Suppression of malignant growth of human breast cancer cells by ectopic expression of integrin-linked kinase
Version of Record online: 20 MAY 2004
Copyright © 2004 Wiley-Liss, Inc.
International Journal of Cancer
Volume 111, Issue 6, pages 881–891, 10 October 2004
How to Cite
Chen, P., Shen, W.-Z. and Karnik, P. (2004), Suppression of malignant growth of human breast cancer cells by ectopic expression of integrin-linked kinase. Int. J. Cancer, 111: 881–891. doi: 10.1002/ijc.20340
- Issue online: 3 AUG 2004
- Version of Record online: 20 MAY 2004
- Manuscript Accepted: 2 FEB 2004
- Manuscript Revised: 16 JAN 2004
- Manuscript Received: 4 SEP 2003
- U.S. Army Medical Research and Materiel Command. Grant Numbers: DAMD17-96-1-6052, DAMD17-00-1-0205
- gene transfer;
- breast cancer;
- cell-cycle arrest;
- nude mouse assay
Allelic loss at the short arm of chromosome 11 is one of the most common and potent events in the progression and metastasis of breast cancer. Here, we present evidence that the Integrin-Linked Kinase (ILK) gene maps to the commonly deleted chromosome 11p15.5 and suppresses malignant growth of human breast cancer cells both in vitro and in vivo. ILK is expressed in normal breast tissue but is downregulated in metastatic breast cancer cell lines and in advanced breast cancers. Transfection of wild-type ILK into the MDA-MB-435 mammary carcinoma cells potently suppressed their growth and invasiveness in vitro and reduced the cells' ability to induce tumors and metastasize in athymic nude mice. Conversely, expression of the ankyrin repeat or catalytic domain mutants of ILK failed to suppress the growth of these cells. Growth suppression by ILK is not due to apoptosis but is mediated by its ability to block cell-cycle progression in the G1 phase and by modulating the levels of integrins. These findings directly demonstrate that ILK deficiency facilitates neoplastic growth and invasion and suggest a novel role for the ILK gene in the suppression of tumor metastasis. © 2004 Wiley-Liss, Inc.
Genetic alterations that occur in breast cancer are believed to be of importance for initiation as well as progression of the disease. These genetic alterations lead to the loss or activation of a number of critical genes, such as those involved in cell proliferation, differentiation, apoptosis and genetic stability. The genetic abnormalities most frequently observed in breast tumors are amplification of proto-oncogenes (MYC, ERBB2 and CCND1), mutations of TP53 and loss of heterozygosity (LOH) on chromosomes 3p, 6q, 7q, 8p, 9p, 11, 13q, 17, 18q and 22q.1, 2 Metastatic phenotypes have been linked to such genes as NME1 (17q), CDH1 (16q), BRMS1 (11q) and KISS1 (1q).1, 3, 4, 5 LOH analyses have defined regions of deletion associated with metastasis on chromosomes 3p21, 15q14, 16q22 and 11p15.2, 6
Frequent genetic alterations on chromosome 11p15 suggest a crucial role for this region in breast6, 7 and other adult8, 9, 10, 11, 12 and childhood cancers.13, 14, 15, 16, 17 More recently, we have mapped 2 distinct regions on chromosome 11p15.5 that are subject to LOH during breast tumor progression and metastasis.6 LOH at region 1 correlated with tumors that contain ductal carcinoma in situ, suggesting that the loss of a critical gene in this region may be responsible for early events in malignancy. LOH at region 2 correlated with a more aggressive tumor and an ominous outlook for the patient, such as aneuploidy, high S-phase fraction and the presence of metastasis in regional lymph nodes. Although considerable advances have been made in the fine-mapping of chromosome 11p15.5, the tumor/metastasis suppressor gene(s) encoded by this region has evaded identification.
Integrin-linked kinase (ILK) is an intriguing serine/threonine kinase that has been implicated in integrin-, growth-factor- and Wnt-signaling pathways.18 It binds to the cytoplasmic domains of β1 and β3 integrins and mediates the downstream signaling events in integrin function.19 Interactions between integrins and their ligands are involved in the regulation of many cellular functions, including embryonic development, cell proliferation, tumor growth and the ability to metastasize.20 In Drosophila, the absence of ILK function causes defects similar to loss of integrin adhesion, and ILK mutations cause embryonic lethality and defects in muscle attachment.21 Although ILK maps to the commonly deleted chromosome 11p, the potential of this gene in tumor/metastasis suppression has not been evaluated. We have therefore analyzed the effect of ILK expression on the in vitro and in vivo tumor growth and invasion of human mammary carcinoma cells.
MATERIAL AND METHODS
Mapping of ILK to LOH region 2
Two YAC clones (847a12 and 696H10), a BAC clone (BAC 1760) and a PAC clone (PAC1331) overlapped the LOH region 2 and were used to construct a genomic contig across this the approx. 336 kb LOH region 2 (data not shown). The contig was assembled by mapping of STSs and polymorphic markers. ILK was mapped to this region using the primers (5′-TGGAACCCTGAACAAACACT-3′ and 5′-AGTCCCTGCTCTTCCTTGTA-3′).
Cell lines and cell cultures
MDA-MB-435 is an estrogen receptor- and progesterone receptor-negative, metastatic, ductal human breast carcinoma cell line derived from a 31-year-old female.22 It has a heterogeneous chromosome complement of from 50–58 chromosomes, including several derivative marker chromosomes. The cell line readily forms tumors when injected into mammary fat pads (mfp) of nude mice, and macroscopic metastases to lungs and regional lymph nodes can be identified 15–18 weeks postinoculation. This pattern closely parallels clinical observations. Therefore, this cell line was chosen for our studies. MDA-MB-435 cells were the generous gift of Dr. Janet Price (University of Texas M.D. Anderson Cancer Center, Houston, TX). Human breast cancer cells MCF-7, T47D, MDA-MB-231, ZR75-1, MDA-MB-134, MDA-MB-468 were obtained from the American Type Culture Collection (Rockville, MD). All human breast cancer cell lines were maintained in DMEM-F12 medium supplemented with 10% fetal calf serum and no antibiotics. The neomycin-resistant ILK transfectants were maintained in DMEM-F12 containing 300 μg/ml G418 (Life Technologies, Gaithersburg, MD). All cell lines were free of Mycoplasma sp. contamination as determined by a PCR-based test (Pan Vera, Madison, WI).
Northern blotting, immunoblotting and immunohistochemistry
Primary tumor and adjacent normal breast tissue samples were obtained from breast cancer patients undergoing mastectomy at the Cleveland Clinic Foundation (CCF). All tumor samples described in this article were determined to contain LOH at 11p15.5 region. Samples of these tumors and corresponding noninvolved tissue from each patient were collected at the time of surgery, snap-frozen and transferred to −80°C. Clinical and histopathologic features of the tumors were determined by the Pathology Department at CCF. An initial cryostat section was stained with H&E to determine the proportion of contaminating normal tissue in the tumor.
Total RNA from cell cultures and tumor tissues was isolated using an RNeasy Kit® (Qiagen, Chatsworth, CA). Total RNA (20 μg) was size-fractionated in a 1% agarose gel containing 2.2 M formaldehyde. After transfer and UV-cross-linking, the nylon membranes were probed with a full-length 32P-labeled ILK cDNA. The band intensities on autoradiographs were quantitated by densitometry within linear range of signal and normalized to ribosomal 18S RNA levels. All the immunohistochemical determinations were performed on representative samples snap-frozen in liquid nitrogen and stored at −80°C until sectioning. Cryostat sections (4–6 μm thick) of tumor blocks were deparaffinized with xylene, rehydrated and microwaved for 10 min in 10 mM citrate buffer (pH 6.0). MDA-MB-435 cells were grown on sterile microscope coverslips (Fisher Diagnostics, Middletown, VA). The coverslips with MDA-MB-435 cells and slides containing tissue sections were fixed in 4% paraformaldehyde solution for 12 min, washed in PBS, preincubated with 10% Triton X-100 in PBS and incubated at room temperature with an affinity-purified rabbit polyclonal Anti-ILK IgG preparation (Upstate Biotechnology, Lake Placid, NY) for 1 hr. After being washed with PBS, bound antibodies were visualized using rhodamine-conjugated goat anti-rabbit IgG secondary antibody. Sections were visualized by fluorescence microscopy. The optimal concentration of primary and secondary antibody was determined by titration and ranged from 1:50–1:200. For negative controls, in all instances, we used nonspecific IgG as the primary antibody.
For ILK immunoblotting, 5 μg of whole cell lysates (1% N-40, 0.5% deoxycholate) were separated by 10% SDS-PAGE, transferred to nylon membrane (PVDF, BioRad, Hercules, CA) and probed with an affinity-purified rabbit polyclonal Anti-ILK antibody.
Construction and transfection of wild-type and mutant forms of ILK
A partial ILK cDNA clone (EST bb36h09.y1, Invitrogen Corp., Carlsbad, CA) was used to obtain a full-length clone from a human placental cDNA library and its identity confirmed by sequence analysis and database comparison. The ILK catalytic domain mutant E359K was created as described earlier.18 Briefly, mutations were introduced into wild-type ILK cDNA with the Promega Altered Sites II (Promega, Madison, WI). A mutant oligomer (with the altered nucleotide underlined) was used to change glutamic acid at position 359 to lysine (E359K, 5′-CTGCAGAGCTTTGGGGGCTACCCAGGCAGGTG-3′). To create the ankyrin repeat deletion mutant Δ ANK (residues 165–451), the 858 bp fragment of ILK cDNA lacking the NH2-terminal ankyrin repeat domain was PCR amplified from the wild-type ILK cDNA. The mutant clones were confirmed by dideoxy sequencing. The full-length and mutant ILK cDNA inserts were cloned in the eukaryotic expression vector pIRES2-EGFP (Clontech, Palo Alto, CA). MDA-MB-435 cells were transfected with plasmid vector alone or containing the cDNA for ILK using LipofectAMINE reagent (Life Technologies). Briefly, cells were grown to approx. 60% confluence in 6 cm tissue culture dishes, rinsed twice with serum-free medium, overlaid with a mixture of 5 μg of DNA and 10 μl of LipofectAMINE reagent diluted in serum-free medium and incubated at 37°C in 5% CO2/95% air for 18 hr. The transfection medium was then replaced with fresh medium, and 36 hr later, the cells were harvested, diluted in growth medium containing 500 μg/ml G418 and split 1:30 for the selection and establishment of clonal cell lines.
Cell growth rate, cell-cycle analysis and apoptosis assay
For growth rate analysis, untransfected, vector transfected and ILK transfected cells were plated at a density of 2 × 105 in DMEM-F12 plus 5% FBS medium in 24-well cluster plates. Viable and dead cells were assessed by counting with Trypan blue exclusion at Day 2, Day 4, Day 6 and Day 8. Stable ILK transfectants or control cells were trypsinized, and the cell numbers per dish were measured by Coulter counting and also by using a hemocytometer.
Flow cytometry was used to determine the cell-cycle distributions as described.23ILK transfected and parental MDA-MB-435 cells were washed with PBS twice, then fixed in 70% ethanol for 30 min at 4°C. The cells were treated with 1 U DNase-free RNase in 1 ml of PBS for 30 min at 37°C and finally resuspended in 0.05 mg ml propidium iodide (made as a 10× stock in PBS). Cells were analyzed by flow cytometry using a FACScan model (Becton Dickenson, San Jose, CA). Ten thousand forward scatter-gated events were collected for each sample. Fluorescence measurements were accumulated to form a distribution curve of DNA content. Fluorescence events due to debris were subtracted before analysis.
The degree of apoptosis in vector control and ILK transfected MDA-MB-435 cells was quantified by using a commercially available kit with fluorescein-labeled annexin V24 (R & D Systems, Minneapolis, MN), according to the manufacturer's instructions. Samples were analyzed on a Becton Dickinson FACScan flow cytometer (a generous gift from Keck Foundation).
Cell migration assay
Cell migration assays were performed as described earlier25 using modified Boyden chambers (tissue culture-treated, 6.5 mm diameter, 10 μm thickness, 8 μm pores, Transwell®; Costar, Cambridge, MA) containing polycarbonate membranes coated on the underside of the membrane with 10 μg/ml vitronectin in PBS. The ILK transfected and control cells were harvested with 0.05% Trypsin-EDTA, washed twice with quenching medium (serum-free medium containing 5% BSA) and then resuspended in quenching medium (106 cells/ml). About 50,000–100,000 cells were then added to the top of each migration chamber and allowed to migrate to the underside of the top chamber for 6 hr at 37°C in a CO2 incubator. The nonmigratory cells on the upper membrane were removed with a cotton swab, and the migratory cells attached to the bottom surface of the membrane were washed with PBS, extracted with 300 μl extraction buffer and absorbance determined at 560 nm. All values have had background subtracted, which represents cell migration on membranes coated with BSA (1%). Each determination represents the average of 3 individual wells, and error bars represent the standard deviation (SD).
Analysis of cell surface integrin profiles
Fluorescence-activated cell analysis26 was used to identify the integrin profiles on MDA-MB-435 cells in response to ILK expression. Monolayer cultures (60–80% confluence) ILK transfected and control cells were trypsinized and washed in culture medium. Briefly, harvested cells were divided into equal aliquots of 2.5 × 105 cells/ml in serum-free medium plus 1% BSA. After 2 washes in this medium, the cells were resuspended in 1:50 dilution of anti- αvβ3 or α5β1 specific antibody (Chemicon, Temecula, CA) in serum-free medium plus 1% BSA for 1 hr on ice. After 2 washes in serum-free medium plus 1% BSA, the cells were incubated in 1:100 dilution of F(ab′)2 secondary anti-goat antibody conjugated with FITC (ICN Biomedicals, Irvine, CA) in this same medium for 1 hr on ice. The cells were washed twice in PBS/0.1% BSA and resuspended in the same solution. These samples were then analyzed using a Becton Dickinson FACScan and the data analyzed using the CellQuest Software.
Tumorigenicity and metastasis assays
Cells (106) were injected into the subaxillary mammary fat pads of 4–6-week-old female athymic nude mice Ncr nu/nu (10–12 mice/group; Taconic Labs, Germantown, NY) as described.27 Mice were maintained under the guidelines of NIH and the Cleveland Clinic Foundation. All protocols were approved and monitored by the Institutional Animal Care and Use Committee. Food and water were provided ad libitum. Tumors were monitored weekly after inoculation. When the mean tumor diameter reached 1.0–1.3 cm, primary tumors were surgically removed under Ketaset-Rompun anesthetic. Mice were then maintained for an additional 4 weeks to allow further growth of lung metastases. After euthanasia, all organs were checked for metastases.
Localization of ILK to the LOH region on chromosome 11p15.5
The LOH region 2 extends between the markers D11S1760 and D11S1331 on chromosomal band 11p15.5 (Fig. 1).6 We constructed a 500 kb genomic contig (data not shown) that includes the critical region between D11S1760 and D11S1331. Using a PCR-based screening method, we initially isolated PAC and BAC clones that contained D11S1760 and D11S1331 markers. The order of the genomic clones in the contig was confirmed by mapping of STSs, ESTs, unigene clusters and known genes that were previously mapped to chromosome 11. Eleven novel transcripts and 7 previously reported genes were PCR-mapped to the critical region between D11S1760 and D11S1331. Three of the known genes, Tata box-binding protein-associated protein (TAF II 30),28 Lysosomal pepstatin insensitive protease (CLN2)29 and Integrin-linked kinase (ILK)30 were previously mapped only at the level of cytogenetic resolution. However, with the current mapping data, we have been able to determine the precise genomic locations of these 3 genes (Fig. 1). ILK and TAF II 30 are known to be in an overlapping transcription unit in a head-to-tail arrangement. The map location and its role in multiple signaling pathways make ILK an attractive candidate tumor/metastasis-suppressor gene.
Loss of ILK expression in human breast carcinomas
To determine whether ILK has a role in breast cancer progression, mRNA expression in normal and tumor breast epithelial cells was compared by Northern blot hybridization and densitometry quantitation normalized to ribosomal 18S RNA (Fig. 2). A single 1.8 kb ILK mRNA is highly expressed in all samples of normal breast epithelial cells. A representative example is shown in Figure 2(N1). In sharp contrast, there is either complete loss or significant downregulation of ILK mRNA expression in invasive breast tumors (Fig. 2a). Comparison of ILK mRNA expression in a panel of well-characterized breast cancer cell lines and in the nonmalignant breast epithelial cell line MCF-10A is shown in Figure 2b. ILK mRNA expression in MCF-10A is comparable to the expression in normal breast tissue (N7, N8) (Fig. 2b). However, there is a 2-5 fold downregulation of ILK mRNA expression in the breast cancer cell lines MCF-7, T47D, ZR75.1, MDA-468, MDA-134, MDA-231 and MDA-435 (Fig. 2b). Immunoblot analysis (Fig. 2c) indicates that the differences at the RNA levels are translated to differences at the protein level. The amount of immunoprecipitable ILK from NP-40 lysates was barely visible in 5 of 7 tumor samples compared to normal breast tissue.
To further confirm these observations, ILK protein expression was also examined using indirect immunofluorescence microscopy in frozen samples of 20 normal and corresponding pathologic human breast tissue samples. Figure 3 shows 4 representative examples. Immunohistochemical staining of normal breast tissue with ILK-specific primary antibody and rhodamine-labeled secondary antibody shows specific staining of the mammary epithelial cells surrounding the lumen in normal breast tissue from breast cancer patients. ILK expression is particularly intense in epithelial cells both within large ducts and within terminal duct lobular units but not in the stromal compartment. Incubation with purified nonspecific rabbit immunoglobulin IgG did not result in any positive staining of the normal epithelium of the breast (control). The normal breast tissue from 4 representative patients were positive (3N, 12N, 6N and 10N), whereas ILK expression was nearly completely lost in the 4 corresponding infiltrating ductal carcinomas (3T, 12T, 6T, 10T) (Fig. 3). These data show that ILK production by breast tumor cells correlates inversely with tumorigenicity and metastatic potential.
We observed a striking clinical correlation between loss of ILK expression in breast tumors and lymphatic invasion (data not shown). These tumors revealed a significantly higher incidence of metastasis to regional lymph nodes (p = 0.01) than tumors that expressed ILK. However, loss of ILK expression did not correlate significantly with ER/PR or p53 status or with any other critical parameters.
All breast tumor samples described in Figures 2 and 3 have previously been identified to contain LOH at the 11p15.5.6 Allelic loss results in the reduction of gene dosage and thus may result in decreased expression. However, as seen in Figure 2, all tumors have LOH for 11p15.5 and yet only some tumors show complete loss of ILK expression. Therefore, intragenic mutations or epigenetic mechanisms might contribute to the biallelic silencing of the ILK gene in breast tumors. These mechanisms are currently under investigation.
ILK suppresses cell growth in human breast carcinoma cells
The inverse correlation between ILK expression and tumorigenicity suggested the hypothesis that elaboration of ILK by tumor cells into their environment may exert an inhibitory effect. To test this hypothesis, we transfected the human breast carcinoma cell line MDA-MB-435 with the ILK cDNA. This cell line synthesizes very low levels of ILK compared to normal mammary epithelial cells (Fig. 2b) and can be injected into the mammary fat pad of nude mice to provide an orthotopic model system for human breast cancer tumorigenicity and metastasis. The MDA-MB-435 cells were transfected with a mammalian expression vector pIRES-EGFP containing full-length ILK cDNA under control of the CMV promoter. A total of 4 stable clones expressing different levels of ILK have been established. Comparison of mRNA expression by Northern blot analyses revealed that the clones TR4 and TR5 expressed slightly higher levels of ILK mRNA compared to the clones TR2 and TR3 (Fig. 4b). Based on Northern blot analysis, ILK expression in clone TR5 is comparable to the expression in the nonmalignant breast epithelial cell line MCF-10A and to the expression in normal mammary epithelial cells (Fig. 2b), suggesting that ILK expression is restored to normal levels upon transfection. The expression of ILK in empty vector controls (data not shown) is comparable to untransfected MDA-MB-435 cells (UT). ILK protein levels in transfected (TR3-TR5) and untransfected cells (UT) were determined by indirect immunofluorescence and Western blotting. The ILK protein levels in these cells are restored to normal breast tissue levels (Fig. 4c), and the expressed ILK protein is localized in the cytoplasm (Fig. 4a, panel c). Most strikingly, corresponding to the low levels of ILK mRNA (Fig. 2b), the highly metastatic MDA-MB-435 cell line exhibited very little detectable ILK protein (Fig. 4a, panel b, and 4c).
To determine whether ILK expression had any effect on the growth properties of the MDA-MB-435 cells, we determined the growth kinetics of the clones TR3 and TR5. ILK expression causes the MDA-MB-435 cells to grow to a low saturation density (Fig. 5a), and there is substantial growth suppression of the TR5 clone compared to untransfected MDA-MB-435 cells. The growth suppression of the transfectants was ILK concentration dependent with TR5 (high-expressing clone) growing to a lower saturation density than TR3 (low-expressing clone). Furthermore, the growth rate of TR5 was decreased by approx. 40% with a cell doubling time of 96 hr compared to the growth rate of cells transfected with vector alone or untransfected cells, which had a doubling time of 48 hr. However, there were no remarkable changes in cell morphology.
The ability of ILK to suppress growth could be either due to a nonspecific lethal effect of protein overproduction or due to a more specific effect on cell proliferation. Alternatively, antisense interference of ILK mRNA with the TAF II 30 may also be responsible, since the overlapping ILK and TAF II 30 transcription units are in a head-to-tail arrangement. To further establish a link between ILK protein function and growth suppression, we tested the growth kinetics of 2 ILK variants. ILK contains 4 ankyrin repeats at the NH2-terminus18 that participate in protein-protein interactions important for integrin-, growth-factor- and Wnt-mediated signaling. First, a deletion mutant, Δ ANK lacking this domain was constructed. In addition, the residue E359 has been shown to be essential for ILK function.18 We therefore constructed an ILK point mutant (E359K) in which the highly conserved Glu359 within the ILK catalytic domain was substituted with lysine. Whether the residue E359 is responsible for kinase activity is unclear;31 however, it is clear that E359K mutation interferes with ILK protein function. In transfected clones, the level of expression of mutant ILK proteins was comparable to that of wild-type ILK. The growth rates of the stably transfected ILK mutant clones Δ ANK and E359K compared to the ILK transfectant TR5 are shown in Figure 5b. As discussed above, expression of the wild-type ILK strongly inhibited growth of the MDA-MB-435 cells. In contrast, both the Δ ANK and E359K mutants lost their capacity to suppress the growth of the MDA-MB-435 cells (Fig. 5b), arguing against a nonspecific effect of protein overproduction as well as antisense interference with TAF II 30 transcription unit.
Expression of ILK in MDA-MB-435 cells leads to a G1 cell-cycle arrest
The observed growth suppression by ILK could be caused by either increased apoptosis or inhibition of cell proliferation. To investigate the mechanisms underlying the growth suppression by ILK expression, we studied apoptosis by fluorescence-activated cell sorting (FACS) analysis of Annexin-V stained ILK and vector transfectants. There was no increase in the rate of apoptosis in ILK-expressing cells compared to vector transfectants (data not shown). Therefore, programmed cell death does not seem to account for the growth suppression of ILK transfected cells.
To test for cell-cycle regulation by ILK, propidium iodide-stained MDA-MB-435 clones were analyzed by flow cytometry. Expression of ILK increased the number of cells in G0/G1 from 64% to 85% (Fig. 5c, VT and TR5-ILK) and decreased inversely the number of cells in S and G2/M phase from 26% and 10% to 9% and 5% (Fig. 5c, VT and TR5-ILK). In contrast, the cell-cycle profiles of the 2 ILK variants Δ ANK and E359K were very similar to the parental MDA-MB-435 cells. These results indicate that ILK growth suppression results from G1 cell-cycle arrest. The accumulation of cells in the G0/G1 phase of the cell cycle suggests arrest predominantly at the G1/S boundary. ILK expression does not induce cell death or apoptosis but induces a very pronounced growth arrest with 85% of the cells in G0/G1, a property that is the hallmark of growth/tumor suppressors.
ILK suppresses the invasive phenotype of human breast carcinoma cells
Cell migration on vitronectin in vitro has been linked to the metastatic capacity of tumor cells in vivo.32, 33 To examine the effects of ILK expression on breast cancer cell invasion, the ability of vector and ILK transfected MDA-MB-435 cells to degrade and invade vitronectin-coated polycarbonate membrane was investigated. As shown in Fig. 6a, a significant reduction in invasive potential was noted in the ILK-expressing clone TR5 compared to vector transfected MDA-MB-435 cells (VT) (Fig. 6a). Cell invasion through membranes coated with vitronectin is decreased by 60% in MDA-MB-435 cells expressing ILK compared to vector transfected MDA-MB-435 cells. In contrast, the 2 ILK variants Δ ANK and E359K have no significant effect on cell invasion under identical conditions (Fig. 6a).
In fact, there is a slight increase in invasive potential of the variant clones (Δ ANK and E359K), suggesting a dominant-negative effect, perhaps due to inhibition of endogenous ILK in the MDA-MB-435 cells. These results indicate that ILK expression abates extracellular matrix invasion of tumor cells in vitro, one of the hallmarks of tumorigenicity and transformed cell growth.
Cell adhesion, migration and invasion are controlled by the levels of integrins and by the amount of fibronectin matrix around the cell.20 Because the α5β1 and αvβ3 integrins have been implicated in the regulation of angiogenesis, tumor cell migration, invasion and metastasis, we speculated that ILK might regulate cell migration via alteration of the cellular composition of integrins. Using specific antibodies against these integrins in flow cytometry analysis, we compared integrin expression patterns in relation to the ILK expression status. The results are shown in Figure 6b. The ILK transfected cells demonstrated a 22% increase in levels of the growth-suppressing integrin α5β1 and a 31% decrease in levels of the growth-promoting integrin αvβ3 compared to the control cells. The changes in levels of αvβ3 and α5β1 expression in ILK transfected cells, although relatively moderate in comparison to control cells, nonetheless are highly significant. Collectively, these observations suggest that ILK reduces the invasive potential of MDA-MB-435 cells by altering their integrin profiles, which changes their ability to perceive and interact with their extracellular environment.
ILK suppresses tumor formation and metastasis in nude mice
The most stringent experimental test of neoplastic behavior is the ability of injected cells to form tumors in nude mice. Yet not all of the cellular growth properties commonly associated with the cellular state in vitro are required for neoplastic growth in vivo and vice versa. Therefore, loss of tumorigenicity under expression of ILK in vivo would be a critical test to substantiate the growth suppressor function of ILK. The mammary carcinoma cell line MDA-MB-435 forms tumors at the site of orthotopic injection, metastasizes in nude mice and closely resembles the course of human breast cancer.34 To investigate whether ILK expression affected tumor formation in nude mice, 2 different ILK transfectant clones (TR5-ILK and TR3-ILK) and 2 vector controls were inoculated into the subaxillary mammary fat pads of 4–6-week-old athymic nude mice. Tumors were measured weekly thereafter to assess the growth rate. All MDA-MB-435 vector transfectants were already palpable 7 days after injection. Subsequently, the tumors of vector transfectants grew steadily, attaining mean volumes of 3.0 cm3 (mean ± S.D.) at 15 weeks (Fig. 7a,b). In contrast, only 2 of 12 mice injected with ILK transfectants developed tumors. The tumor growth of ILK transfectants was significantly slower than that of control transfectants (p < 0.005, Fisher variance analysis). At sacrifice (15 weeks), the ILK tumors reached a mean volume of only 0.45 cm3 (mean ± S.D.), which was significantly smaller than control tumors (p < 0.001, Student's t-test). Vector transfected MDA-MB-435 cells developed an average of 12–24 lung metastases per mouse (Fig. 7c). Additional tumor masses were present in central venous blood vessels, the diaphragm and lymph nodes of vector transfectants (data not shown). In contrast, with the ILK transfectants, only 1 of the 2 animals that developed tumors exhibited a single metastatic colony in the lung. The presence of additional microscopic metastases in random lung sections was not observed by H&E staining (data not shown). These results clearly demonstrate that the expression of ILK in human MDA-MB-435 breast carcinoma cells significantly suppresses tumorigenicity and metastatic ability in athymic nude mice.
Growth-inhibitory functions of ILK
Our study reveals that expression of ILK potently suppresses in vitro growth and invasiveness and in vivo tumorigenicity of the human mammary carcinoma cells. The MDA-MB-435 cells are a model for deficient ILK protein expression, and transfection of the ILK gene is designed to restore this deficiency. The growth-suppression activity requires a functional ILK protein, since expression of wild-type ILK, but not the ankyrin repeat or the catalytic domain mutants, resulted in growth suppression of MDA-MB-435 cells. These results suggest a possible role for ILK in the suppression of tumor growth and metastasis and directly implicate its loss in processes regulating the malignant phenotype in human breast cancer. ILK seems to play a dual role in the MDA-MB-435 model system. First, it regulates cell-cycle progression at the G1/S boundary, and second, it modulates the levels of integrins, transmembrane receptors that have been shown to regulate growth, differentiation and invasiveness of cells. During this process, the neoplastic cells cease to proliferate and lose their ability to migrate through vitronectin membranes (Fig. 6) and to induce tumor growth and metastasis in nude mice (Fig. 7). Our results are consistent with earlier micro-cell-mediated chromosome transfer experiments showing that introduction of human chromosome 11 into MDA-MB-435 cells suppressed metastasis in nude mice.35
Our observations on the role of ILK in the highly invasive mammary carcinoma cell line differ from earlier findings of ILK overexpression in normal epithelial cells. Overexpression of ILK in normal epithelial cells (Scp2 mouse mammary epithelial cells and IEC-18 rat intestinal epithelial cells) results in anchorage-independent growth,36 cell-cycle progression37 and tumorigenicity in nude mice.38 The apparent contradiction between these 2 models may be due to the cellular context in which ILK was expressed. Normal epithelial cells express optimal levels of ILK, whereas the MDA-MB-435 cells have sub-threshold ILK expression. Thus, the effects we report are due to restoration of the cellular ILK levels, whereas the results of ILK transfection in normal cells likely represent the effects of ILK overexpression. These differences could account for distinct phenotypic changes in the 2 experimental models. For instance, normal epithelial cells have a different integrin composition compared to breast cancer cell lines,39, 40 and any changes in these cells that perturbs the ratio of various integrins (such as ILK overexpression) can be expected to differently affect the fate of these cells. The α5β1 integrin is expressed by the majority of epithelial cell types, including those of breast, skin, lung, gastrointestinal system and the genitourinary tracts.39, 41 Whereas the normal breast ducts and ductules express high levels of α5β1 integrin,41 the expression pattern of integrins in invasive mammary carcinoma cells is atypical, and abnormalities observed include loss, downregulation or improper localization α5β1 integrin at the cell surface.42, 43 The α5β1 integrin is diminished in moderately differentiated carcinoma, and its expression is markedly reduced or undetectable in poorly differentiated adenocarcinoma of the of the breast.40, 41, 42, 43 The α5β1 integrin has been directly implicated in the growth inhibition of tumor cells.40 In contrast, expression of the αvβ3 integrin positively regulates tumor cell proliferation.40 Integrin αvβ3 is minimally expressed in normal epithelial cells and in normal blood vessels and is significantly upregulated within human tumors.44 In several malignancies, tumor cells express αvβ3, and this expression correlates with tumor progression in melanoma, glioma and ovarian, prostate and breast cancer.40, 45 In breast, αvβ3 characterizes the metastatic phenotype as this integrin is upregulated in invasive tumors and in distant metastases.45, 46 Thus, the pattern of integrin expression in the tumor cell is implicated in the enhanced proliferation that is a characteristic of tumor cells. The data we have obtained with the MDA-MB-435 cells are consistent with a role for ILK in the modulation of integrin expression and integrin-regulated cellular proliferation. The MDA-MB-435 studied here were chosen for the metastatic property that easily formed tumors in nude mice. We speculate that the growth-suppression role described for ILK using this model is a general phenomenon that occurs in all cell lines in response to changes in ILK levels.
In the context of a highly invasive mammary carcinoma cell, we have demonstrated that ILK transfection causes a reversal in ratio of expression of integrin α5β1 to αvβ3. However, it is highly likely that the levels of other integrins such as α6β1 may also change either directly in response to ILK or indirectly due to changes in the availability of β1 or β3 integrins. This change is associated with a decrease in the cells' ability to transmigrate in vitro and metastasize in vivo. It is well documented that changes in how cells interact with their environment via altered, modulated or regulated integrin interaction can have dramatic and far-reaching consequences for both normal and pathologic conditions. Studies have shown that perturbations of certain integrins (by either ligation or treatment with certain anti-integrin antibodies) can generate signals which result in an increase in intracellular pH, and Ca2+ levels, changes in inositol lipid synthesis, tyrosine phosphorylation of pp125FAK, activation of p34/cdc2 and cyclin A, activation of protein kinase C, mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase, Ras, and NF-κB.40, 47 Furthermore, the αvβ3 on melanoma cells has been shown to bind and localize proteolytically active MMP-2 at the cell surface which appears to facilitate cell mediated collagen degradation and directed cellular invasion.48 Clearly, integrins are key molecules to integrate intrinsic and extrinsic events of the cellular behavior. They profoundly influence oncogenesis and the metastatic process. Thus, changes in integrin composition brought about through restoration of ILK protein expression in invasive breast carcinoma cells can be expected to bestow completely different, and even opposite effects, compared to the effects of ILK protein overexpression in normal cells.
A likely mechanism for the growth and metastasis of the breast cancer cells MDA-MB-435 is shown in Figure 8. Focal contact with the extracellular matrix is a fundamental mechanism by which cells initiate intracytoplasmic signaling in order to regulate differentiation, growth, attachment, migration, invasion and metastasis. The stoichiometry of integrins, cytoskeletal elements and ILK is probably an essential factor for the proper functioning of these cellular processes. Regulation of cellular ILK level is critical for maintaining the integrin repertoire of the cell that is essential for normal interactions between cells and the extracellular matrix. The highly invasive and metastatic cell line MDA-MB-435 has an abnormal integrin repertoire and altered cell matrix interaction due to loss of ILK expression. The increased expression of αvβ3 and reduced expression of α5β1 integrin and perhaps changes in other integrins as well make these cells highly proliferative, invasive and metastatic both in vitro and in vivo. These cells become less invasive and metastatic upon ILK gene transfer, which restores the stoichiometry of the integrins and cytoskeletal elements. ILK downregulates αvβ3 and upregulates the expression of αvβ1 integrin and could as well affect the expression levels of other integrins not examined here. This in turn leads to G1 cell-cycle arrest, decreased migration and cues for tissue preservation. It is highly likely that the effects seen are accompanied by composite changes in many other cell surface integrins and receptors as well. In breast cancer, altered cell matrix contact due to altered integrins has been shown to be a prerequisite for metastasis.49 Based on this model, we speculate that ILK transfection in normal cells alters their integrin profile and integrin-cytoskeletal stoichiometry, leading to a decrease in cell extracellular matrix interactions. This in turn induces anchorage-independent growth,36 cell-cycle progression37 and tumorigenicity in nude mice.38 Clearly, the functions of ILK are more complex than previously envisioned, and the divergent and often paradoxical effects mediated by ILK may depend on the particular cell type, cell environment and the cell-specific integrins that are activated.
In conclusion, we have shown that the loss of ILK expression is associated with the acquisition of a malignant breast tumor phenotype and that ILK may directly act as a growth suppressor, presumably by controlling cell division and by modulating the levels of integrins. The absence of the ILK may promote uncoordinated G1 cell-cycle progression, allowing cells to bypass the normal signaling processes regulated by growth factors and cell anchorage, leading to tumorigenesis.
We thank Dr. J. Price for providing the MDA-MB-435 cells, Dr. J. Drazba for assistance with fluorescence microscopy and Dr. H. Ramirez for help with the animal experiments. DNA was sequenced by the Cleveland Clinic Sequencing Core. We are grateful to Dr. B. Williams for his encouragement and support. This work was supported by grants from the U.S. Army Medical Research and Materiel Command under DAMD17-96-1-6052 and DAMD17-00-1-0205 to P.K.
- 17Neoplasms associated with the Beckwith-Wiedemann Syndrome. Perspect Pediatr Pathol 1976; 3: 255–72., .
- 43Cellular growth and survival are mediated by β1 integrin in normal human breast epithelium but not in breast carcinoma. J Cell Sci 1999; 108: 1945–57., , .