -6 are expressed in a wide range of tissues and function in the formation of most, if not all, organs during embryonic development (10) . GATA-4 and GATA-6 are first expressed during the formation of extraembryonic endoderm differentiated from the pluripotent embryonic stem cells of the inner cell mass during early embryonic development (13, 14, 15) . In vitro analysis has shown GATA-4 and GATA-6 to be two of the most upstream factors during the primitive endoderm (an epithelial cell type) differentiation of pluripotent embryonic stem cells (16) . GATA-4 and GATA-6 are expressed in the heart (17) , liver (18) , lung (19) , gastric epithelium (20) , intestine and colon (21) , testis (22) , and ovary (23 , 24) and play critical roles in the development of these organs. GATA factors are not tissue specific but rather function in the specification and differentiation of cell lineages within an organ, such as the differentiation of an epithelial cell lineage from stromal cells. In a study of the differentiation of embryonic stem cells to endoderm cells, Dab2, GATA-4, and collagen IV were among 10 of the genes identified to be regulated by GATA-6 (25) . Dab2 is a candidate tumor suppressor of ovarian cancer expressed mainly in the surface epithelial cells, and its expression is often lost in ovarian tumors at an early stage of tumor development (7 , 26 , 27) . The loss of Dab2 closely correlates with the morphological transformation of the ovarian surface epithelial cells (8) and the disorganization of the primitive endoderm, the first polarized epithelial structure of the early embryos (9) .
In adult tissues, GATA factors likely function in maintaining the differentiated states of cells (10) . One possibility is that the loss of GATA factors or their cognate regulatory pathways leads to dedifferentiation of epithelial cells and contributes to tumorigenicity. Previously, the expression of GATA factors has been investigated in tumor cells. GATA-4 is expressed in sex cord-derived ovarian and gonad tumors (24) and gastric cancer cell lines (28) . It was found that GATA-4 and GATA-6 are reciprocally altered in adrenal tumors (29) .
In this report, we have investigated the expression of GATA-4 and GATA-6 in ovarian tumors and cell lines, and examined the correlation with dedifferentiation of the tumor cells using the expression of Dab2, collagen IV, and laminin as markers for epithelial differentiation. We observe the inappropriate expression and subcellular localization of the GATA transcription factors in ovarian tumors and cancer cells, and propose that alterations of GATA transcription factor expression and nucleocytoplasmic trafficking account for the dedifferentiation in ovarian tumorigenicity. Additionally in vitro suppression of GATA-6 led to cell dedifferentiation.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
HOSE and "Immortalized" Lines (HIO).
HOSE cells were derived from freshly dissected nontumor ovarian tissues obtained from women undergoing prophylactic oophorectomies. Briefly, the surface epithelium was scraped from the ovarian surface, and the cells were cultured in medium 199 and MCDB-105 (1:1) supplemented with 4% FBS and 0.2 units/ml of insulin (Novagen). These early passage HOSE cells were obtained in small quantities and were used mainly for immunofluorescence analysis, but were not sufficient in quantity for various biochemical characterization.
After culturing of the primary cultures for 1–2 months but before the cells entered replicative senescence, the HOSE cells were transfected with an SV40 large T antigen (SV40Tag) expression vector. These cells are referred to as HIO cells and can undergo an additional 20–30 population doublings before ceasing proliferation. The cells have been characterized and verified to be of epithelial cell type as described previously (5) . RNA used in the Northern blots was isolated from HIO cells at the following passages (p): HIO-103, p11; HIO-105, p10; HIO-107, p7; HIO-114, p11; and HIO-118, p34.
ROSE and Tumor (Nutu) Cell Lines.
ROSE cell lines were derived and characterized as described previously (30
, 31)
. Briefly, 20 ovaries were aseptically removed from 10 mature female Fisher rats (12–16 weeks of age) and were trypsinized to selectively release cells from the surface epithelial layer. Cell suspensions washed out from treated ovaries were pooled together, filtered with two sheets of cheesecloth, and transferred to tissue culture flasks. The cells were incubated overnight at 37°C, 5% CO2, in DMEM with 10% serum, and antibiotic supplement. The next day, the floating cells and debris were washed off with warm PBS, and the attached cells were further subcultured at 1- or 2-week intervals at a split ratio of 1:5 (for early passage, i.e., <10 subcultures). The cells were characterized and shown to be epithelial in more than 95% of the cell populations. The early passage of these cells is known as ROSE cells. ROSE 23 is an early culture of the cells and was tested to be nontumorigenic (30
, 31)
.
Continuous passaging and subculturing resulted in spontaneously transformed cell populations raised as foci in monolayer cultures. The cells from individual foci were collected, expanded, and examined for tumorigenicity by injecting these individual cell populations into female athymic nude mice. Late passages of ROSE cell subcultures (numbers 12, 14, 19, and 26), formed tumors by 3–6 weeks after inoculations. Histopathologically, tumors were as adenocarcinomas with different degrees of morphological dedifferentiation (30 , 31) . Tumors from cell populations 19 and 26 were poorly differentiated adenocarcinomas. Tumors from cell population 12 appeared to be well differentiated, and another tumor from cell population 14 was moderately differentiated. The cell lines derived from the tumor were named Nutu (number 12, 14, 19, and 26) lines (30 , 31) .
Immunohistochemistry.
The immunostaining of ovarian tissues and tumors was performed and analyzed as reported previously (5
, 7
, 8
, 32)
. Sections were first deparaffinized and rehydrated. Antigen retrieval was performed for 5 min at 120°C in citrate buffer (pH 6) using an electric steamer cooker for 5 min, followed by cooling for 10 min before immunostaining. All of the tissues were then exposed to 3% hydrogen peroxide for 5 min, primary antibodies for 25 min, biotinylated secondary linking antibodies for 20 min, streptavidin enzyme complex for 20 min, diaminobenzidine as chromogen for 5 min, and hematoxylin as counterstain for 1 min. These incubations were performed at room temperature; between incubations, sections were washed with Tris-Buffered Saline (TBS) buffer. An avidin-biotinylated enzyme complex kit (DAKO LSAB2) was used in combination with the automated DAKO AUTOSTAINER (DAKO Corp.). Coverslipping was performed using the Tissue-Tek SCA (Sakura Finetek USA, Inc., Torrance, CA) automatic coverslip. The sources of the primary antibodies were: monoclonal mouse anti-Dab2 IgG (Transduction Lab); anti-GATA-4 and anti-GATA-6 rabbit antiserum (Santa Cruz Biotechnology). The immunostaining of GATA-4 and GATA-6 was verified for specificity using blocking peptides on tissue sections containing GATA-4- and GATA-6-positive ovarian surface epithelia.
The slides were scored independently by three persons (C. C., X. X. X., I. H. R.) including a pathologist (C. C.). Staining in both cytoplasm and nuclear areas were scored separately. Positive scoring was given when the epithelial staining of collagen IV and laminin was higher than 10% on the slide. If tumor cytosolic Dab2 staining was positive in more than 10% of the cells, the tumor was scored as Dab2-positive. The results from three independent determinations were then compared, any differences in scoring results were discussed, and the slides were further examined to reach a common conclusion.
Cell Culture, Western and Northern Blot Analysis.
Ovarian epithelial and tumor cell lines were previously established (the OVCAR lines; Ref. 34
), or obtained from American Type Culture Collection (A2780, ES2, SKOV-3, and OV1016). The cells were cultured in DMEM with 10% FBS. Total cell lysate was used for Western blotting using antibodies against GATA-4, GATA-6, Dab2, and laminin.
Total RNA was isolated from 100-mm plates of 80% confluent monolayers using the guanidinium isothiocyanate/phenol/chloroform extraction procedure as described previously (5)
. Northern blot analysis was performed using 32P-labeled cDNA fragments with a random prime labeling kit, Prime-It II (Stratagene). cDNA for human GATA-6 was a generous gift from Dr. David Wilson (Washington University, St. Louis, MO) and from Dr. Kenneth Walsh (St. Elizabeth’s Medical Center, Boston, MA). Plasmids containing the partial cDNA of collagen IV
1 and -2, and laminin ß-1 were EST clones obtained from American Type Culture Collection. Human Dab2 cDNA was reported previously (35)
. The gel-purified cDNA fragments from restriction digestion were used as probes. All of the cDNA fragments were sequenced to verify their identity before use.
Cell Transfection.
Human GATA-4 (GenBank accession no. D78260; Ref. 36
) and GATA-6 cDNAs were cloned either in pcDNA3 vectors or in an expression vector (pMT-CB6+) under an inducible metallothionein promoter. Transfection of plasmid DNA were performed using Mirus Trans1T-LT1 reagent (Mirus Co., Madison, WI) according to the manufacturer’s protocols. GATA-4 was transfected into ES2 cells, which are positive for GATA-6 and Dab2 but negative for GATA-4 expression. GATA-4 was transfected in SKOV3 cells, which are negative for GATA-4 and Dab2, but are positive for GATA-6 protein expression. A plasmid-construct containing the GFP under the metallothionein promoter was used as a positive control for the cell transfection. After 72 h of transfection, the cells were cultured for 21 days in DMEM containing 10% FBS and 1 mg/ml G418. This selection medium was changed every 48 h. The stable clones selected were used to analyze GATA-4, GATA-6, and Dab2 expression before and after treatment with zinc sulfate (200 µM) for 2 days to induce gene expression.
Suppression of Gene Expression by siRNA Approach.
The expression of GATA-6 in cultured epithelial and tumor cells was reduced/suppressed using siRNA technique. Three 21-bp oligonucleotide sequences specific for human GATA-6 were tested in pSuppressorNeo vector (IMG:800; Imgenex, San Diego, CA) for suppression of GATA-6 expression. The most successful sequence found, hG6-110, is a 21-bp sequence 110 base 3' of ATG site-specific to human GATA-6 without significant similarity to other genes. A synthetic double-strand oligonucleotide was inserted (see below
) into pSuppressorNeo plasmid according to the manufacturer’s recommendation (Imgenex).
|
Immunofluorescence Microscopy.
Cells were plated on 22 x 40 mm cover slides in 6-well dishes and were fixed with 4% paraformaldehyde when they had reached 60% confluence. Cells were permeabilized with 0.5% Triton X-100 in PBS for 5 min, were washed with PBS, and were blocked with 3% BSA in PBS containing 0.1% Tween 20 (at room temperature for 30 min). GATA-4, GATA-6, Dab2, and laminin antibodies were used at 1:200 dilution in 1% FBS in PBS containing 0.1% Tween 20 and were incubated at 37°C for 2 h. The cellular localization of the antigens was revealed by fluorescein- or Texas Red-conjugated secondary antibodies (Jackson Immuno-Research Laboratories, West Grove, PA) at 1:200 dilution. The secondary antibodies were: donkey antimouse IgG conjugated with Texas Red and donkey antirabbit IgG conjugated with fluorescein. Rabbit anti-GATA-4 or anti-GATA-6 antibodies were used with mouse anti-Dab2 antibodies for double labeling. Goat anti-GATA-4 and rabbit anti-GATA-6 antibodies were used for detection of GATA-4 and GATA-6 on the same slides. Nuclei were marked by Dapi staining. The Nikon Eclipse E 800 epifluorescence microscope with x60 oil immersion objective linked to a Roper Quantix CCD (charged coupled device) camera were used for observation and image acquisition. A Nikon Eclipse E800 fluorescence microscope with x60 water immersion objective linked to a Bio-Rad Radiance 2000 LSCM (laser scanning confocal microscope) camera was also used to examine the slides. Images were merged by overlaying, using the Adobe-Photoshop program.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. All of the surface epithelial cells are intensely positive for both GATA-4 and GATA-6 staining in the nucleus. Some cells scattered in the ovarian stromas are also GATA-4 and/or GATA-6 positive.
|
. Dab2 staining exhibited a cytoplasmic speckled pattern, in agreement with its association with endocytic vesicles reported previously (33)
. Similarly, primary ROSE cells were found positive for GATA-4, GATA-6, and Dab2 expression by both Western blotting and immunofluorescence microscopy (not shown). Ovarian cancer cell lines of known GATA-4, GATA-6, and Dab2 expression status (detailed below), A2780, ES2, and SKOV-3, were used as positive and negative controls, respectively (detailed below).
Frequent Loss of GATA-6 in HOSE Cells and Tumors.
Previously, the loss of Dab2 was found to be an early event in ovarian tumorigenicity, correlating closely with dysplastic morphological transformation of ovarian surface epithelia (7
, 8) . Because Dab2 and collagen IV are regulated by GATA factors during the differentiation of embryonic stem cells to epithelial-like extraembryonic endoderm cells (25)
, we hypothesized that GATA factors might also function in the maintenance of ovarian surface epithelial differentiation by regulating expression of Dab2 and collagen IV. Thus, we investigated the expression of GATA-4 and GATA-6 in ovarian tumors to determine whether the loss of Dab2 might be caused by a dysfunction of these factors in ovarian tumor cells. Immunostaining of archived ovarian tumor tissues showed that GATA-6 is completely lost in cancer cells, in 15 (30%) of the 50 tumors analyzed (Fig. 1, C and D
; Tables 1
and 2
). In an example shown in Fig. 1C
, morphologically normal ovarian surface epithelial cells were positive for both GATA-4 and GATA-6. However, in malignant areas of the same tumor, cells were positive for GATA-4, but GATA-6 staining is absent (Fig. 1C)
. Additional examples are shown for five tumors that are negative for GATA-6 and positive for GATA-4 in the nucleus (Fig. 1D
, Tumors 1–5), two tumors that were negative for both GATA-4 and GATA-6 in the nucleus (Fig. 1D
, Tumors 6 and 7), and one tumor that was positive for both GATA-4 and GATA-6 (Fig. 1D
, Tumor 8). Unlike GATA-6, however, GATA-4 was present in most of the tumors, and only 7 (14%) of the 50 tumors analyzed were GATA-4 negative (Table 2)
.
|
|
. The expression of Dab2, collagen IV, and laminin, which are indicators of epithelial differentiation or dedifferentiation of the tumor cells, was also determined. Most of the tumor lines, including A2780, OVCAR-2, -3, -4, -8, and -10, and OV1016, have little or no expression of either GATA-4 or GATA-6, consistent with the observations of the loss of GATA-6 in tumor tissues. ES2, in which both GATA-6 and Dab2 are expressed strongly, is one of the few exceptions among tumor cell lines. Even in two of the five nontumorigenic HIO lines, GATA-6 is already absent, correlating with the loss of expression of Dab2, collagen IV, and laminin (Fig. 1E)
. In all of these tumor and HIO cell lines, with the exception of SKOV-3, the loss of GATA-6 expression correlates well with the loss of Dab2 expression. The expression of collagen IV, however, is present in OVCAR-2 and -10 despite the absence of GATA-6. Remarkably, GATA-4 expression is found to be absent in most of the tumor and nontumorigenic cell lines, in contrast to the observation from tumor tissues.
Loss of Nuclear Localization of GATA-6 in Ovarian Tumors.
In some of the ovarian tumors in which GATA-6 expression was classified as positive, the staining appears to be cytoplasmic rather than nuclear (Table 1)
. Examples of the cytoplasmic staining of GATA-6 are shown in four tumors (Fig. 2A)
. In these tumors, GATA-6 staining is absent in the nucleus of the tumor cells, which are counterstained blue by hematoxilyn (instead of brown or dark staining for the GATA-6 antigen). Brown staining, indicating the presence of GATA-6 protein, is visible around the nucleus, demonstrating the cytoplasmic localization of GATA-6 in these tumors. In contrast, GATA-4 staining is present in the nucleus of tumor cells from adjacent sections.
|
and confocal (Fig. 2C)
imaging. GATA-6 positive (primary HOSE cells, ES2, and SKOV-3) and GATA-6 negative (A2780) cell lines were analyzed (Fig. 2)
. In SKOV-3 cells, GATA-6 is not nuclear localized (Fig. 2, B and C)
, whereas it is nuclear in both HOSE (Fig. 2, B and C)
and ES2 (Fig. 2B)
cells. Thus, examples exist in cell lines to confirm the observation of GATA-6 nuclear exclusion in tumor tissues. Laminin staining was observed in HOSE and ES2 cells but not in A2780 and SKOV-3 cells (Fig. 2B)
, consistent with Northern blot results (Fig. 1E)
. Presumably, the expressed laminin protein detected by immunofluorescence signal is associated with the outer cell membranes.
Both GATA-4 and GATA-6 are transcription factors, and their function is believed to reside in the nucleus. Furthermore, GATA-4 and GATA-6 are known to work in collaboration in regulating myocardial gene expression (40)
and embryonic cell differentiation (16)
. Thus, by combining the number of tumors that have lost or have nuclear exclusion of one of the GATA factors, the majority (41 of 50, or 82%) of the tumors analyzed have lost the function of GATA factors (Table 2)
.
Heterogeneity in GATA-6 Expression among Tumor Cells.
In the tumors classified as GATA-6-positive, heterogeneity in the expression of GATA factors among cancer cells of the same tumors is a general feature, as shown by examples of GATA-4 and GATA-6 staining in three ovarian tumors (Fig. 3A)
. In tumor 1 (Fig. 3A)
, 
20% of the tumor cells are positive for GATA-6 staining, interspersed with the rest of GATA-6-negative tumor cells. All of the tumor cells appear to have GATA-4 staining, both nuclear and cytoplasmic. In this tumor, there is no detectable morphological difference between the GATA-6 positive and negative cells. In the second example (Fig. 3A
, tumor 2), the tumor cells are stained heterogeneously for GATA-4, about 30% positive and 70% negative. Uniquely, the GATA-6 staining in this tumor appears to locate in the nucleolus, whereas the nuclei are free of GATA-6 staining. Such a staining feature is deviant from the general observation for GATA-6 expression in ovarian surface epithelial cells. In the last example (Fig. 3A
, tumor 3), most tumor cells are positive for GATA-4 but negative for GATA-6. Interestingly, about 10% of the tumor cells that are GATA-6 positive in the nucleus are scattered among GATA-6-negative tumor cells. Thus, we observe variable GATA factor expression in morphologically indistinguishable tumors cells. This heterogeneity in GATA factor expression may contribute to the heterogeneity among cancer cells within a tumor mass (1
, 41)
.
|
. Examples are shown for the loss of GATA-6 but not of GATA-4 nuclear expression (Fig. 3B
, tumor 1), cytoplasmic instead of nuclear staining of GATA-6 in the epithelial cells of the morphologically nonneoplastic epithelium of another tumor (Fig. 3B
, tumor 2), and loss of both GATA-4 and GATA-6 expression in the nuclei of epithelial cells of ovarian surface epithelia (Fig. 3B
, tumor 3). Such results indicate that the loss of GATA-6 function, either by the absence of expression or by its nuclear exclusion, is an early event in ovarian epithelial transformation. This conclusion is consistent also with the loss of GATA-6 in some of the nontumorigenic HIO cells (Fig. 1E)
.
Frequent Loss of GATA Factors during Tumorigenicity of Rodent Ovarian Epithelial Cells.
We next determined whether the loss of nuclear GATA factors in ovarian tumors could be recapitulated in a rodent model for ovarian epithelial transformation. Previously, ROSE cells that were isolated and cultured in vitro were shown to transform to tumorigenic cells by repeated passage in culture (30
, 31)
. Thus, we also examined a panel of ovarian surface epithelial cells and derived tumorigenic cell lines (Nutu lines) for the expression of GATA transcription factors. Primary ovarian surface epithelial cells were able to grow continuously in passage as monolayer cultures for up to 2–3 months. Rose 23, a nontumorigenic line derived from expansion of the primary ROSE cells over an extensive period of culture (2 months), exhibits well-organized cell-cell adhesive morphology in a monolayer culture (Fig. 4A)
. The four tumorigenic lines, Nutu 12, -14, -19, and -26, appear to have lost the organized cell-cell contacts (Fig. 4A)
. ROSE 23, the nontumorigenic control line, exhibits low GATA-6 expression (Fig. 4C)
but has lost GATA-4 expression (Fig. 4B)
. The continuous culturing and passaging of the ROSE cells led to formation of foci on the monolayers, and eventually tumorigenic sublines were produced, as determined by the Nutu assay (30
, 31)
. The four Nutu lines, numbers 12, 14, 19, and 26, were derived independently from four ROSE cell preparations. Three of the Nutu lines, numbers 12, 14, and 19, have lost their GATA-4 expression (Fig. 4B)
but retain a low level of GATA-6 expression (Fig. 4C)
. In both ROSE 23 and Nutu lines, GATA-6 is not exclusively nuclear (Fig. 4, E and F)
. However, one of the lines (Nutu 26) expresses GATA-4, whereas GATA-6 is absent or greatly reduced (Fig. 4, B and C)
. All of the Nutu lines have lost the expression of Dab2 and laminin, and ROSE 23 expresses Dab2 and laminin only at low levels (Fig. 4, D–F)
. Thus, it seems that GATA-4 or GATA-6 is lost during the tumorigenic transformation of ROSE cells and that the GATA transcription factors are often lost early, before the tumorigenic phenotype is presented.
|
by Western blotting and in Fig. 5B
by immunofluorescence microscopy, the vector had no effect on either GATA-6 or Dab2 expression in ES2 cells; however, the hG6-110 construct suppressed both GATA-6 and Dab2 expression. Laminin expression was also down-regulated, but ß actin as control, determined by using the same blot, was not affected (Fig. 5A)
. The transfection efficiency was estimated to be >85%. In some cells in which GATA-6 is positive, presumably untransfected by the GATA-6 siRNA suppression construct, Dab2 is also positive (Fig. 5B
, arrow).
|
. The nuclear morphology of HIO-118 but not of ES2 cells was affected by the GATA-6 suppression construct, indicating that the loss of GATA-6 expression and subsequent change in gene expression drastically alters the biology of the cells. Perhaps, additional mutations in ES2 tumor cells allow better tolerance of the changes in gene expression profile caused by the abrupt loss of GATA-6 than do additional mutations in the nontumorigenic HIO-118 cells.
Restoration of GATA Factor Expression in Tumor Cells.
In tumor cell lines with varied expression status of GATA-4 and GATA-6, we examined the effect of restoration of GATA factor expression. We were able to establish several lines of inducible GATA-4-expressing clones from OVCAR cells, SKOV-3 (GATA-4-negative, GATA-6-positive but localized in cytoplasm; Fig. 6A
), and ES2 (GATA-4-negative, GATA-6 positive; Fig. 6C
) using an inducible expression vector (pMT-CB6+). The expression of GATA-4 is under the control of the metallothionein promoter that is inducible by ZnSO4. The induction of GATA-4 expression in SKOV-3 cells did not result in the induction of Dab2 (Fig. 6A)
, a transcriptional target of GATA-4 and GATA-6 (35)
. Induction of GATA-4 in all clones of SKOV-3 cells, however, dramatically altered cell morphology to small and spindle-like, greatly inhibited cell growth, and induced cell death by day 4 of zinc addition, as illustrated by representative images of the cells on dishes (Fig. 6B)
. All three clones of GATA-4-expressing cells, clone 4, -5, and -6, behaved similarly, and control cells of a GFP-inducible clone were not affected by the addition of zinc.
|
. In these GATA-4-transfected ES2 cells, GATA-4 was detected in the nucleus by immunofluorescence microscopy (Fig. 6D)
. In two clones (clones 4 and 5) tested, the induction of GATA-4 resulted in an increased expression of Dab2, with a smaller increase in the expression of the endogenous GATA-6 (Fig. 6E)
. Expression of GATA-4 also dramatically decreased the growth of ES2 cells, as determined by both MTT assay and cell counting (Fig. 6F)
. Thus, we conclude that reexpression of GATA-4 in either GATA-6 cytoplasmic-localized (SKOV-3) or nuclear-localized (ES2) cells greatly inhibits cell growth and/or survival under cell culture conditions. In three attempts to introduce GATA-6 expression into the ovarian tumor cell line A2780, which is negative for both GATA-4 and GATA-6, no GATA-6-expressing cell lines were obtained after the analysis of more than 50 clones of G418-selected cells in each transfection. The efficiency of transfection/expression of GATA-6 in these cells was monitored by immunofluorescence microscopy of the newly transfected cells. In all of these transfections, 3–10% of the cells in cultures were positive for GATA-6 4 days after transfection (not shown). However, after selection and expansion over a 4-week period, the G418-resistent cells were no longer positive for GATA-6 expression, as determined by either immunofluorescence microscopy or Western blotting. Thus the GATA-6-positive cells were presumably lost or the GATA-6 expression in the cells was suppressed during the selection and expansion of the transfected clones. We speculate that the reexpression of GATA-6 is incompatible for maintenance of transformed ovarian surface epithelial cells in culture. Further study of the effect of GATA-6 reexpression in tumor cells will be needed.
We also transfected A2780 cells with a mixture of both GATA-4 and GATA-6 expression constructs simultaneously (Fig. 7)
. Four days after transfection, cells expressing both GATA-4 and GATA-6 could be detected by immunofluorescence microscopy (Fig. 7A)
, and GATA-4 and GATA-6 were found to express in the same cells (Fig. 7B)
. Dab2 expression was not recovered after GATA-4 and GATA-6 expression, because no Dab2 staining (red) was detectable in cells expressing GATA-4 and/or GATA-6 (Fig. 7A)
. Thus, ectopic expression of GATA-4 and GATA-6 in transformed epithelial cells could not revert cells to the differentiated state based on the absence of Dab2 expression. Again, the clones of GATA-4 and GATA-6 positive cells could not be expanded and maintained in cultures.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
GATA Factors Act in Collaboration to Maintain Epithelial Differentiation.
The GATA factors are known to function in epithelial cell lineage determination during organ formation in embryonic development (17
, 18
, 20
, 21
, 42
, 43)
. In extraembryonic endoderm differentiation, both GATA-4 and GATA-6 are expressed and participate in gene induction (13
, 44
, 45)
, and ectopic expression of one induces the other in embryonic stem cells (16)
. Presumably, a proper makeup of GATA factors leads to differentiation toward a particular cell lineage. Both GATA-4 and GATA-6 are expressed in many epithelial cell types (10)
. However, in adult cells, the functions of these transcription factors have not been as extensively investigated. One can speculate that the expression of a proper ratio of GATA-4 and GATA-6 functions to maintain the differentiated states of cells in postdevelopment tissues.
Consistent with this notion, ovarian surface epithelial-derived tumor cells lose the expression of either GATA-4 (often in rat cells) or GATA-6 (often in human cells). The loss of one of the GATA factors often correlates with the loss of collagen IV expression, an indicator of epithelial differentiation. Thus, the loss of GATA factors and dedifferentiation would result in basement membrane-independence of epithelial tumor cells. In several tumor cell lines examined, either GATA-4 or GATA-6 is absent, although collagen IV and laminin are expressed. It is possible that tumor cells lose the expression of a GATA factor and collagen IV initially, and collagen IV expression is a gain-of-function alteration in later stages of tumor development. This would be consistent with the observation that ovarian cancer cells often lose extracellular collagen IV and laminin initially, and the restoration of collagen IV and laminin expression correlates with tumor cell spreading in later stages (5) .
The loss of Dab2 expression, however, closely correlates with the loss of GATA-6 function. Dab2 is another known transcriptional regulatory target of GATA factors (25)
. Dab2 is specifically expressed in ovarian surface epithelial cells (7)
, and its loss in tumor cells closely correlates with morphological transformation and disruption of the epithelia (8)
. It was thus suggested that Dab2 functions in the maintenance of epithelial organization (32)
. Dab2 deficiency in mice results in early embryonic lethality (9)
. The Dab2 (-/-) phenotype was characterized as disorganization of the visceral endoderm layer, an epithelial structure in early embryos (9)
. Thus, the phenotype of Dab2 deficiency in mouse embryos supports the idea that Dab2 functions in epithelial cell positioning organization and Dab2 acts in establishing epithelial polarity.5
Dab2 expression is lost in most of the ovarian tumor cell lines, correlating with the loss of GATA-6 (see Fig. 4
), and Dab2 expression is a more consistent marker of epithelial differentiation. Therefore, loss of Dab2 is a hallmark of epithelial dedifferentiation because of the deviant expression and function of GATA factors.
Mechanism of GATA Factors in Regulating Cell Lineage Specification and Differentiation.
Chromosome remodeling and chromatin structure are thought to be responsible for establishing and maintaining states of differential gene expression and thus cell functional differentiation during embryonic development (46)
. A role for GATA factors and other zinc finger transcription factors in interaction with chromatin and in gene regulation during cell lineage determination and development has been speculated (47)
. One intriguing idea developed by Zaret and colleagues [Bossard and Zaret (18)
, Zaret (47
, and Cirillo et al. (48)
in their investigation of hepatocyte cell fate determination from the precursor cells of the gut endoderm is the concept of the genetic potentiation by GATA-4. It was shown by in vivo footprinting experiments that in precursor cells of the gut endoderm, GATA-4 and HNF3 can alter chromatin conformation and occupy the binding sites in the regulated genes (such as albumin) without initiating their transcription (18
, 48) . The observation was interpreted as the binding of the GATA-4 transcription factors to chromatin as a gain of competence to differentiate while maintaining pluripotency. A secondary event leads to the activation of gene transcription and, thus, to the commitment of the cells to a GATA-4-positive differentiated cell fate, whereas the uncommitted cells can adopt an alternative cell fate.
In postdevelopment adult cells, GATA-4 and GATA-6 may be required to maintain chromatin structure of the differentiated cells (49) . In tumor cells, the expression of either GATA-4 or GATA-6 is often lost, or they are mislocated in cellular compartments and are unable to perform their nuclear function as transcription factors in maintaining chromatin conformation. The absence of GATA factors may allow the chromatin to drift from a "differentiated" conformation, and the tumor cells may thus lose their epithelial differentiation. The link between chromatin structure and cancer has been recognized (50) , and the functional losses of GATA factors may be the underlying mechanism for the abnormalities in chromatin structure associated with malignancy.
Another obvious question is how the expression of GATA factors are regulated during development and how their expression is lost in neoplastic transformation. The differentiation of embryonic stem cells to visceral endoderm cells in vitro has provided some insights, that GATA-4 and GATA-6 are upstream genes in the regulatory cascade (16) , and that environmental cues such as aggregation or morphogens such as retinoic acid (44) can induce the expression of GATA factors. However, the transcriptional regulatory mechanisms during embryonic development for GATA factors are complex and are yet to be explored (51) .
Implication of GATA Factors in the Mechanism of Tumorigenicity.
In normal tissues, differentiation of epithelial cells renders the growth and survival of the cells to regulation by tissue architecture organization and endocrine signaling. The proliferating tumor cells no longer obey the rules imposed on differentiated epithelial cells; these cells no longer depend on a basement membrane for growth and survival, and they often invade and colonize in adjacent (tumor spreading) or distal (metastasis) tissues (5
, 32)
. Such a lack of epithelial properties in morphology, behavior, and gene expression is known as dedifferentiation (1)
. Presumably a large number of epithelial-determining genes are deregulated in the tumor cells. The present thinking holds that only a few genetic mutations are required for the development of neoplasia (52)
, and, thus, epigenetic mechanisms may account for the vast deregulation of gene expression in cancers. We propose that the inappropriate expression of GATA factors in epithelial cells of postdevelopmental tissues occurs as an error in the execution of epigenetic program for the maintenance of differentiated cell lineages. The loss of GATA transcription factors may be a fundamental mechanism for the tumor cells to abandon the differentiated states and thus escape the regulation imposed on epithelial cells in normal tissues.
The loss of GATA factors is an early event in epithelial cell transformation because the changes have occurred in preneoplastic epithelium. Even in several lines of nontumorigenic, immortalized ovarian surface epithelial cells, either GATA-4 or GATA-6 is already lost. Therefore, the loss of GATA factors, and thus the dedifferentiation of epithelial cells, may be an earlier event in tumorigenicity that correlates with the morphological disruption of epithelial layer structure. Additional genetic and epigenetic changes may further drive the development of the epithelial cells into a carcinoma.
In most established epithelial and tumor cell lines, GATA-4 is absent, although it is present in most of the tumors. Additionally, in three of four rat tumor cell lines, GATA-4 rather than GATA-6 is lost. Even in the cultured nontumorigenic ROSE cells, ROSE 23, GATA-4 expression is absent. It is noted that, different from the development of human ovarian tumors, the rat ovarian epithelial cells first undergo tissue culture adaptation before tumorigenic transformation in vitro, in tissue culture condition. Thus, the correlation exists between the loss of GATA-4 and adaptation of the epithelial cells to in vitro tissue culture condition. Nevertheless, the loss of GATA-6 appears to be required for tumorigenic transformation in both cell lines and tumor tissues, whereas the loss of GATA-4 is often associated with the adaptation of both tumor and nontumorigenic ovarian surface epithelial cells to culture conditions (Fig. 8A)
.
|
. The present study of ovarian surface epithelial tumors presents a scheme that the loss of expression or function of GATA-6 and/or GATA-4 are general events associated with and account for dedifferentiation in the processes of adaptation of epithelial cells to cultures and/or neoplastic transformation. This study concludes that GATA-4 and GATA-6 determine epithelial lineage and reveals a possible mechanism underlying the phenomenon of epithelial dedifferentiation in carcinomas.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by NIH Grants R01 CA79716, R01 CA75389, and R01 CA095071 (to X-X. X.) from National Cancer Institute, by funds from Ovarian Cancer Research Foundation (OCRF, New York, NY; to X-X. X.), and by an appropriation from the Commonwealth of Pennsylvania.. T. C. H., A. K. G., and X-X. X. are also supported by funding from Ovarian Cancer SPORE P50 CA83638. 
2 Current address: Novartis Oncology/Pharmacology, Summit, NJ 07901.
3 To whom requests for reprints should be addressed, at Ovarian Cancer and Tumor Cell Biology Programs, Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia, PA 19111-2497; Phone: (215) 728-2188; Fax: (215) 728-2741; E-mail: X_XU{at}fccc.edu
4 The abbreviations used are: Dab2, disabled-2; FBS, fetal bovine serum; GFP, green fluorescence protein; HIO, immortalized HOSE (cell); HOSE, human ovarian surface epithelial (cell); Nutu, nude mice tumorigenic; OVCAR, ovarian carcinoma; ROSE, rat ovarian surface epithelial (cell); Dapi, 4',6-diamidino-2-phenylindole; siRNA, small interfering RNA; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide.
5 D-H. Yang and X-X. Xu, unpublished observations.
Received 12/30/02. Revised 5/29/03. Accepted 6/ 5/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. D. Capo-chichi, K. Q. Cai, J. R. Testa, A. K. Godwin, and X.-X. Xu Loss of GATA6 Leads to Nuclear Deformation and Aneuploidy in Ovarian Cancer Mol. Cell. Biol., September 1, 2009; 29(17): 4766 - 4777. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. A. Ramos-Vara and M. A. Miller Immunohistochemical Evaluation of GATA-4 in Canine Testicular Tumors Vet. Pathol., September 1, 2009; 46(5): 893 - 896. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M.E.I. Hellebrekers, M. H.F.M. Lentjes, S. M. van den Bosch, V. Melotte, K. A.D. Wouters, K. L.J. Daenen, K. M. Smits, Y. Akiyama, Y. Yuasa, S. Sanduleanu, et al. GATA4 and GATA5 are Potential Tumor Suppressors and Biomarkers in Colorectal Cancer Clin. Cancer Res., June 15, 2009; 15(12): 3990 - 3997. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Y. Hui and H. A. LaVoie GATA4 Reduction Enhances 3',5'-Cyclic Adenosine 5'-Monophosphate-Stimulated Steroidogenic Acute Regulatory Protein Messenger Ribonucleic Acid and Progesterone Production in Luteinized Porcine Granulosa Cells Endocrinology, November 1, 2008; 149(11): 5557 - 5567. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. S. Viger, S. M. Guittot, M. Anttonen, D. B. Wilson, and M. Heikinheimo Role of the GATA Family of Transcription Factors in Endocrine Development, Function, and Disease Mol. Endocrinol., April 1, 2008; 22(4): 781 - 798. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. V. Iorio, R. Visone, G. Di Leva, V. Donati, F. Petrocca, P. Casalini, C. Taccioli, S. Volinia, C.-G. Liu, H. Alder, et al. MicroRNA Signatures in Human Ovarian Cancer Cancer Res., September 15, 2007; 67(18): 8699 - 8707. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kanematsu, A. Ramachandran, and R. M. Adam GATA-6 mediates human bladder smooth muscle differentiation: involvement of a novel enhancer element in regulating {alpha}-smooth muscle actin gene expression Am J Physiol Cell Physiol, September 1, 2007; 293(3): C1093 - C1102. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kamnasaran, B. Qian, C. Hawkins, W. L. Stanford, and A. Guha From the Cover: GATA6 is an astrocytoma tumor suppressor gene identified by gene trapping of mouse glioma model PNAS, May 8, 2007; 104(19): 8053 - 8058. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A.-M. Bleau and E. C. Holland Trapping the mouse genome to hunt human alterations PNAS, May 8, 2007; 104(19): 7737 - 7738. [Full Text] [PDF]
|
|
|
|
 |
|
 J. K. Divine, L. J. Staloch, H. Haveri, C. W. Rowley, M. Heikinheimo, and T. C. Simon Cooperative interactions among intestinal GATA factors in activating the rat liver fatty acid binding protein gene Am J Physiol Gastrointest Liver Physiol, August 1, 2006; 291(2): G297 - G306. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Oskouian, J. Mendel, E. Shocron, M. A. Lee Jr., H. Fyrst, and J. D. Saba Regulation of Sphingosine-1-phosphate Lyase Gene Expression by Members of the GATA Family of Transcription Factors J. Biol. Chem., May 6, 2005; 280(18): 18403 - 18410. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. T. Kimchi, M. C. Posner, J. O. Park, T. E. Darga, M. Kocherginsky, T. Karrison, J. Hart, K. D. Smith, J. J. Mezhir, R. R. Weichselbaum, et al. Progression of Barrett's Metaplasia to Adenocarcinoma Is Associated with the Suppression of the Transcriptional Programs of Epidermal Differentiation Cancer Res., April 15, 2005; 65(8): 3146 - 3154. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Divine, L. J. Staloch, H. Haveri, C. M. Jacobsen, D. B. Wilson, M. Heikinheimo, and T. C. Simon GATA-4, GATA-5, and GATA-6 activate the rat liver fatty acid binding protein gene in concert with HNF-1{alpha} Am J Physiol Gastrointest Liver Physiol, November 1, 2004; 287(5): G1086 - G1099. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
