| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Pediatric Oncology [C. K., J. P., J. C., C. Y., L. H.] and Medicine Branches [P. N., J. T.], National Cancer Institute, and Cancer Genetics Branch, Human Genome Research Institute, NIH [J. K., P. M.], Bethesda, Maryland 20892
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In K7M2, spontaneous pulmonary metastases from an orthotopic primary tumor develop in >90% of mice compared with K12, where only 33% of mice develop metastases (P < 0.001).
|
To define genetic determinants of metastasis in OSA, we have used cDNA microarrays to compare gene expression between the clonally related high metastatic (K7M2) and low metastatic (K12) primary tumors. Array comparisons of K7M2 and K12 have defined 53 of 3166 unique printed cDNA probes (genes) that are differentially expressed. To focus our attention on a smaller group of potentially important genes, we have taken a functional approach to determine the significance and relevance of these differentially expressed genes vis à vis pulmonary metastasis. By review of the literature, we have assigned each differentially expressed gene to six nonmutually exclusive metastasis-associated categories including proliferation and apoptosis, motility and cytoskeleton, invasion, immune surveillance, adherence, and angiogenesis (7) . K7M2 and K12 were then compared in studies that independently examined each of these metastasis-associated processes. These studies (some presented in the initial description of the model; Ref. 2 ) demonstrated increased cellular motility and cytoskeletal changes suggestive of motile cells, earlier heterotypic adherence, and enhanced tumor angiogenesis in the more aggressive K7M2 compared with the less aggressive K12 model. On the basis of this functional and metastasis-related characterization, 10 genes that were assigned to the cell motility and cytoskeleton, heterotypic adherence, and angiogenesis categories were considered to be most likely to describe the aggressive behavior of K7M2 compared with the K12 cells. Ezrin, a member of the ERM gene family was 1 of these 10 genes. Similar to other members of its gene family, ezrin plays a role in linking the actin cytoskeleton to the cell membrane (8 , 9) . Ezrin has been associated with cell motility, invasion, and adherence but has not been described previously in OSA. Differential expression of ezrin between K7M2 and K12 was confirmed at both mRNA and protein levels. In situ differences in the distribution of ezrin protein suggested a role of ezrin in the motility of K7M2 cells. The relevance of ezrin in human OSA was supported by Northern analysis that demonstrated its expression in five of five human OSA cell lines.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
by two cycles of pulmonary metastasis implantation to the proximal tibia. The model conditions and characterization of the highly metastatic cell line, K7M2, and less metastatic cell line, K12, are summarized in Fig. 1
. K7M2 and K12 cells were maintained in vitro using complete culture media [DMEM; Celox Co., Hopkins, MN; 100 µg/ml penicillin-streptomycin; and 2 mM L-glutamine (Sigma Chemical Co., St. Louis, MO)] with 10% FCS (Sigma Chemical Co.) at 37°C at 5% CO2. All cell lines used for in vitro and in vivo studies were from the 3rd to 15th passages. For all in vitro assays, cells were harvested, using Trypsin/Versene, from near-confluent cultures. Cell viability was assessed using trypan blue, and experiments were not continued if cell viability was <90%.
RNA Extraction.
RNA was extracted from in vitro tumor cell lines using the Qiagen RNeasy Midi kit (Qiagen, Inc., Valencia, CA) according to the manufacturer’s specification. RNA was extracted from grossly dissected primary tumor and from pulmonary metastases using the Trizol (Life Technologies, Inc., Rockville, MD) reagent according to the manufacturer’s specification. Extracted RNA was quantitated by spectrophotometry and then examined by 1% agarose gel (Seakem GTG; FMC Bioproducts, Rockland, ME) electrophoresis. Intact RNA was then reextracted in 1.0 ml of Trizol for use in microarray experiments.
Microarrays, Probes, Hybridization, and Scanning.
The mouse array is composed of 3899 detector elements (probes). Of these, 315 are unclustered ESTs, 630 are clustered ESTs, and 3004 are clustered, named genes. There is significant redundancy in the named gene portion of the set, with 2221 unique clusters (probes) represented. All clones (including clustered and unclustered ESTs) were obtained from Research Genetics (Huntsville, AL); selection of probes was based only on availability. Clustering was undertaken by Research Genetics. PCR products from these clones were prepared and printed onto glass slides according to protocols described previously (11
, 12)
. Fluorescent cDNA targets (samples) were labeled with either Cy3 or Cy5 (Amersham Pharmacia Biotech, Piscataway, NJ) from 100 to 200 µg of total RNA by oligo dT-primed polymerization using SuperScript II reverse transcriptase (Life Technologies, Inc., Rockville, MD) as described previously (11
, 12)
. Imaging and image analysis was as described previously (11
, 12)
. Normalization for differential efficiencies of labeling and detection was performed using all spots (probes) that performed well as defined by criteria embedded in the DeArray software developed by Chen et al. (13)
. The normalization constant from these genes was then used to calculate the calibrated ratio for every cDNA probe within the image. Furthermore, the ratio variation of the these genes determined the confidence interval in which ratios are to be considered as no difference from 1.0. The 99% confidence interval was used throughout the experiments to test the significance of differentially expressed genes. Additional information on the image analysis as well as the raw data including the hybridization images for each sample is available on the Internet.3
Microarray outliers were defined as those genes that were significantly differentially expressed and that had a MaxMean signal intensity (defined as the maximum of the mean intensity of either the red or green channel for each gene) >2000 in two distinct array experiments (using samples from different mice bearing the K7M2 and K12 primary tumors). Differential expression of randomly selected genes was confirmed by Northern analysis. Northern blots consisted of 25 µg of total RNA from K7M2 and K12 cell lines, primary tumors, and pulmonary metastases. Plasmid probes, purified by MaxiPrep (Qiagen, Inc., Valencia, CA) containing sequence-confirmed cDNAs, were labeled by nick translation (Amersham, Life Science) with [
-32P]dCTP. Hybridization of Northern blots was carried out using Express Hyb (Clontech, Palo Alto, CA), according to the manufacturer’s recommendations. Equal loading of RNA was verified by reprobing all Northern blots with an [-32P]dCTP-labeled murine ß-actin plasmid probe. Northern analysis for ezrin in human OSA cell lines (HOS, MG63, U2, G292, and SaOS) was undertaken using a nick-translated plasmid provided generously by Dr. Richard Lamb (8)
.
Assignment of Microarray Outliers to Protein Function and Metastasis-related Gene Process Categories.
Differentially expressed genes, identified from cDNA microarray comparisons, were assigned to a modification of the NCBI Clusters of Orthologous Gene classification by searching the OMIM and PubMed databases by gene name.4
Genes were then assigned to six nonmutually exclusive metastasis-associated processes (proliferation/apoptosis, motility/cytoskeleton, invasion, adhesion, immune surveillance, and angiogenesis) using a PubMed database search of the gene name and each of the following terms: cancer, metastasis, proliferation, apoptosis, invasion, motility, immune surveillance, adhesion, and angiogenesis. Each gene was categorized with a single and mutually exclusive function, whereas gene assignment to metastasis-associated processes was not mutually exclusive.
Actin Cytoskeleton.
Cells were grown on sterile coverslips fixed with 3.7% formaldehyde in PBS for 10 min and then extracted with 0.2% Triton X-100 for 10 min at room temperature. After fixation and extraction, one unit of rhodamine phalloidin (Molecular Probes, Eugene, OR) was added to the coverslip prepared cells, as recommended by the manufacturer, and incubated at room temperature. After 1 h, the coverslip was washed two times for 2 min with PBS. Cells were stained with 0.4 µg/ml of 4',6-diamidino-2-phenylindole (Sigma Chemical Co.) for 10 min at room temperature. After incubation, the coverslip was washed with PBS, rinsed quickly with water, air-dried, and mounted onto slides using SlowFade (Molecular Probes). Cells were visualized on a Zeiss Axiovert microscope using a x63 objective, and images were captured with an Optronics CCD camera.
Motility and Invasion Assays.
Preincubation of 12-mm polycarbonate Transwell (Costar, Cambridge MA) plates for 4 h at 37°C was undertaken by adding complete culture medium to the upper chamber and complete culture medium with 10% FCS to the lower chamber before the addition of cells. Five thousand K7M2 or K12 cells were added in a volume of 200 µl of serum-free complete medium to the upper Transwell chamber. The Transwell upper chamber was placed into lower chambers that had been filled with 500 µl of complete medium with 10% FCS. For the Matrigel (Collaborative Biomedical Products, Bedford. MA) invasion assay, 100 µl of a 1:3 dilution of Matrigel in serum-free culture medium was added to the upper chamber of the Transwell and incubated for 2 h at 37°C (after the preincubation step). Transwell plates (motility assay) and Transwell plates coated with Matrigel (invasion assay) were incubated for 4, 12, 24, and 48 h at 37°C. At the completion of the incubation period, culture medium was suctioned from upper and lower chambers without disturbing cells. The total number of (motile or invasive) cells moving through the Transwell membrane was determined in replicates of six by wiping the apical surface of the Transwell membrane with a cotton swab and then staining the Transwell insert with DiffQuick (American Scientific Products, McGraw Park, IL). The Transwell membrane was then cut out using a No. 11 scalpel blade (Becton Dickinson Acute Care, Franklin Lakes, NJ), mounted, and coverslipped on a microscopic slide for cell enumeration. The total number of plated cells was evaluated at each time point (in replicates of six) by eliminating the "cotton swab step" and repeating the staining and membrane preparation procedure described above. The percentage of motility and percentage of invasion = (total number of cells on basilar surface of membrane)/(total number of cells on both apical and basilar surfaces of the membrane). Motility and invasion experiments were each repeated three times. Representative results from experimental conditions, repeated in replicates of six, are presented.
Gelatinase Zymography.
Gelatinase zymography was used to evaluate MMP2 and MMP9 activity in tissue culture supernatant and primary tumor as described previously (14)
. Serum-free medium was placed on confluent cultures of K7M2 and K12 for 24 h. Protein content in culture supernatants was quantified using the bicinchoninic acid assay (Pierce, Rockford, IL). Up to 25 µg of total protein were evaluated by gelatin electrophoresis (Novex Gel; Invitrogen, Carlsbad, CA). Standards for the pro and active forms of MMP2 and MMP9 (Calbiochem, La Jolla, CA) allowed identification of MMP forms.
Heterotypic Adherence Assay.
K7M2 and K12 cells were added to 96-well, flat-bottomed, substrate-coated plates (collagen type IV precoated plates and Matrigel precoated plates; Becton Dickinson, Bedford, MA) in quadruplicate. Cells were incubated at 37°C in 5% CO2 for 0.5, 1.5, 4, 16, 24, or 48 h. After incubation, medium was carefully suctioned out of each well. Using a multichannel pipette in a controlled manner, each well was washed three times with PBS (Biofluids, Inc., Rockville, MD). Between each wash, the plate was manually rocked back and forth three times. PBS was carefully suctioned out of each well. After the washes, the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay was used to determine the number of remaining cells (adherent cells; Ref. 15
). The percentage of adherence for each time point = (mean number of cells remaining in a well after washing)/(mean number of cells in wells that were not washed).
Immune Surveillance.
Female beige-scid mice (Charles River Laboratories, Wilmington, MA), 4–5 weeks of age, were housed under pathogen-free conditions with a 12-h light/12-h dark schedule and fed autoclaved standard chow and water ad libitum. In vitro passaged tumor cell lines (K7M2 and K12) were harvested and prepared for injection as described previously (16)
. Cells were brought to a final concentration of 1 x 107 cells/ml in phenol-free HBSS and kept at 4°C. Cells were enumerated, and viability was assessed using trypan blue (BioWhittaker, Walkersville, MD) staining. Experiments were continued if cell viability was >90%. A volume of 100 µl (1 x 106 cells) was injected into the lateral tail vein of beige-scid mice (10 mice/cell line). Mice were monitored at least three times weekly for evidence of morbidity associated with pulmonary metastases. Criteria for morbidity associated with metastases in mice included ill thrift, anorexia, dehydration, decreased activity and grooming behavior, and dyspnea. Sacrifice of mice with presumed pulmonary metastases was primarily based on the development of dyspnea. All mice that were sacrificed because of presumed pulmonary metastases had necropsy confirmation of diffuse metastases. Statistical comparisons were made using the Kruskal-Wallace nonparametric t test calculated with InStat for the MacIntosh. Statistical significance was defined as P < 0.05. Animal care and use were in accordance with guidelines of the NIH Animal Care and Use Committee (17)
.
Ezrin Immunocytostaining.
Coverslip preparations were fixed and extracted as described above. Nonspecific binding sites were blocked by incubating the cells with 1% BSA in PBS for 1 h at 4°C before processing for immunofluorescence labeling. Anti-ezrin antibody (Santa Cruz Biotechnology, Santa Cruz Biotechnology, CA) was added to the coverslip and incubated for 1 h at 4°C. After antibody incubation, the coverslip was washed two times for 2 min with PBS and incubated at 4°C with Cy3-conjugated rabbit antigoat immunoglobulin (Jackson Immunoresearch Laboratories, Inc). After 1 h, the coverslip was washed two times for 2 min with PBS. Cells were stained with 0.4 µg/ml of 4',6-diamidino-2-phenylindole (Sigma Chemical Co.) for 10 min at room temperature. The coverslip was then washed with PBS, rinsed quickly with water, air-dried, and mounted onto slides using SlowFade (Molecular Probes). Cells were visualized on a Zeiss Axiovert microscope using a x63 objective, and images were captured with an Optronics CCD camera.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
75% of outlier genes found to be common in replicate experiments. Thirty-one genes common to replicate experiments were overexpressed in K7M2 compared with K12 primary tumors, and 22 genes were overexpressed in K12 compared with K7M2 primary tumors (Tables 1
and 2
).
|
|
and 2
). The second classification of each gene was to six nonmutually exclusive metastasis-associated processes (proliferation/apoptosis, motility/cytoskeleton, invasion, adhesion, immune surveillance, and angiogenesis). The assignment of each gene into both classification systems is presented in Tables 1
and 2
. Twenty-nine of the 53 differentially expressed genes could be categorized into metastasis-related functions. Northern analysis confirmed the pattern of differential gene expression identified by cDNA microarray for selected genes (Fig. 2)
.
|
and 2
, we returned to the K7M2 and K12 models to uncover specific differences in their biologies for each of the six metastasis-associated processes (proliferation/apoptosis, motility/cytoskeleton, invasion, adhesion, immune surveillance, and angiogenesis). The six metastasis-associated processes examined were considered to be important steps in the cascade of events that follows a tumor cell from its primary site to a distant metastatic site. The metastasis-associated processes of proliferation/apoptosis and angiogenesis were examined in the original report of the K7M2/K12 model (2)
. These studies demonstrated no difference in apoptosis or proliferation between K7M2 and K12 but did demonstrate significant differences in tumor angiogenesis favoring K7M2.
Motility and Cytoskeleton.
To define differences in the cytoskeleton of the OSA cells, we examined their F-actin cytostructural architecture. Significant differences in the size and morphology of the K7M2 and K12 cells was demonstrated by F-actin staining (Fig. 3)
. The K7M2 cells were larger, had more numerous cellular extensions (filapodia and pseudopodia), greater substrate contact points, and enhanced spread phenotype than the K12 cells. Cellular extensions were less common and were significantly shorter in the K12 compared with the K7M2 cells. K12 cells contained organized actin stress filaments in the body of the cells, whereas in K7M2 actin stress filaments were concentrated within cellular extensions. In K7M2 cells, cytoplasmic actin was primarily punctate with very little in polymerized forms. Foot processes from cellular extensions in the K7M2 cells had very high levels of F-actin. Actin-rich fragments, not associated with the cells, were common in preparations of the K7M2 cell line and absent from the K12 cells. These cell-free fragments demonstrated similar patterns of actin expression as the K7M2 foot processes. F-actin staining seemed to suggest a greater potential for motility in the K7M2 compared with the K12 cells. To examine this potential difference in cellular motility, we performed in vitro motility experiments using Transwell tissue culture plates. In these assays, motility was defined as migration of cells from the apical surface of the Transwell membrane through 8-µm pores to the basal surface of the membrane. Significantly greater in vitro motility was demonstrated in the K7M2 cells compared with the K12 cells for all time points evaluated (4, 12, 24, 48, and 72 h of culture; P < 0.001; Fig. 4
). The number of cells leaving the basal surface of the Transwell membrane into the lower chamber was small (<2% of plated cells) and was similar for K7M2 and K12 cell lines (data not shown).
|
|
Heterotypic Adhesion.
Differences in the heterotypic adherence (cell-to-substrate adherence) of K7M2 and K12 cells were examined using type IV collagen and Matrigel as substrates. Adherence was assessed at early (0.5 and 1.5 h) and late time points (12, 24, and 48 h). The ability to initiate early adherence has been described as a distinct process from the maintenance of adhesion by cancer cells. Heterotypic adherence to both type IV collagen and Matrigel was significantly greater in K7M2 cells compared with K12 cells after 0.5 and 1.5 h of culture (early adherence; P < 0.01; Fig. 5
). However, no significant difference in heterotypic adhesion was observed between K7M2 and K12 cells for the late adherence (maintenance of adherence) time points for either type IV collagen or Matrigel. Heterotypic adhesion for both K7M2 and K12 increased significantly from 12 to 24 to 48 h (P < 0.01). Microscopic evaluation of 96-well plates suggested that homotypic adhesion (cell-to-cell adhesion) may have contributed to the total adhesion measured at time points after 48 h.
|
On the basis of the studies comparing the K7M2 and K12 model for each of six metastasis-associated processes, we concluded that differences in motility/cytoskeleton, heterotypic adherence, and angiogenesis (tumor vascularity, as assessed in the original report of the model; Ref. 2
) were most likely to explain the more aggressive behavior of the K7M2 compared with the K12 models. For this reason, special attention was directed to the microarray outliers that had been assigned to the motility/cytoskeleton, heterotypic adherence, and angiogenesis categories (Tables 1
and 2
). The 10 microarray outliers (taken from Tables 1
and 2
) assigned to these three metastasis-associated categories are presented in Table 3
. Ezrin, a gene not described previously in mesenchymal tumors, was examined further based on its described functions in several metastasis-associated processes and its novelty to OSA.
|
. The strongest ezrin staining signal was seen at the cytoplasmic boundaries, cellular extensions, and foot processes of the K7M2 cells. The absence of ezrin staining in the K12 cellular extensions was notable. This pattern of ezrin staining was similar to the distribution of F-actin seen in K7M2 cells (Fig. 3)
. The potential relevance of ezrin in human OSA was assessed by Northern analysis of human OSA cell lines. High expression of ezrin was found in four of five human OSA cell lines (Fig. 7)
. Ezrin expression in the fifth cell line, G292, was low but specific for ezrin in repeat Northern hybridizations.
|
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
is the result of microarray analysis, assignment of genes to metastasis processes, and detailed functional studies that compared the K7M2 and K12 models in each of these six metastasis-associated processes. The outlier genes identified in the comparison of K7M2 and K12 tumors included several that have not been described previously in OSA (e.g., galectin-3, ezrin, and clusterin). This is to be expected and is an advantage of a cDNA microarray that included a wide variety of genes. Microarrays that only include genes from a specific field (i.e., OSA or cancer) would not share this advantage. It is not surprising that >50% of known genes (excluding ESTs with no known functions) identified in this analysis were placed into at least one metastasis-related category. This high rate of assignment of genes is related to the design of the experiments (use of clonally related models differing in spontaneous pulmonary metastatic potential) and the bias of the method (PubMed literature search) by which genes were assigned to categories. Another important feature of the arrays used herein is the inclusion of ESTs. Their inclusion has allowed three ESTs, not currently in a UniGene cluster, to be associated with the metastatic behavior observed in the model system. A common concern with data generated by cDNA microarray analysis is the validation of the gene outlier list. We have reported previously on the high concordance of gene expression defined by our cDNA microarray and Northern analysis (11) . This concordance was confirmed by Northern for a selected number of genes defined in our study and was further validated at the protein level, using immunostaining, for ezrin.
The importance of motility during metastasis has been reviewed recently (21, 22, 23, 24) . In vitro motility assays and differences in actin cytoskeleton staining strongly supported the enhanced motility phenotype of K7M2 compared with K12. The presence of cellular extensions, including filapodia and lamellipodia, have been associated previously with tumor motility, invasiveness, and metastasis in several tumor models and supports the more aggressive (motility) phenotype of the K7M2 cells. K7M2 cells appear to be "primed" for motility, with little polymerized cytoplasmic F-actin, awaiting environmental cues with stress filaments organized at the periphery of the cell. Support for this hypothesis comes from studies of motile renal epithelial cells that demonstrate increased motility and decreased F-actin stress cable formation after stimulation of the hepatocyte-growth factor pathway (25) . The presence of actin-rich fragments, found free from the K7M2 cells, may also support enhanced motility. These fragments may represent a passive fracture of cytoplasmic foot processes seen in highly motile cells. It is also possible that they represent an active process of cytoplasmic shedding associated with tumor stroma formation.
Differences in the motility of K7M2 and K12 cells focused our attention on the cDNA microarray outliers (genes) associated with motility. Interesting members of the motility category include ezrin and galectin-3. Neither of these genes have been reported previously in OSA. Ezrin is a member of the ERM and merlin family of genes (26) . The protein products of the ERM genes are part of the band 4.1 protein superfamily (27) . Ezrin is a downstream effector of the Rho kinase signaling pathway, acting as a cytoplasmic linker of F-actin with the cell membrane (28) . The importance of the RhoC and the Rho/Rho kinase signaling pathway in cell motility, actin cytoskeleton, and metastasis has been demonstrated recently (20) . The parallel distribution of F-actin and ezrin protein at cytoplasmic boundaries, within cellular extensions, and at foot processes suggests the role of ezrin in the motility and actin cytostructure of K7M2 cells. The immunocytostaining and distribution of ezrin in the K7M2 cells lend support for the biological role of ezrin in the more aggressive K7M2 cells. The potential importance of ezrin in human OSA was supported by high expression in four of five and detectable levels in five of five human OSA cell lines. This is the first report of ezrin expression in OSA, a gene more commonly associated with epithelial tissues. Further work should examine the role of ezrin in other mesenchymal tissues and malignancies.
The ability of a cancer cells to adhere to substrates at distant sites (i.e., the pulmonary arterioles and venules) is essential for successful metastasis. The importance and timing of heterotypic adhesion in this process are currently under debate (29, 30, 31) . Using a simple in vitro assay of heterotypic adherence, the increased ability of the K7M2 cells to adhere to type IV collagen and Matrigel at 0.5 and 1.5 h of culture (suggestive of enhanced early heterotypic adhesion) was demonstrated. No differences in heterotypic adhesion were seen after culture periods of 12 h, suggesting equivalent abilities of the K7M2 and K12 cells to maintain adherence. After longer culture times, it is possible that homotypic (cell-to-cell) adhesion contributed to the total adhesion measured in the in vitro assay. Interestingly, the ERM family of genes (including ezrin) have been associated with early heterotypic adherence and not necessarily associated with the ability to maintain late adherence (30) . Another gene with adherence-associated functions is galectin-3. Galectin-3 is a lectin binding protein that has been associated previously with a malignant phenotype in several epithelial cancers including prostate and colon carcinomas (32 , 33) . Galectin-3 may be also associated with metastasis through its demonstrated roles in tumor motility and invasion (34) . The diverse metastasis-associated functions of galectin-3, including heterotypic adherence, make it an interesting candidate for evaluation in sarcomas including OSA.
In our previous characterization of the K7M2 and K12 models, enhanced angiogenesis of the K7M2 model was documented by CD31 and factor VIII staining of the primary tumor and the pulmonary metastases (2)
. The presence of ill-formed vascular structures, which interacted with nearly all tumor cells, was evident in sections of the K7M2 primary tumor and pulmonary metastasis. Conversely, the K12 tumor cells were densely packed with very few vascular structures present. In this earlier work, expression patterns for several angiogenesis-associated genes, including flt1, flt4, TIE1, TIE2, CD31, and VEG-F, were similar in cell lines, primary tumors, and metastases from both K7M2 and K12 models. The strong difference in the angiogenic phenotype of the models and the lack of differences in gene expression for the "classical" angiogenesis-associated genes suggest the possibility that a less well-recognized angiogenesis-associated gene, potentially defined by our microarray analysis [connective tissue growth factor, integrin V (CD51), or galectin-3], may be important in defining the angiogenic phenotype of the more aggressive K7M2 model.
On the basis of our functional characterization of the K7M2 and K12 cell lines and tissues, genes associated with proliferation and apoptosis, tumor invasion, and immune surveillance were considered to be less important determinants of metastasis in this model. This is not to say that these processes are not important for metastasis; rather, that these processes are less likely to characterize the differences in metastases observed between K7M2 and K12. Defining the importance of genes using this functional approach depends on valid and relevant assays for each metastasis-related process. It is also dependent on the correct assignment of a gene within a functional group (using PubMed database searches). The proliferation and apoptotic rate of the tumor models was assessed in vitro (doubling time) and in vivo (immunohistochemical staining with Ki67 and terminal deoxynucleotidyltransferase-mediated nick end labeling) in our initial characterization of the model (2) . The concordance of findings in both in vitro and in vivo assays suggested that differences in resting proliferation and apoptosis rates do not provide a metastatic advantage for K7M2. Tumor invasion was assessed using a simple Matrigel invasion assay and by measuring MMP activities in K7M2 and K12 cells. Similar assays have been used to assess the invasive potential of tumor cells (35) . Both assays failed to demonstrate significant differences in the invasive phenotype of K7M2 and K12. The sensitivity of the Matrigel invasion assay may be low and may not have detected subtle differences in the invasive phenotype of the K7M2 and K12 cells. An in vivo system was used to examine whether differences in immune surveillance could account for differences in the biology of K7M2 and K12. It is possible that minor histocompatibility differences could emerge in two clonally related cell lines and that these differences could result in the more effective immune recognition and destruction of one tumor cell line compared with another. Our results demonstrate that the relative metastatic phenotype of the K7M2 and K12 cell lines was maintained in significantly immunocompromised animals (beige-scid mice). Therefore, genes associated with the immune surveillance and immune rejection of cancer are less likely to characterize the differences in the biology of K7M2 and K12.
Using a well-characterized murine model, a microarray that includes a large and diverse number of cDNA probes, and a functional approach to analyze microarray outliers, we have defined several potentially important genes associated with pulmonary metastasis in OSA. Genes identified in this work include those not described previously in OSA as well as potentially novel metastasis-associated genes (ESTs). Functional studies suggest that 10 genes associated with motility/cytoskeleton, heterotypic adherence, and angiogenesis are most likely to be associated with differences in the metastatic behavior of the high and low metastatic OSA model. Genes identified in this analysis may have relevance to OSA and other solid tumors with high rates of pulmonary metastasis. A motility, adherence, and invasion gene called ezrin was identified using this approach. The differential expression of ezrin protein was confirmed in K7M2 and K12 cells and the potential relevance of ezrin in human OSA suggested by finding its expression in five of five human OSA cell lines. This work represents a rationale approach to the evaluation of microarray data and will be useful to identify genes that may be causally associated with metastasis.
|
FOOTNOTES |
|---|
1 To whom requests for reprints should be addressed, at Pediatric Oncology Branch, National Cancer Institute, Bethesda, Maryland 20892. 
2 The abbreviations used are: OSA, osteosarcoma; ERM, ezrin, radixin, and moesin; EST, expressed sequence tag; MMP, matrix metalloproteinase; scid, severe combined immunodeficient.
3 Internet address: http://www.nhgri.nih.gov/DIR/LCG/arraydb/.
4 Internet address: http://www.ncbi.nlm.nih.gov/cgi-bin/COG/palog?fun=all.
Received 12/29/00. Accepted 3/19/01.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H. Osawa, C. A. Smith, Y. S. Ra, P. Kongkham, and J. T. Rutka The role of the membrane cytoskeleton cross-linker ezrin in medulloblastoma cells Neuro-oncol, January 1, 2009; 11(4): 381 - 393. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. D. Leaner, J. F. Chick, H. Donninger, I. Linniola, A. Mendoza, C. Khanna, and M. J. Birrer Inhibition of AP-1 Transcriptional Activity Blocks the Migration, Invasion, and Experimental Metastasis of Murine Osteosarcoma Am. J. Pathol., January 1, 2009; 174(1): 265 - 275. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Monni, L. Haddaoui, A. Naba, I. Gallais, M. Arpin, P. Mayeux, and F. Moreau-Gachelin Ezrin is a target for oncogenic Kit mutants in murine erythroleukemia Blood, March 15, 2008; 111(6): 3163 - 3172. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Cantiani, M. C. Manara, C. Zucchini, P. De Sanctis, M. Zuntini, L. Valvassori, M. Serra, M. Olivero, M. F. Di Renzo, M. P. Colombo, et al. Caveolin-1 Reduces Osteosarcoma Metastases by Inhibiting c-Src Activity and Met Signaling Cancer Res., August 15, 2007; 67(16): 7675 - 7685. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Gordon, N. V. Koshkina, S.-F. Jia, C. Khanna, A. Mendoza, L. L. Worth, and E. S. Kleinerman Corruption of the Fas Pathway Delays the Pulmonary Clearance of Murine Osteosarcoma Cells, Enhances Their Metastatic Potential, and Reduces the Effect of Aerosol Gemcitabine Clin. Cancer Res., August 1, 2007; 13(15): 4503 - 4510. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Cavanna, E. Pokorna, P. Vesely, C. Gray, and D. Zicha Evidence for protein 4.1B acting as a metastasis suppressor J. Cell Sci., February 15, 2007; 120(4): 606 - 616. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. C. Whiteford, S. Bilke, B. T. Greer, Q. Chen, T. A. Braunschweig, N. Cenacchi, J. S. Wei, M. A. Smith, P. Houghton, C. Morton, et al. Credentialing Preclinical Pediatric Xenograft Models Using Gene Expression and Tissue Microarray Analysis Cancer Res., January 1, 2007; 67(1): 32 - 40. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. G. Marguiles, V. S. Klimberg, S. Bhattacharrya, D. Gaddy, and L. J. Suva Genomics and Proteomics of Bone Cancer. Clin. Cancer Res., October 15, 2006; 12(20): 6217s - 6221s. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Terabe, C. Khanna, S. Bose, F. Melchionda, A. Mendoza, C. L. Mackall, L. J. Helman, and J. A. Berzofsky CD1d-Restricted Natural Killer T Cells Can Down-regulate Tumor Immunosurveillance Independent of Interleukin-4 Receptor-Signal Transducer and Activator of Transcription 6 or Transforming Growth Factor-{beta}. Cancer Res., April 1, 2006; 66(7): 3869 - 3875. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Yu, E. Davicioni, T. J. Triche, and G. Merlino The Homeoprotein Six1 Transcriptionally Activates Multiple Protumorigenic Genes but Requires Ezrin to Promote Metastasis Cancer Res., February 15, 2006; 66(4): 1982 - 1989. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. de Silva Rudland, L. Martin, C. Roshanlall, J. Winstanley, S. Leinster, A. Platt-Higgins, J. Carroll, C. West, R. Barraclough, and P. Rudland Association of S100A4 and Osteopontin with Specific Prognostic Factors and Survival of Patients with Minimally Invasive Breast Cancer Clin. Cancer Res., February 15, 2006; 12(4): 1192 - 1200. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S.-W. Kim, M. Hayashi, J.-F. Lo, C. Fearns, R. Xiang, G. Lazennec, Y. Yang, and J.-D. Lee Tid1 Negatively Regulates the Migratory Potential of Cancer Cells by Inhibiting the Production of Interleukin-8 Cancer Res., October 1, 2005; 65(19): 8784 - 8791. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Wan, A. Mendoza, C. Khanna, and L. J. Helman Rapamycin Inhibits Ezrin-Mediated Metastatic Behavior in a Murine Model of Osteosarcoma Cancer Res., March 15, 2005; 65(6): 2406 - 2411. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Cao, Z. Zhou, B. de Crombrugghe, K. Nakashima, H. Guan, X. Duan, S.-F. Jia, and E. S. Kleinerman Osterix, a Transcription Factor for Osteoblast Differentiation, Mediates Antitumor Activity in Murine Osteosarcoma Cancer Res., February 15, 2005; 65(4): 1124 - 1128. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. L. Eckhardt, B. S. Parker, R. K. van Laar, C. M. Restall, A. L. Natoli, M. D. Tavaria, K. L. Stanley, E. K. Sloan, J. M. Moseley, and R. L. Anderson Genomic Analysis of a Spontaneous Model of Breast Cancer Metastasis to Bone Reveals a Role for the Extracellular Matrix Mol. Cancer Res., January 1, 2005; 3(1): 1 - 13. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. J. Belbin, B. Singh, R. V. Smith, N. D. Socci, V. B. Wreesmann, M. Sanchez-Carbayo, J. Masterson, S. Patel, C. Cordon-Cardo, M. B. Prystowsky, et al. Molecular Profiling of Tumor Progression in Head and Neck Cancer Arch Otolaryngol Head Neck Surg, January 1, 2005; 131(1): 10 - 18. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Zou, K. S. Famulski, R. S. Parhar, E. Baitei, F. A. Al-Mohanna, N. R. Farid, and Y. Shi Microarray Analysis of Metastasis-Associated Gene Expression Profiling in a Murine Model of Thyroid Carcinoma Pulmonary Metastasis: Identification of S100A4 (Mts1) Gene Overexpression as a Poor Prognostic Marker for Thyroid Carcinoma J. Clin. Endocrinol. Metab., December 1, 2004; 89(12): 6146 - 6154. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Pinkas, S. S. Martin, and P. Leder Bcl-2-Mediated Cell Survival Promotes Metastasis of EpH4 {beta}MEKDD Mammary Epithelial Cells Mol. Cancer Res., October 1, 2004; 2(10): 551 - 556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Abdel-Ghany, H.-C. Cheng, R. C. Elble, H. Lin, J. DiBiasio, and B. U. Pauli The Interacting Binding Domains of the {beta}4 Integrin and Calcium-activated Chloride Channels (CLCAs) in Metastasis J. Biol. Chem., December 5, 2003; 278(49): 49406 - 49416. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Gorlick, P. Anderson, I. Andrulis, C. Arndt, G. P. Beardsley, M. Bernstein, J. Bridge, N.-K. Cheung, J. S. Dome, D. Ebb, et al. Biology of Childhood Osteogenic Sarcoma and Potential Targets for Therapeutic Development: Meeting Summary Clin. Cancer Res., November 15, 2003; 9(15): 5442 - 5453. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. B. Sneddon, C. A. Syme, A. Bisello, C. E. Magyar, M. D. Rochdi, J.-L. Parent, E. J. Weinman, A. B. Abou-Samra, and P. A. Friedman Activation-independent Parathyroid Hormone Receptor Internalization Is Regulated by NHERF1 (EBP50) J. Biol. Chem., October 31, 2003; 278(44): 43787 - 43796. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Kakiuchi, Y. Daigo, T. Tsunoda, S. Yano, S. Sone, and Y. Nakamura Genome-Wide Analysis of Organ-Preferential Metastasis of Human Small Cell Lung Cancer in Mice Mol. Cancer Res., May 1, 2003; 1(7): 485 - 499. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Houghton, P. C. Adamson, S. Blaney, H. A. Fine, R. Gorlick, M. Haber, L. Helman, S. Hirschfeld, M. G. Hollingshead, M. A. Israel, et al. Testing of New Agents in Childhood Cancer Preclinical Models: Meeting Summary Clin. Cancer Res., December 1, 2002; 8(12): 3646 - 3657. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Bendre, D. Gaddy-Kurten, T. Mon-Foote, N. S. Akel, R. A. Skinner, R. W. Nicholas, and L. J. Suva Expression of Interleukin 8 and not Parathyroid Hormone-related Protein by Human Breast Cancer Cells Correlates with Bone Metastasis in Vivo Cancer Res., October 1, 2002; 62(19): 5571 - 5579. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Abdel-Ghany, H.-C. Cheng, R. C. Elble, and B. U. Pauli Focal Adhesion Kinase Activated by beta 4 Integrin Ligation to mCLCA1 Mediates Early Metastatic Growth J. Biol. Chem., September 6, 2002; 277(37): 34391 - 34400. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Stapleton, A. Malliri, and B. W. Ozanne Downregulated AP-1 activity is associated with inhibition of Protein-Kinase-C-dependent CD44 and ezrin localisation and upregulation of PKC theta in A431 cells J. Cell Sci., January 7, 2002; 115(13): 2713 - 2724. [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 |
, K7M2;
, K12. Bars, SD.
