Introduction

ParvA (-parvin) and ParvB (-parvin) are ubiquitously expressed members of a recently identified family of adaptor proteins, which have been implicated in integrin-mediated cell adhesion through binding to the ILK. ParvA has also been called actopaxin or CH-ILKBP1, and ParvB is also known as affixin (Nikolopoulos and Turner, 2001; Tu et al., 2001; Yamaji et al., 2001). Interactions of ParvA or ParvB with ILK are mediated by one of two conserved calponin homology (CH) domains, a structural motif that also mediates protein interactions with actin (Banuelos et al., 1998; Gimona and Winder, 1998). The Caenorhabditis elegans Parvin/Pat-6 homologue interacts genetically with ILK/Pat-4 and is essential for the assembly of integrin/Pat-3-dependent muscle attachment structures (Mackinnon et al., 2002; Lin et al., 2003), which are analogous to focal adhesions of mammalian cells (Hresko et al., 1994; Williams and Waterston, 1994). In worms, parvin/Pat-6 is an essential protein, which functionally links the actin cytoskeleton to integrin-mediated cell adhesion.

ParvA and ParvB are encoded by different genes in humans and their highly conserved intron–exon structure suggests that they arose from a relatively recent gene duplication event. The human ParvA gene maps to chromosome 11p15.5 and ParvB maps to chromosome 22q13.31 (Korenbaum et al., 2001; Olski et al., 2001). Regulation of ParvB expression is complex, and alternative transcripts involving differential 5' and 3' splicing have been identified (Korenbaum et al., 2001; Olski et al., 2001). The biological significance of different ParvB proteins is unknown; however, all known isoforms retain the tandem arrangement of CH1 and CH2 domains, and thus, are all likely to be capable of binding to ILK. BLAST searches (Altschul et al., 1997) indicate C. elegans and Drosophila melanogaster each have a single parvin gene, implying the evolution of distinct functions of ParvA and ParvB in mammals. In addition to modulating cell adhesion, ParvA potentiates ILK signaling (Attwell et al., 2003; Fukuda et al., 2003). However, microdeletions near the ParvB locus on chromosome 22q13.31 are associated with colon and breast cancers (Castells et al., 2000), suggesting that ParvB expression may be dysregulated in some cancers.

Two key regulatory targets of ILK signaling are the serine/threonine (S/T) kinases, PKB and glycogen synthase kinase 3 (GSK3) (Delcommenne et al., 1998), both of which regulate cell proliferation and apoptosis (Coffer et al., 1998; Kim and Kimmel, 2000). ILK-mediated phosphorylation of PKB on S473 activates PKB and increases cell survival (Attwell et al., 2000), while phosphorylation of GSK3 on S9 inhibits its activity, thereby promoting cell proliferation through stabilization of -catenin and cyclin D1 (Radeva et al., 1997; D'Amico et al., 2000). In the MMTV-ILK transgenic mouse model of mammary carcinoma, tumors display increased levels of PKB pSer473 and GSK3 pSer9, as well as increased cyclin D1 expression (White et al., 2001). Silencing of either ILK or ParvA expression inhibits PKB S473 and GSK3 S9 phosphorylations, and thus, ParvA appears to play a role in promoting ILK signaling of these downstream kinases (Attwell et al., 2003; Fukuda et al., 2003).

Inhibition of cellular ILK activity by small molecules or dominant-negative mutants suppresses the transformed phenotype (Persad and Dedhar, 2003), stimulating interest in ILK as a target for anticancer drug development (Edwards et al., 2004). Physiologic inhibitors of ILK signaling include PTEN, a lipid phosphatase that is the major antagonist of PI3K activity (Myers et al., 1998; Maehama and Dixon, 1999; Mills et al., 2001; Leslie and Downes, 2002). Dephosphorylation of PI(3,4,5)P₃ by PTEN indirectly inhibits ILK activity, which is accordingly elevated in PTEN null cells (Persad et al., 2001). We have recently discovered ILKAP, a S/T phosphatase that selectively binds to and inhibits ILK, blocking phosphorylation of GSK3 S9 and suppressing downstream transactivation by -catenin/TCF factors (Leung-Hagesteijn et al., 2001). Herein, we report the inhibition of ILK signaling and suppression of cellular transformation by ParvB, and provide new evidence that it is downregulated in advanced breast cancers.

Results

Identification and genomic characterization of ParvB variants

In a yeast two-hybrid screen for ILK-interacting proteins, we identified a cDNA corresponding to ParvB. In order to generate ParvB expression vectors for functional studies, we amplified ParvB mRNA from a human rhabdomyosarcoma (RMS) cell line, and cloned a ParvB cDNA sequence with an open-reading frame encoding a protein sequence of 397 amino acids ('CLINT', GenBank Accession number AAL08219). This product was subcloned into pcDNA3.1, as described in Materials and methods. Other deposited ParvB cDNAs translate to protein sequences of 350 (Accession number AAH46103) and 364 (Accession number AAG27171) amino-acid residues, indicating we had identified a third ParvB isoform. We refer to these isoforms as ParvB1–3, in order of increasing length; thus, the isoform we identified as CLINT is hereafter called ParvB3.

Genomic analysis indicated that ParvB3 arises from transcription of two unique 5' exons, located on chromosome 22q13.1 (Figure 1a). The three ParvB isoforms share an identical COOH region of 327 amino-acid residues, comprised mainly of the tandem CH domains, encoded by exons 3–13 (Figure 1a). These isoforms diverge in their NH₂-terminal regions of 23, 37 and 70 amino acids (Figure 1b), arising through differential splicing (ParvB3), or differential translation start site usage (ParvB1 and B2) (Figure 1a). In order to confirm that ParvB3 is a bona fide gene product, we used ParvB3-specific primers to amplify RNA isolated from HEK293, RMS and PC3 prostate carcinoma cells. In all these cells, we amplified a specific ParvB3 product, confirming expression of ParvB3 (Figure 1c).

Figure 1.

ParvB genomic organization, and expression of different isoforms in human cells. (a) Structure of ParvB gene locus at chromosome 22q13.1 is indicated schematically. Novel ParvB3 exons 1 and 2 are indicated. The Genethon marker, D22S1171, delineates the centromeric boundary of deletions associated with breast cancers (Castells et al., 2000). Genomic clones are indicated above gene schematic, with clone sizes (kilobase pairs) within parentheses. Exon 14 is untranslated. (b) Amino-acid residues 1–70 of ParvB3 are encoded by exons 1 and 2. Sequence alignment was performed using Gene Inspector v. 1.5.11 (Textco) DNA analysis software. (c) ParvB3-specific RT–PCR was used to amplify total RNA (1 g) isolated from HEK293, RMS and PC3 prostate carcinoma cells

Full figure and legend (77K)

ParvB mRNA expression is downregulated in breast cancer cell lines and tumors

We performed quantitative RT–PCR on RNA isolated from nine patient-matched breast tumors (Bloom–Richardson Grade 2 or 3) and normal mammary tissues, using ParvB primers that would recognize all known ParvB isoforms (pan-ParvB). ParvB mRNA was downregulated by 60% or greater, in four tumors, compared to their respective matched normal mammary gland levels (Figure 2a). In order to pursue functional studies of ParvB, we characterized ParvB expression in a number of well-studied breast cancer cell lines, using the pan-ParvB primers. We observed two RT–PCR products of 650 and 500 basepairs (bp) in the normal mammary gland, as well as MCF7, MCF10A, SK-BR-3, T-47D, MDA-MB-435 and MDA-MB-436 cells. These products were sequenced and verified as ParvB. Interestingly, the 650 bp product was preferentially amplified in Hs578T, BT549 and BT474 cells. We did not detect any amplified ParvB product in two tumorigenic breast cancer cell lines MDA-MB-231 and ZR-75-1. RT–PCR amplification of 2-microglobulin control RNA was equally robust in all cell lines (Figure 2b). Thus, ParvB is expressed in the normal mammary gland, and our data suggest that it is downregulated in advanced breast cancers, and breast cancer cell lines.

Figure 2.

ParvB mRNA transcipt levels are reduced in human breast tumors and cell lines. (a) Q-RT–PCR of ParvB expression in paired human breast tumors and the adjacent normal mammary gland (see Materials and methods). Error bars indicate standard deviation for triplicate determinations. Pairwise t-test across nine samples indicated significant reduction of ParvB transcript levels (P=0.005). (b) RT–PCR analysis of ParvB expression was performed for the human mammary gland and breast cell lines, using primers spanning the terminal four exons (10–13) on the 3'-end of the PARVB cDNA (pan-ParvB). All RT minus control samples revealed no PCR product. L1 is normal liver. M1 is liver metastasis

Full figure and legend (90K)

As the expression of ParvB was significantly downregulated in a number of breast tumors, we next wished to examine the potential biological significance of this downregulation to breast cancer progression. In order to pursue functional studies of ParvB in breast cancer cells, we transfected 'empty' pcDNA3.1 or pcDNA-ParvB3 (full-length) into the poorly differentiated MDA-MB-231 line, and into the more epithelial-like MCF7 cells. Stable transfectants of each line were selected. Isoform-specific RT–PCR indicated that ParvB1 and B2 transcripts were expressed in MCF7 but not MDA-MB-231 cells. Neither parental cell line expressed detectable levels of ParvB3; therefore, RT–PCR readily confirmed the expression of the exogenous ParvB3 gene in both MCF7 and MDA-MB-231 transfectant lines. In addition, both cell lines expressed ParvA (Figure 3). These lines were used to assay modulation of oncogenic cell behaviors by ParvB.

Figure 3.

Expression of ParvB3 in transfected breast cell lines. (a) RT–PCR analysis of total RNA (1 g) isolated from MDA/ParvB3, MCF/ParvB transfectants and vector control cells, using ParvB3- or PARVA-specific primers. (b) RT–PCR was performed on ParvB3-transfected cells, using ParvB1- or ParvB2-specific primers, as described in Materials and methods

Full figure and legend (49K)

Identification of ParvB-deficient breast cancer cell lines

To facilitate the study of ParvB protein expression, we raised a polyclonal antibody against recombinant ParvB3, containing residues 1–142, which include the common CH1 domain (Figure 4). HEK293 cell lysates were subjected to Western blotting using antibody preparations that had been preadsorbed, either with ParvB3 or GST recombinant proteins. The affinity purified antibody recognized endogenous proteins of approximately 40 and 45 kDa in 12% SDS–PAGE of HEK293 cell lysates, probably representing two ParvB isoforms (Figure 4a). Preadsorption with ParvB3, but not GST, abrogated antibody recognition of these proteins. GFP-tagged ParvB3 fusion proteins were transiently expressed in HEK293 cells in order to further test for antibody specificity. The affinity purified antibody recognized recombinant full-length ParvB3 protein, residues 1–142, but not the COOH-terminal, residues 157–397, thus confirming its specificity (Figure 4b and c). These results confirmed that our antibody detected endogenous and exogenous ParvB proteins.

Figure 4.

Expression of endogenous and recombinant ParvB protein. (a) To determine specificity, affinity purified ParvB antibody (raised against recombinant ParvB3 residues 1–142, including the common CH1 domain) was incubated with recombinant ParvB3 or GST protein, then used to probe 50 g HEK293 cell lysates resolved on 12% SDS–PAGE. The anti-ParvB antibody specifically detected two endogenous protein bands (45 and 40 kDa) that can be absorbed by ParvB3-GST fusion peptide. (b) Cellular proteins (50 g/lane) from parental and EGFP-ParvB3 HEK293 transient transfectants were resolved by 12% SDS–PAGE. Affinity purified ParvB antibody detected the exogenous ParvB3 and CH1 EGFP proteins, in addition to the endogenous ParvB protein bands. The membranes were subsequently probed with GFP and GAPDH antibodies. (c) Domain structure of ParvB3. Full-length ParvB3 cDNA (residues 1–316) and two truncated variants, CH 1 and CH 2, corresponding to residues 1–142 and 157–316, respectively, were cloned into pEGFP-C3 vector as described in Materials and methods

Full figure and legend (76K)

The observed downregulation of ParvB mRNA in breast tumors suggested that ParvB protein levels would be reduced as well. Using the affinity purified ParvB antibody, we detected very low protein levels in lysates from MDA-MB-231 and MCF7 cells compared to HEK293 or PC3 prostate carcinoma cells (Figure 5a). Interestingly, ILK protein levels were higher in the more invasive MDA-MB-231 line, compared to the MCF7 cells, consistent with its role in promoting cell invasiveness. We then wished to analyse breast tumors for the downregulation of ParvB protein. Patient-matched, normal and tumor tissue lysates were subjected to Western blots and ParvB signals analysed by densitometry. ParvB protein levels were decreased by >90% in five of seven tumors, relative to the levels in patient-matched normal breast tissue (Figure 5b–d). These blots were then reprobed for ILK levels. Reduced ParvB protein levels correlated with increased ILK protein in four of the seven tumor samples. To confirm that increased ILK protein correlated with increased ILK activity, three matched tumor/normal pairs were subjected to ILK immune complex kinase assays. In accord with increased protein levels, ILK activity was elevated in these three tumors (Figure 5c), suggesting that in some breast cancers, downregulation of ParvB contributes to the upregulation of ILK signaling.

Figure 5.

ParvB protein expression is downregulated in human breast tumors and breast cancer cell lines. Whole-cell lysates (50 g/lane) of (a) human cell lines and (b) paired normal and tumor breast tissue lysates were subjected to 12% SDS–PAGE and immunoblotted with ParvB antibody. Blots were then sequentially stripped and reprobed with ILK and GAPDH antibodies. Values represent ratios of tumor to normal ParvB levels, determined by densitometry of ParvB doublet bands, each internally normalized against the GAPDH signal (see Materials and methods). (c) Normal and tumor breast tissue lysates (700  g each) were analysed in an ILK immune complex kinase assay using MBP as exogenous substrate. (d) A total of seven patient-matched tumor/normal mammary gland samples were analysed for ParvB protein levels. Results are presented as relative levels for each patient sample, normalized to GAPDH expression, by densitometry. The three left-most samples are derived from the blots in (b)

Full figure and legend (89K)

ParvB inhibits anchorage-independent growth of MCF7 and MDA-MB-231 cells

The inverse correlation between ParvB and ILK levels in breast tumors prompted us to ask whether ParvB can suppress the transformed cell phenotype. Moreover, as ParvB is an ILK-binding protein, we examined the potential of ParvB to inhibit oncogenic ILK signaling. For these studies, we used the MDA-MB-231/ParvB and MCF7/ParvB transfectant cell lines. Lysates of stable transfectant cell lines were analysed for ParvB protein expression. Western blotting with either affinity purified ParvB antibodies (Figure 6a) or anti-myc (not shown) confirmed appropriate expression from the ParvB3 plasmid.

Figure 6.

ParvB expression suppresses anchorage-independent cell growth. (a) ParvB3 recombinant protein expression in MDA-MB-231 and MCF7 cells stably expressing ParvB3. Cell lysates (50 g/lane) were analysed by Western blotting with ParvB antibody. Membranes were then stripped and reprobed with anti-GAPDH antibodies. (b) Stable vector and ParvB cells (5 10³/well) were incubated in soft agar for 13 days at 37°C. Colonies were counted in five randomly selected fields ( 10 objective) per well. Bars represent the mean colonies/fields.e.m. of triplicate wells (P<0.05). (c) Vector control and MDA/ParvB3 cells were cultured on collagen I and assayed for differences in proliferation using the MTT dye assay. Similar results were obtained comparing cells that were plated on tissue culture plastic. These results are representative of two independent determinations, with error bars denotings.e.m. of quadruplicate wells

Full figure and legend (67K)

As ILK induces anchorage-independent growth (Hannigan et al., 1996), we examined ParvB effects on anchorage-independent growth of MDA-MB-231 and MCF7 cells. Compared with vector controls, both the MDA/ParvB3 and MCF7/ParvB3 cell lines demonstrated significant suppression of colony formation in soft agar (>60% inhibition, Figure 6b). We then asked if the growth suppressive effect is specific to anchorage-independent growth. MTT dye assays indicated that proliferation of adherent MDA/ParvB cultures was not inhibited (Figure 6c) nor did the cell cycle profile of adherent MDA/ParvB3 cells differ from the vector controls (Table 1). We conclude that ParvB does not affect the intrinsic proliferative capacity of these cells. These results indicate that ParvB suppresses anchorage-independent growth of both MDA-MB-231 and MCF7 cell lines.

Table 1 - ParvB3 does not effect MDA-MB-231 cell cycle distribution.

Full table

We originally reported that the expression of ILK in epithelial cells decreased cell adhesion to extracellular matrix (ECM) proteins (Hannigan et al., 1996), and ILK has also been shown to stimulate ECM invasion by epithelial cells (Troussard et al., 2000). We therefore directly assayed the effects of ParvB3 expression on cell adhesion and ECM invasion. First, we seeded vector control and ParvB3 transfectants into collagen I-coated microwells and quantified the relative proportion of cells attached to the substrate after 1 h. ParvB expression increased MDA-MB-231 cell adhesion to collagen by about twofold and increased MCF7 adhesion by about 1.5-fold (Figure 7a). These effects on cell adhesion are similar to what has been reported for ParvB in other cells (Yamaji et al., 2001).

Figure 7.

ParvB inhibits matrigel invasion by MDA-MB-231 cells. (a) Vector () and ParvB3 () cells were assayed for adhesion to ECM. Cells were plated on collagen I for 1 h at 37°C, and adherent cells visualized by crystal violet staining. Error bars indicate s.e.m. of four experimental values (P<0.05). (b) MDA/ParvB3 and vector control cells were plated on Matrigel, in the upper chamber of invasion plate wells, with bottom chambers containing medium with or without 100 ng/ml EGF. Results are presented as number of cells per microscope field ( 10 objective) from five randomly selected fields. Data are expressed as the mean cells/fields.e.m. of three chambers (P<0.05). Each experiment is representative of two independent determinations

Full figure and legend (16K)

We were then interested to examine the ability of ParvB3 to suppress cell invasion through an ECM. Although derived from a metastatic site, MCF7 cells retain an epithelioid morphology with intact cell–cell junctions, and are non-invasive. Conversely, MDA-MB-231 cells are highly invasive. We tested ParvB3 effects on ECM (Matrigel) invasion by MDA-MB-231 cells, using a modified Boyden chamber assay. EGF is a potent chemoattractant for the MDA-MB-231 cells (Price et al., 1999; Andl et al., 2003); thus, we determined whether ParvB3 modulates EGF-stimulated invasiveness in these cells. Cells were seeded onto Matrigel-coated membranes in the top chamber of each well, and allowed to invade for 20 h through the Matrigel toward the bottom chamber, containing medium with or without 100 ng/ml EGF. The vector control and parental cells showed similar numbers of invaded cells, while the number of invading MDA/ParvB3 cells was suppressed by approximately 50% (Figure 7b). EGF stimulated control cell invasiveness by an additional 50%; however, it did not stimulate invasiveness of MDA/ParvB3 transfectants. These results demonstrate that ParvB3 can suppress basal and EGF-stimulated ECM invasion by breast tumor cells.

ParvB inhibits EGF-induced phosphorylation of cellular ILK targets, PKB and GSK3

Inhibition of EGF-stimulated invasiveness suggests that ParvB inhibits signaling downstream of the EGF receptor (EGFR). ILK stimulates ECM invasion of mammary epithelial cells (Troussard et al., 2000), and we have found that EGF is a potent inducer of ILK activity in kidney epithelial cells (unpublished data). Therefore, we sought to determine whether ParvB inhibited ILK signaling in MDA/ParvB3 and MCF7/ParvB3 cells. For these studies, we examined EGF induction of GSK3 Ser9 phosphorylation by phosphospecific Western blot analysis of the MDA/ParvB3, MCF7/ParvB3 and respective vector control cells. These experiments showed that EGF-induced GSK3 Ser9 phosphorylation was suppressed by about 70% in MDA/ParvB3 cells, and by about 30% in MCF7/ParvB3 cells, compared to their respective vector controls (Figure 8a). This quantitative difference likely reflects the substantially lower levels of ILK- and EGF-induced GSK3 Ser9 phosphorylation observed in the MCF7 cells compared to MDA-MB-231 (Figures 5 and 8a).

Figure 8.

ParvB3 suppresses EGF signaling and ILK kinase activity in breast carcinoma cells. (a) MDA/ParvB3, MCF7/ParvB3 and vector control cells were serum starved for 24 h, and stimulated or not with EGF (100 ng/ml) for 10 min. Total cellular proteins (50 g/lane) were resolved by 12% SDS–PAGE. Levels of phosphorylated GSK3 (GSK3-ser9), total GSK3 (GSK3), ILK and GAPDH were determined by Western blot. These results represent four independent determinations. (b) Vector control and ParvB3 MDA-MB-231 cells were serum starved and stimulated with or without EGF. Cytoplasmic lysates (700 g) were analysed in an ILK immune complex kinase assay using MBP as exogenous substrate

Full figure and legend (59K)

We next wished to determine if ParvB inhibits signaling at the level of ILK activity, thereby explaining the decreased levels of GSK3 phosphorylation. We therefore assayed EGF-induced ILK immune complex kinase activity in MDA/vector and MDA/ParvB3 cells. EGF rapidly (<15 min) induced ILK activity in the vector control cells; however, there was no detectable increase of ILK activity in EGF-treated MDA/ParvB3 cells (Figure 8b). Thus, ParvB-mediated inhibition of GSK3 signaling appears to be due to suppression of ILK kinase activity.

In order to test further the role of ParvB in suppressing EGF and ILK signaling, we infected cells with adenovirus expressing ILK. Infection of MDA/vector control and MDA/ParvB3 cells with Ad-ILK resulted in an approximately threefold overexpression of ILK (Figure 9a). We assayed for ILK signaling in the Ad-ILK-infected cells by Western blots, using antibodies to phospho-GSK3(S9) and phospho-PKB(S473). Infection of vector control cells with Ad-ILK induced PKB(S473) and GSK3(S9) phosphorylation, as expected. However, Ad-ILK- or EGF-induced GSK3(S9) phosphorylation in MDA/ParvB3 cells was only to about 30% of the levels that were induced in the vector control cells. There was a similar trend toward inhibition of ILK-induced phosphorylation of PKB(S473); however, it was consistently less pronounced than the effect on GSK3 signaling (Figure 9b). Phosphorylation of PKB and GSK3 was effectively suppressed in Ad-ILK-infected, EGF-treated MDA/ParvB cells.

Figure 9.

ParvB3 is an inhibitor of EGFR and ILK signaling. (a) MDA/ParvB3 and vector control cells were infected with Ad-ILK at an MOI of 0, 1, 5 or 10 and immunoblotted for ILK expression after 48 h. (b) MDA/ParvB3 and vector control cells were infected with Ad-ILK for 24 h. After infection, cells were serum starved overnight and treated with or without EGF as indicated for 10 min. Total cellular proteins (50 g/lane) were analysed by immunoblotting with antibodies recognizing ILK, phosphorylated GSK3 (GSK3-ser9) and PKB (PKB-ser473), total GSK3 (GSK3) and PKB (PKB), and GAPDH. (c) EGFR activation was confirmed by Western blotting with phospho-EGFR(Tyr1068) and EGFR antibodies. Each experiment represents at least two independent determinations

Full figure and legend (99K)

The possibility of integrin and EGFR crosstalk (Yu et al., 2000) suggests that ParvB3 could be inhibiting signaling at the level of EGFR activation, and thus, we examined EGFR expression and activation in the MDA/ParvB3 cells. ParvB did not affect EGFR expression level, or EGF-induced phosphorylation at Tyr1068, indicating normal ligand-induced EGFR activation. These controls also indicated that ILK overexpression had no effect on EGFR expression or activation (Figure 9c). Together, with data showing inhibition of ILK kinase activity (Figure 8b), these results indicate that ParvB inhibits EGF-ILK signaling at the level of ILK. Directed yeast two-hyrid assays confirmed that the ParvB3 protein interacts with ILK, requiring an intact ILK catalytic domain (Figure 10). Thus, ParvB may inhibit ILK activity through direct association with catalytic domain residues.

Figure 10.

ParvB3 directly associates with ILK. (a) EGY48 strain transformed with plasmids encoding ParvB3 cDNA, lacZ reporter and the indicated ILK baits. LexILK^WT is kinase active, while LexILKE359K is a kinase-deficient variant, due to a point mutation in the catalytic domain. LexILK^cat is a truncated ILK (containing residues 1–276), lacking the catalytic domain. Both ILKE359K and ILK^cat mutants showed very weak interactions, as compared to ILK^WT, indicated by the decreased intensity of blue colonies. (b) Expression of LexA bait proteins in EGY48. Western blotting of transformed yeast lysates using a LexA monoclonal antibody indicated equivalent expression of ILK^WT, ILKE359K and ILK^cat bait proteins. Results shown are representative of six colonies, isolated from each of the three independent cotransformations of each ILK bait plasmid

Full figure and legend (64K)

Discussion

We have identified a novel function of ParvB, to restrict ILK signal transduction, and have shown that loss of this physiologic ILK inhibitor occurs in some advanced breast cancers. ILK protein levels and activity are upregulated in a number of cancers, including Ewing's sarcoma (Chung et al., 1998) and melanoma (Dai et al., 2003), as well as prostate (Graff et al., 2001), gastric (Ito et al., 2003), colon (Marotta et al., 2001; Marotta et al., 2003) and ovarian (Ahmed et al., 2003) carcinomas. In prostate cancer cells, loss of the PTEN tumor suppressor upregulates ILK signaling (Persad et al., 2001). PTEN is frequently inactivated in prostate carcinoma (Cairns et al., 1997; Li et al., 1997), indicating that loss of an indirect ILK inhibitor contributes to cancer progression. ParvB may inhibit ILK activity through a direct interaction. Therefore, ParvB loss may be a significant progression factor in breast, and possibly other cancers, where PTEN is not frequently inactivated. Importantly, targeted expression of ILK in the mammary gland of transgenic mice induces focal tumor formation (White et al., 2001), indicating that mammary epithelium is sensitive to oncogenic ILK signaling. As our data suggest loss of ParvB is one mechanism (others including loss of PTEN or ILKAP, for example) that may contribute to elevated ILK activity in tumors, it will be important to determine the significance of this association from larger samples of breast and other cancers.

Parvins are important regulators of integrin-mediated cell adhesion through a conserved interaction with ILK that serves to recruit parvins to sites of cell attachment (Tu et al., 2001; Nikolopoulos and Turner, 2002; Lin et al., 2003). They have also been implicated in regulating cell shape changes and ECM-directed cell migration, highlighting their involvement with integrin-dependent actin reorganization at the cell periphery (Zhang et al., 2002). Online gene expression data (http://nciarray.nci.nih.gov/ca
rds/) indicate that ParvA and ParvB are widely coexpressed. Recent evidence, including that presented herein, indicates that parvin regulation of cell function is complex. Gene expression studies and siRNA-mediated silencing of ParvA have shown that it promotes cell proliferation and survival, through increased ILK signaling of PKB and GSK3 (Attwell et al., 2003; Fukuda et al., 2003). Both ParvA and ParvB show similar, CH2 domain-dependent binding to the COOH terminal region of ILK (Nikolopoulos and Turner, 2002; Yamaji et al., 2001), thus raising the question of how these molecules differentially regulate ILK signaling. It is possible that ParvA and ParvB exert antagonistic effects on signaling through competitive interactions with ILK. Alternatively, ILK-mediated phosphorylation of each might have opposite effects in promoting ILK-ParvA and ILK-ParvB complexes, under conditions of growth factor or integrin-activated ILK signaling. Indeed, it has been shown that inhibition of ILK activity dissociates ILK–ParvA complexes, suggesting that ILK activity is required for positive signaling through ParvA (Attwell et al., 2003). Our data, demonstrating negative regulation of ILK by ParvB in breast cancer cells, suggests these two parvins may be antagonistic regulators of ILK.

Regulation of Parvin expression in mammals is complex, and both ParvA and ParvB gene products express multiple isoforms, generated through differential splicing and polyadenylation (Korenbaum et al., 2001; Olski et al., 2001). The unique N-terminal ParvB3 residues we have described are encoded on two exons in the ParvB locus on chromosome 22q13.2–13.3 (Figure 2a). Online databases indicate that these unique exons match expressed sequence tags from the skin, skeletal muscle, pituitary gland, brain and lung. However, the biological significance of different ParvB isoforms is unknown, raising the question of whether ILK inhibition is an isoform-specific function. We used transfection of ParvB3 for our functional studies, but the basic structure of tandemly arranged CH1 and CH2 domains is conserved in all verified ParvB variants. The CH2 domain mediates ILK binding and localization of ParvB to focal adhesion sites (Yamaji et al., 2001). Our analyses of mRNA expression utilized pan-ParvB primers for quantitative RT–PCR. It is likely that the decreased ParvB levels in breast tumors result, at least in part, from post-transcriptional downregulation, affecting all ParvB isoforms. This notion is supported by our protein data. ParvB Western blots were performed with an antibody that recognizes the CH1 domain, common to all ParvB isoforms. Thus, inhibition of ILK signaling is likely a common function of ParvB proteins.

We do not rule out the possibility of mutational inactivation of ParvB in breast cancer, suggesting a possible tumor suppressor role. Previously, we mapped a minimal region of deletion on chromosome 22q13.3 in breast cancer (Castells et al., 1999, 2000). The centromeric boundary of this region is demarcated by the Genethon probe, D22S1171, which lies in the intron between ParvB exons 2 and 1A (Figure 1a). Loss of this marker would indicate deletion of ParvB exons encoding the CH1 and CH2 domains. Our data suggest that this deletion would disrupt the ILK/ParvB interaction, and thereby upregulate ILK signaling. Thus, although dysregulation of actin cytoskeletal remodeling can contribute to aberrant cell migration and metastasis, modulation of ILK activity by ParvA (Attwell et al., 2003) and ParvB identifies an important, emerging role for actin-binding proteins in mediating intracellular signal transduction.

We are currently examining whether inhibition of ILK activity promotes, or if phosphorylation destabilizes, ILK-ParvB complexes as potential mechanisms for differential regulation of ILK signaling by ParvA and ParvB. In addition, it seems unlikely that ParvB inhibits all functions of ILK. The CH1 domain of ParvB mediates its interaction with -PIX (Rosenberger et al., 2003), a guanine nucleotide exchange factor for Rac1 and Cdc42 GTPases. We find that cardiac expression of ILK in transgenic mice activates Rac1 in cardiomyocytes (H Lu, G Hannigan, J Coles, unpublished data), suggesting a role for ParvB/-PIX in mediating ILK-Rac-dependent actin cytoskeletal reorganization. Thus, ParvB may positively regulate ILK-dependent events at the cell periphery, while inhibiting cytoplasmic ILK signaling.

ParvB is one of the three identified cellular inhibitors of ILK signaling, including PTEN and the S/T phosphatase, ILKAP (Leung-Hagesteijn et al., 2001; Kumar et al., 2004). ILK upregulation promotes nuclear translocation of -catenin, likely through inhibitory phosphorylation of GSK3 on Ser9 (Persad et al., 2001; Tan et al., 2001). -catenin then associates with TCF/Lef transcription factors, activating target genes such as cyclin D1 and c-myc (He et al., 1998; Novak and Dedhar, 1999; Shtutman et al., 1999; Barker and Clevers, 2000; D'Amico et al., 2000). There is a direct correlation between elevated -catenin and cyclin D1 expression levels in breast cancer (Jonsson et al., 2000; Lin et al., 2000). Our preliminary data suggest that ParvB modulates nuclear accumulation of -catenin and cyclin D1 protein in MDA-MB-231 cells (not shown). This result suggests that ParvB inhibition of ILK-GSK3 signaling suppresses -catenin transactivation, thereby blocking the transcription of cyclin D1 and c-myc genes. Both immunocytochemistry and live imaging with GFP-ParvB3 indicate that a fraction of ParvB localizes to the cell nucleus (not shown). It will be very important to identify the nuclear targets that underlie suppression of the oncogenic phenotype by ParvB.

Materials and methods

Construction of ParvB expression plasmids

For cloning into pcDNA3.1, a full-length ParvB3 cDNA was isolated from a human adult heart library (Clontech) by screening with a ParvB3 RT–PCR-derived probe. The following primers were then used to amplify cDNA inserts, using high-fidelity Pfu DNA polymerase (MBI Fermentas): 5'-GGCCGGTACCATGCACCATGTGTTTAAA (forward); and 3'-GGCCGATATCCTCCACGTTCTTGTACTT (reverse).

The PCR product was digested with KpnI and EcoRV restriction enzymes, gel purified and ligated into KpnI/EcoRV-digested pcDNA3.1 vector. Following transformation of Escherichia coli DH5, colonies were screened for the presence of inserts by KpnI and EcoRV restriction digestions. Positive clones were sequenced to verify cloning in-frame with the myc-His epitope tag.

For the expression of GFP-ParvB fusion proteins, full-length PARVB3 (residues 1–397) and two truncated cDNAs, CH1(1–142) and CH2(157–397) corresponding to residues 1–142 and 157–397, respectively, were generated. The pcDNA-ParvB3 plasmid was used as template for PCR cloning into the pEGFP-C3 expression vector (Clontech). The PCR products were generated with high-fidelity Pfu DNA polymerase (Stratagene) using the following primers: ParvB3, 5'-GGCCGAGCTCATGCACCATGTGTTTAAAG-3', 5'-GGCCCCGCGGTCACTCCACGTTCTTGTAC-3'; CH1(1–142), 5'-GGCCGAGCTCATGCACCATGTGTTTAAAG-3', 5'-GGCCCCGCGGCGTTACATGCTCAGGA-3'; and CH2(157–397), 5'-GGCCGAGCTCTCCAGCCACATCTCGGAG-3', 5'-GGCCCCGCGGTCACTCCACGTTCTTGTAC-3'.

PCR products were digested with SacI and SacII restriction enzymes (MBI), gel purified and ligated into SacI/SacII-digested pEGFP-C3 vector. Following transformation, colonies were screened for the presence of inserts by SacI and SacII restriction digestions. Positive clones were sequenced to verify cloning in-frame with the EGFP epitope tag.

Construction of ILK adenovirus

Full-length ILK cDNA (Hannigan et al., 1996) was subcloned into a shuttle vector, pAdTrack-CMV (AdEasy System, Stratagene). The plasmid was linearized by restriction digestion with PmeI and cotransformed into E. coli BJ5183 cells with pAdEasy-1. Recombinants were selected for kanamycin resistance and recombination confirmed by restriction analysis. The linearized recombinant plasmid was transfected into an adenovirus packaging cell line, HEK293, using Lipofectamine (Invitrogen) in T-25 flasks (2 10⁶ cells/flask) according to the manufacturer's instructions. Transfections and viral productions were monitored by GFP expression from the pAdTrack vector. For viral purification, cells were harvested and subjected to four freeze/thaw cycles in dry ice/methanol and centrifuged to remove cell debris. Supernatant containing virus (8 ml) was combined with 4.4 g CsCl and centrifuged at 32 000 r.p.m., 10°C, for 24 h. The virus fraction was collected and titered on HEK293 cells.

Production of polyclonal ParvB antibodies

A polyclonal -parvin (PARVB) antiserum was generated by immunizing New Zealand white rabbits with sarkosyl-solubilized GST-ParvB3 fusion protein, comprising amino-terminal residues 1–142 of ParvB3. Amino-acid numbering is in reference to the ParvB3 translation product, which we initially deposited to the GenBank as CLINT (Accession number AAL08219). For immunodepletion of GST antibodies, crude serum from the immunized rabbits was incubated with recombinant GST bound to glutathione–sepharose 4B (Pharmacia Biotech). The depleted serum was then incubated with GST-ParvB3 recombinant protein immobilized on glutathione–sepharose 4B. ParvB antibodies were subsequently eluted from the washed column with 100 mM glycine (pH 2.8) and immediately neutralized in 1 M Tris-HCl (pH 8.0) as described (Youssoufian, 1998).

Cell culture and derivation of stable cell lines

Human breast cancer cell lines were maintained in Dulbecco's -minimal essential medium (-MEM), supplemented with 10% fetal calf serum (FCS) (Invitrogen) and antibiotic/antimycotic (100 ) (Invitrogen). Wild-type ParvB pcDNA3.1/Myc-His B clone or the vector alone was introduced into MDA-MB-231 and MCF7 cells using Fugene 6 Transfection Reagent (Roche). At 2 days post-transfection, cells were replated and grown in -MEM containing 10% FCS and 1.4 mg/ml G418 (Invitrogen). After 21 days of selection, neomycin-resistant ParvB3 and empty vector cells were expanded.

Cell adhesion and matrigel invasion assays

Cell adhesion assays were performed as described previously (Hannigan et al., 1996). Briefly, cells (5 10⁵) were resuspended in DMEM supplemented with 1 HEPES (Invitrogen) and seeded in quadruplicate wells in a 96-well plate, and incubated at 37°C in a humidified CO₂ cabinet for 1 h. Cells were washed with PBS once, fixed in 5% glutaraldehyde (Sigma) for 20 min at room temperature and stained with 0.1% crystal violet solution (Sigma) at room temperature for 1 h. Cell adhesion was quantified by measuring the absorbance at optical density 570 nm using a Spectra Max 250 ELISA plate reader.

In vitro cell invasion was assayed in BD BioCoat matrigel invasion chambers (BD Biosciences, 24 wells, 8 m pore size). The top chamber was seeded with 5 10⁴ cells in DMEM. The bottom chamber was filled with DMEM supplemented with EGF (100 ng/ml) as a chemoattractant. Chambers were incubated for 20 h in a humidified tissue culture incubator, 37°C, 5% CO₂ atmosphere. Noninvasive cells were removed from the upper surface of the membrane with a cotton swab, and cells on the lower surface of the membrane were fixed and stained with Diff-Quick and mounted on glass slides. Five random fields/well ( 10 objective) were counted for quantitation of cell invasion. Triplicate wells were counted for each assay. Data were analysed using Prism v. 3.0 statistical software (GraphPad Inc.).

Analysis of cell growth and cell cycling

Adherent cell growth was assessed by MTT dye conversion at 570 nm according to the manufacturer's recommendations (Roche Applied Science). Briefly, ParvB3 or vector control cells were seeded into 96-well plates (2 10³ cells/well) and incubated in either 0.5 or 10% FBS-containing medium for 96 h. After the incubation period, MTT labeling reagent was added to each well. Following a 4-h incubation at 37°C, formazan salt crystals were solubilized overnight in a humidified atmosphere. The solubilized formazan product was quantified spectrophotometrically using an ELISA reader (OD 570 nm).

For the analysis of anchorage-independent growth, cells (5–10 10⁴) stably transfected with vector or myc-tagged ParvB3 were suspended in soft (0.3%) Bacto agar (Difco) in -MEM with 10% FBS and placed onto a solidified layer of 0.5% agar, in triplicate wells of a six-well plate. Colonies were allowed to develop for 10 days. Average colony numbers were determined by counting colonies greater than five cells in five random fields in triplicate wells at 10 magnification.

Analysis of ParvB mRNA and protein expression

Snap-frozen human breast cancer and adjacent normal mammary gland specimens were obtained from the Cooperative Human Tissue Network (CHTN) at the Hospital of the University of Pennsylvania in accordance with the Institutional Review Board (IRB) standards and guidelines. Total RNA was prepared from human breast tissues using the RNeasy Lipid Tissue Midi Kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions. Total RNA was prepared from breast cell lines using Trizol reagent (Invitrogen) according to the manufacturer's instructions. Single-stranded cDNA was synthesized from 5 g total RNA using the Superscript™ First Strand Synthesis System for RT–PCR (Invitrogen) and resuspended in DEPC-treated water (100 l total volume). PARVB PCR products were generated by touchdown PCR amplification of 5 l of cDNA template (in a total volume of 50 l) using MasterTaq polymerase (Eppendorf, Westbury, NY, USA) and a forward oligonucleotide primer in exon 10 (PARVB1Fn, 5'-TTTGGAGGTGACGGAACTGGA-3') paired with a reverse primer in exon 13 (PARVB1Rn, 5'-TGAAGGCCTGTGATCGCTAAC-3'). Internal control PCR reactions for human 2-microglobulin were performed using a forward primer in exon 1 (m2E1sl, 5'-AGATGTCTCGCTCCGTGGCCT-3') paired with a reverse primer in exon 2 (m2E2asl, 5'-CCCACTTAACTATCTTGGGCTGT-3'). PCR products were ligated into the pCRII vector using the TA cloning system (Invitrogen) followed by transformation of XL-1 Blue bacterial cells. White colonies were screened for the presence of PARVB cDNA insert by PCR amplification. Plasmid DNA was prepared using the QIAprep Spin Miniprep Kit (Qiagen) and the inserts sequenced.

Quantitative real-time PCR (Q-PCR) reactions were performed using SYBR green reagent (Applied Biosystems) in 25 l total volume and an ABI PRISM^® 7000 Sequence Detection System (Applied Biosystems) according to the manufacturer's instructions. PARVB primers were designed to lie in different exons. For PARVB, 1 l of cDNA template was used with forward primer hPARVBrealF (5'-CATCCGCCTTCCTGAGCAT-3') paired with reverse primer hPARVBrealR (5'-AGCAGGCCTTCCCGTTTC-3') or with forward primer hPARVBreal2F (5'-TGAATTTGGAGGTGACGGAACT-3') paired with reverse primer hPARVBreal2R (5'-CAGAAGGCCCATGAGCAGAA-3'). -actin was used as an internal control, where 0.025 l of cDNA template was used with forward primer (-actinRTF, 5'-CCT GGC ACC CAG CAC AAT-3') and reverse primer (-actinRTR, 5'-GCC GAT CCA CAC GGA GTA CT-3'). Optimal PCR conditions were determined by performing primer matrix reactions and generating standard curves for both PARVB and -actin. PCR reactions were performed in triplicate and the relative expression level of PARVB mRNA in normal and tumor tissue was calculated by normalizing to -actin mRNA expression levels using the comparative C_T (C_T) method, where C_T represents the cycle number at which the amplification reaches a threshold level chosen to lie in the exponential phase of all PCR reactions. Relative expression levels were calculated using the formula 2^-C_T, where C_T represents the difference between the average PARVB C_T value and the average -actin C_T value within a given tissue. C_T represents the difference between the C_T values for a matched normal and tumor pair. The normalized PARVB expression level for the normal mammary gland was set to 1. Data were analysed using ABI PRISM^® 7000 sequence detection system software (Applied Biosystems).

RT–PCR was performed to amplify PARVB products from 1 g of total RNA extracted by Trizol method from transfectant and parental cells. RT–PCR products were resolved by 1% agarose gel electrophoresis.

The sequences of the primers used are: ParvB1, 5'-ATGAAGAAGGACGAGTCGTTCCTG-3' (forward), 5'-TCACTCCACGTTCTTGTACTTGGTGAA-3' (reverse); ParvB2, 5'-ATGTCCTCCGCGCCGCGCTCG-3' (forward), 5'-TCACTCCACGTTCTTGTACTTGGTGAA-3' (reverse); ParvB3, 5'-ATGCACCATGTGTTTAAAGATCACCAA-3' (forward), 5'-TCACTCCACGTTCTTGTACTTGGTGAA-3' (reverse); ParvA, 5'-TCGAATTCAATGGCCACCTCCCCGCAGAA-3' (forward), 5'-TGCTCTAGATCACTCCACGTTACGGTACTT-3' (reverse); and -actin, 5'- GGGACCTGACTGACTACCT-3' (forward), 5'- CTAGAAGCATTTGCGGTGGA-3' (reverse).

Specimens of human breast tumors and the normal mammary gland were homogenized in 1.5 ml of RIPA lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate, 5 mM EDTA, 5 mM sodium orthovanadate and complete protease inhibitor cocktail (Roche)) per 400 mg of tissue. Lysates were incubated for 30 min (4°C) with shaking, centrifuged and the supernatants snap-frozen in liquid nitrogen. Protein concentration was determined using the BCA™ Protein Assay (Pierce).

Analysis of protein expression, signaling and ILK activity

Western blotting and ILK kinase assay was performed according to our previously published methods (Leung-Hagesteijn et al., 2001). Affinity purified rabbit polyclonal ParvB antibody was used in primary incubations at 3 g/ml. HRP-conjugated goat anti-mouse (Jackson ImmunoResearch) secondary antibody was incubated at 1 : 2000 for 1 h at room temperature. HRP-conjugated GFP antibody (Santa Cruz) and GAPDH (Ambion) were used to detect the GFP tag and as loading control, respectively. Bands were visualized with chemiluminescence substrate (ECL, Amersham). Relative band intensities were quantified using Kodak 1D2.02 software, and normalized against the GAPDH value, for each sample. Phospho-specific antibodies recognizing PKB S473 (New England Biolabs Inc.) and GSK-3 S9 (Cell Signaling Technology) were used for analyses of the phosphorylation status of PKB and GSK-3, as described previously (Leung-Hagesteijn et al., 2001). Relative intensity values were determined by normalizing to the most intense band, for each antibody.

Yeast two-hybrid binding assays

Yeast two-hybrid assays were conducted using a LexA-based system, and expression in EGY48 strain of Saccharomyces cerivisiae, as described previously (Hannigan et al., 1996). Full-length ILK^WT, ILKE359K or ILK^cat (amino-acid residues 1–142, lacking the catalytic domain) cDNAs were cloned into the pEG202 bait vector, for directed interaction in yeast cells cotransformed with prey plasmid, expressing the CH2 domain of ParvB (Figure 3c).

References

1. Ahmed N, Riley C, Oliva K, Stutt E, Rice GE and Quinn MA. (2003). J. Pathol., 201, 229–237. | Article | PubMed | ISI | ChemPort |
2. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W and Lipman DJ. (1997). Nucleic Acids Res., 25, 3389–3402. | Article | PubMed | ISI | ChemPort |
3. Andl CD, Mizushima T, Nakagawa H, Oyama K, Harada H, Chruma K, Herlyn M and Rustgi AK. (2003). J. Biol. Chem., 278, 1824–1830. | Article | PubMed | ISI | ChemPort |
4. Attwell S, Mills J, Troussard A, Wu C and Dedhar S. (2003). Mol. Cell. Biol., 14, 4813–4825. | Article |
5. Attwell S, Roskelley C and Dedhar S. (2000). Oncogene, 19, 3811–3815. | Article | PubMed | ISI | ChemPort |
6. Banuelos S, Saraste M and Carugo KD. (1998). Structure, 6, 1419–1431. | Article | PubMed | ISI | ChemPort |
7. Barker N and Clevers H. (2000). BioEssays, 22, 961–965. | Article | PubMed | ISI | ChemPort |
8. Cairns P, Okami K, Halachmi S, Halachmi N, Esteller M, Herman JG, Jen J, Isaacs WB, Bova GS and Sidransky D. (1997). Cancer Res., 57, 4997–5000. | PubMed | ISI | ChemPort |
9. Castells A, Gusella JF, Ramesh V and Rustgi AK. (2000). Cancer Res., 60, 2836–2839. | PubMed | ISI | ChemPort |
10. Castells A, Ino Y, Louis DN, Ramesh V, Gusella JF and Rustgi AK. (1999). Gastroenterology, 117, 831–837. | Article | PubMed | ISI | ChemPort |
11. Chung DH, Lee JI, Kook MC, Kim JR, Kim SH, Choi EY, Park SH and Song HG. (1998). Virchows Arch., 433, 113–117. | Article | PubMed | ISI | ChemPort |
12. Coffer PJ, Jin J and Woodgett JR. (1998). Biochem. J., 335, 1–13. | PubMed | ISI | ChemPort |
13. D'Amico M, Hulit J, Amanatullah DF, Zafonte BT, Albanese C, Bouzahzah B, Fu M, Augenlicht LH, Donehower LA, Takemaru K, Moon RT, Davis R, Lisanti MP, Shtutman M, Zhurinsky J, Ben-Ze'ev A, Troussard AA, Dedhar S and Pestell RG. (2000). J. Biol. Chem., 275, 32649–32657. | Article | PubMed | ISI | ChemPort |
14. Dai DL, Makretsov N, Campos EI, Huang C, Zhou Y, Huntsman D, Martinka M and Li G. (2003). Clin. Cancer Res., 9, 4409–4414. | PubMed | ISI | ChemPort |
15. Delcommenne M, Tan C, Gray V, Rue L, Woodgett J and Dedhar S. (1998). Proc. Natl. Acad. Sci. USA, 95, 11211–11216. | Article | PubMed | ChemPort |
16. Edwards LA, Shabbits JA, Bally M and Dedhar S. (2004). Cancer Treat. Res., 119, 59–75. | PubMed | ChemPort |
17. Fukuda T, Guo L, Shi X and Wu C. (2003). J. Cell Biol., 160, 1001–1008. | Article | PubMed | ISI | ChemPort |
18. Gimona M and Winder SJ. (1998). Curr. Biol., 8, R674–R675. | Article | PubMed |
19. Graff JR, Deddens JA, Konicek BW, Colligan BM, Hurst BM, Carter HW and Carter JH. (2001). Clin. Cancer Res., 7, 1987–1991. | PubMed | ISI | ChemPort |
20. Hannigan GE, Leung-Hagesteijn C, Fitz-Gibbon L, Coppolino MG, Radeva G, Filmus J, Bell JC and Dedhar S. (1996). Nature, 379, 91–96. | Article | PubMed | ISI | ChemPort |
21. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B and Kinzler KW. (1998). Science, 281, 1509–1512. | Article | PubMed | ISI | ChemPort |
22. Hresko MC, Williams BD and Waterston RH. (1994). J. Cell Biol., 124, 491–506. | Article | PubMed | ISI | ChemPort |
23. Ito R, Oue N, Zhu X, Yoshida K, Nakayama H, Yokozaki H and Yasui W. (2003). Virchows Arch., 442, 118–123. | PubMed | ISI | ChemPort |
24. Jonsson M, Borg A, Nilbert M and Andersson T. (2000). Eur. J. Cancer, 36, 242–248. | Article | PubMed | ISI | ChemPort |
25. Kim L and Kimmel AR. (2000). Curr. Opin. Genet. Dev., 10, 508–514. | Article | PubMed | ISI | ChemPort |
26. Korenbaum E, Olski TM and Noegel AA. (2001). Gene, 279, 69–79. | Article | PubMed | ISI | ChemPort |
27. Kumar AS, Naruszewicz I, Wang P, Leung-Hagesteijn C and Hannigan GE. (2004). Oncogene, 23, 3454–3461. | Article | PubMed | ISI | ChemPort |
28. Leslie NR and Downes CP. (2002). Cell Signal., 14, 285–295. | Article | PubMed | ISI | ChemPort |
29. Leung-Hagesteijn C, Mahendra A, Naruszewicz I and Hannigan GE. (2001). EMBO J., 20, 2160–2170. | Article | PubMed | ISI | ChemPort |
30. Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, Puc J, Miliaresis C, Rodgers L, McCombie R, Bigner SH, Giovanella BC, Ittmann M, Tycko B, Hibshoosh H, Wigler MH and Parsons R. (1997). Science, 275, 1943–1947. | Article | PubMed | ISI | ChemPort |
31. Lin SY, Xia W, Wang JC, Kwong KY, Spohn B, Wen Y, Pestell RG and Hung MC. (2000). Proc. Natl. Acad. Sci. USA, 97, 4262–4266. | Article | PubMed | ChemPort |
32. Lin X, Qadota H, Moerman DG and Williams BD. (2003). Curr. Biol., 13, 922–932. | Article | PubMed | ISI | ChemPort |
33. Mackinnon AC, Qadota H, Norman KR, Moerman DG and Williams BD. (2002). Curr. Biol., 12, 787–797. | Article | PubMed | ISI | ChemPort |
34. Maehama T and Dixon JE. (1999). Trends Cell. Biol., 9, 125–128. | Article | PubMed | ISI | ChemPort |
35. Marotta A, Parhar K, Owen D, Dedhar S and Salh B. (2003). Br. J. Cancer, 88, 1755–1762. | Article | PubMed | ISI | ChemPort |
36. Marotta A, Tan C, Gray V, Malik S, Gallinger S, Sanghera J, Dupuis B, Owen D, Dedhar S and Salh B. (2001). Oncogene, 20, 6250–6257. | Article | PubMed | ISI | ChemPort |
37. Mills GB, Lu Y, Fang X, Wang H, Eder A, Mao M, Swaby R, Cheng KW, Stokoe D, Siminovitch K, Jaffe R and Gray J. (2001). Semin. Oncol., 28, 125–141. | Article | PubMed | ISI | ChemPort |
38. Myers MP, Pass I, Batty IH, Van der Kaay J, Stolarov JP, Hemmings BA, Wigler MH, Downes CP and Tonks NK. (1998). Proc. Natl. Acad. Sci. USA, 95, 13513–13518. | Article | PubMed | ChemPort |
39. Nikolopoulos SN and Turner CE. (2001). J. Biol. Chem., 276, 23499–23505. | Article | PubMed | ISI | ChemPort |
40. Nikolopoulos SN and Turner CE. (2002). J. Biol. Chem., 277, 1568–1575. | Article | PubMed | ISI | ChemPort |
41. Novak A and Dedhar S. (1999). Cell. Mol. Life Sci., 56, 523–537. | Article | PubMed | ISI | ChemPort |
42. Olski TM, Noegel AA and Korenbaum E. (2001). J. Cell Sci., 114, 525–538. | PubMed | ISI | ChemPort |
43. Persad S and Dedhar S. (2003). Cancer Metast. Rev., 22, 375–384. | Article | ISI | ChemPort |
44. Persad S, Troussard AA, McPhee TR, Mulholland DJ and Dedhar S. (2001). J. Cell Biol., 153, 1161–1174. | Article | PubMed | ISI | ChemPort |
45. Price JT, Tiganis T, Agarwal A, Djakiew D and Thompson EW. (1999). Cancer Res., 59, 5475–5478. | PubMed | ISI | ChemPort |
46. Radeva G, Petrocelli T, Behrend E, Leung-Hagesteijn C, Filmus J, Slingerland J and Dedhar S. (1997). J. Biol. Chem., 272, 13937–13944. | Article | PubMed | ISI | ChemPort |
47. Rosenberger G, Jantke I, Gal A and Kutsche K. (2003). Hum. Mol. Genet., 12, 155–167. | Article | PubMed | ISI | ChemPort |
48. Shtutman M, Zhurinsky J, Simcha I, Albanese C, D'Amico M, Pestell R and Ben-Ze'ev A. (1999). Proc. Natl. Acad. Sci. USA, 96, 5522–5527. | Article | PubMed | ChemPort |
49. Tan C, Costello P, Sanghera J, Dominguez D, Baulida J, de Herreros AG and Dedhar S. (2001). Oncogene, 20, 133–140. | Article | PubMed | ISI | ChemPort |
50. Troussard AA, Costello P, Yoganathan TN, Kumagai S, Roskelley CD and Dedhar S. (2000). Oncogene, 19, 5444–5452. | Article | PubMed | ISI | ChemPort |
51. Tu Y, Huang Y, Zhang Y, Hua Y and Wu C. (2001). J. Cell Biol., 153, 585–598. | Article | PubMed | ISI | ChemPort |
52. White DE, Cardiff RD, Dedhar S and Muller WJ. (2001). Oncogene, 20, 7064–7072. | Article | PubMed | ISI | ChemPort |
53. Williams BD and Waterston RH. (1994). J. Cell Biol., 124, 475–490. | Article | PubMed | ISI | ChemPort |
54. Yamaji S, Suzuki A, Sugiyama Y, Koide Y, Yoshida M, Kanamori H, Mohri H, Ohno S and Ishigatsubo Y. (2001). J. Cell Biol., 153, 1251–1264. | Article | PubMed | ISI | ChemPort |
55. Youssoufian H. (1998). Biotechniques, 24, 198–200 202. | PubMed | ISI | ChemPort |
56. Yu X, Miyamoto S and Mekada E. (2000). J. Cell Sci., 113 (Part 12), 2139–2147. | PubMed | ISI | ChemPort |
57. Zhang Y, Guo L, Chen K and Wu C. (2002). J. Biol. Chem., 277, 318–326. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank HSC colleagues, Dr S Egan for critical reading of the manuscript, Dr C Pace-Asciak for the gift of MDA-MB-231 and MCF7 cells, Drs Rae Yeung and Trang Duong for providing -actin primers and Sherry Zhao (HSC FACS facility) for her expert cell cycle analyses. CNJ was supported by grants from the American Association for Cancer Research and US Department of Defense. AKR received an R01 grant from the National Institutes of Health and grants from the Hansen Foundation and NCCRA/EIF. GH was supported by grants from the National Cancer Institute of Canada (with funds from the Terry Fox Run) and the Canadian Institutes of Health Research. GH was a CIHR scholar.