cDNA, complementary DNA; DAL-1, differentially expressed in adenocarcinoma of the lung (gene); ECIS, electrical cell-substrate impedance sensor system; ERM, ezrin, radixin and moesin; FERM, 4.1/ezrin/radixin/moesin; LOH, loss of heterozygosity; NF2, neurofibromatosis 2 gene.
The tumor suppressor gene differentially expressed in adenocarcinoma of the lung-1 (DAL-1), which shares significant homology with members of the 4.1/ezrin/radixin/moesin/neurofibromatosis 2 (4.1 ERM/NF2) protein family, was first identified for its ability to suppress growth in lung cancer cell lines.1 Analysis of chromosomal band 18p11.3, to which DAL-1 localizes, showed that loss of heterozygosity (LOH) in this region occurred not only in nonsmall cell lung carcinomas but also in invasive ductal carcinomas (IDC) of the breast and astrocytic tumors.2 More recently, chromosome 18p LOH events have been identified in more than 55% of microdissected ductal carcinomas in situ (DCIS) tumors,3 suggesting that DAL-1 alteration may be important early in the progression of breast disease. In addition, DAL-1 protein loss has been documented in 60% of sporadic meningiomas,4 indicating that DAL-1 loss may have broader implications for tumor development.
Like other 4.1/ERM/NF2 family members, endogenous DAL-1 has been localized to the plasma membrane near points of cell-cell contact.1 Maintenance of cell-cell and cell-substratum interactions, as well as cytoskeletal organization, is important for controlling cellular growth and differentiation. Disruption of attachment via loss of β1-integrin5–7 or disturbance of adherens junctions through loss of E-cadherin or β-catenin is associated with progression of breast tumors.8–11 The integrity of cellular adhesion interactions and associated intracellular signaling events can also be influenced by proteins that connect the cell membrane to the cytoskeleton such as 4.1 family member NF2 (merlin/schwannomin).12–15
Normal cellular growth can also be regulated by signals affecting the rate of apoptosis. In normal breast tissue, highly regulated apoptosis is important for maintaining normal homeostasis.16–18 Disruption of cytoskeletal organization can occur at the early stages of epithelial cell apoptosis.19 Therefore, changes in proteins that serve to link the plasma membrane to the cytoskeleton, such as members of the expanding protein 4.1/ERM/NF2 superfamily, may also play a role in the process of apoptosis. A specific role for ezrin in regulating apoptosis is supported by its loss along with α-actinin at the plasma membrane/cytoskeleton interface in neutrophil apoptosis.20 In addition, ERM proteins are dephosphorylated and redistributed to the cytoplasm as a result of human FasL-induced apoptosis in mouse L fibroblast cells.21
To investigate whether DAL-1 is indeed an important growth suppressor in breast cancer cells and to begin to characterize the cellular processes affected by re-expression of this gene, constitutive and inducible DAL-1-expressing MCF-7 cell lines were generated. Assays were performed to analyze growth suppression and apoptosis as well as alterations in population doubling times and attachment to extracellular matrix proteins. In addition, steady-state adhesive behavior was measured using the new electrical cell substrate impedance system.22–24 Our investigations revealed that expression of DAL-1 protein suppresses growth in breast cancer cells in part by inducing apoptosis but also increases the ability of MCF-7 cells to adhere to a variety of matrices and maintain a higher level of steady-state adhesion.
MATERIAL AND METHODS
The following breast cancer cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD): MCF-7, SkBr-3, T-47D and MDA-MB-231. Cells were maintained as recommended by ATCC. Medium was replenished every 2–3 days, and cells were kept at less than 80% confluency.
Growth suppression and cell proliferation assays
The effect of DAL-1 expression on growth of breast cancer cells was analyzed by calcium phosphate transfection of equimolar amounts of pcDNA3-2A3, containing the DAL-1 cDNA and empty control vector pcDNA3 as described previously.1 Transfected cells were selected in 400 μg/ml of G418 (Geneticin; Life Technologies, Bethesda, MD) for 10–14 days. At least 2 independent transfections were performed for each cell line.
For proliferation assays, 6-well plates were inoculated in triplicate with 10,000 cells/well of clone 27. DAL-1 protein expression was induced in 1 set of wells by the addition of 1 μM muristerone. Cultures were maintained under the appropriate conditions for 6 days, and on every second day, 3 independent wells were harvested and the cell number determined. Trypan blue (InVitrogen, La Jolla, CA) staining was used to exclude dead cells. The log10 of live cells was plated versus the number of days in culture to determine whether reintroduction of DAL-1 protein expression affected the proliferation rate of MCF-7 cells.
Generation of DAL-1-expressing MCF-7 cell lines
Constitutive DAL-1-expressing MCF-7 and pcDNA3 vector-only control cell lines were established by calcium phosphate transfection with 25 μg of plasmid.1 Transfected cells were placed in 400 μg/ml G418 selective growth medium for 10–14 days to allow individual colony formation. An average of 24 G418-resistant clones for each construct were examined for expression of DAL-1 protein by immunohistochemistry and Western analysis. MCF-7 DAL-1 clones 10 and 20 and MCF-7 pcDNA3-only clone 1 were chosen for further analysis based on the level and distribution of DAL-1 protein expression.
The ecdysone-inducible expression system (InVitrogen) was used to generate an MCF-7 DAL-1-inducible cell line. Generation of this cell line required sequential transfection of the retinoid X receptor-containing pVgRXR vector followed by the pIND-DAL-1 construct. Doubly transfected clones were selected in 50 μg/ml zeocin and 400 μg/ml G418. Approximately 150 individual clones were phenotypically screened by immunohistochemistry for tightly controlled, inducible DAL-1 protein expression after exposure to 1 μM muristerone for 48 hr. Only a single inducible DAL-1-expressing clone, clone 16, expressed DAL-1 protein in a tightly controlled fashion, however nonhomogeneously. Additional subcloning by limiting dilution resulted in the isolation of inducible cell line clone 27, in which more than 80% of cells stained positive for DAL-1 protein after muristerone induction. This cell line was used in our study for analysis of DAL-1-associated effects.
Assays for apoptosis
MCF-7 and MCF-7 DAL-1-inducible clone 27 cells were plated at 10,000 cells/12 mm coverslip and allowed to grow in the presence or absence of 1 μM muristerone for 1–4 days. Cells were assayed for apoptosis by either the terminal deoxynucleotidyl transferase-mediated nick end labeling (TUNEL) assay (Roche Diagnostics, Indianapolis, IN) or annexin V-fluorescein isothiocyanate (FITC) reactivity (BD PharMingen, San Diego, CA) using the manufacturers' recommended protocols. Briefly, for the TUNEL assay, cells were fixed for 30 min in fresh 4% paraformaldehyde followed by permeabilization with 0.1% Triton X-100. Cells were then incubated with terminal transferase and FITC-conjugated dUTP for 90 min at 37°C. For annexin V staining, cells were incubated for 20–30 min at room temperature in 1:100 annexin V-FITC, washed and fixed in 4% paraformaldehyde at room temperature. In both assays, cells were then incubated in Hoescht 33258 (2 μg/ml) for 5 min, placed cell side down in DABCO (1,4-diazabicycol[2.2.2]octane 2.5% in glycerol; Sigma, St. Louis, MO) on microscope slides and sealed. At least 200 cells per coverslip were counted using Hoescht-stained nuclei. The number of apoptotic-positive cells was scored as a percentage of the total cell number. At least 9 coverslips from 3 independent experiments were analyzed by these methods.
Western analysis of DAL-1 expression
Western analysis was performed on all cell lines to confirm the presence or absence of the transfected DAL-1 protein as previously described. DAL-1 protein was detected with an immunopurified polyclonal rabbit anti-DAL-1 antibody1 and visualized with the Immun-Star Chemiluminescent Protein Detection System (Bio-Rad, Richmond, CA) following the manufacturer's instructions except that PVDF filters were blocked using 5% nonfat dry milk in TBS (20 mM Tris/0.5 M NaCl, pH 7.5), and primary and secondary antibodies were diluted in 1% nonfat dry milk in TBS prior to application. Exposure to Kodak BioMax Film determined the presence or absence of DAL-1 protein.
Immunocytochemistry was used to detect the location and level of expression of DAL-1 protein. Cells were grown on 12 mm coverslips and fixed in 100% methanol for 20 min at −20°C. DAL-1 antibody 3A1, diluted 1:1,000 in antibody buffer (Hanks' balanced salt solution, 5% donor calf serum, 0.1% NaAzide) was applied to the cells for 30 min at room temperature, detected using a 1:100 dilution of goat anti-rabbit IgG secondary fluorescein-labeled antibody (Alexa 488; Molecular Probes, Eugene, OR) and incubated at room temperature for 1 hr. Nuclei were stained with Hoechst 33258 (2 μg/ml). Coverslips were placed cell side down on a microscope slide in DABCO (2.5% in glycerol; Sigma) and sealed. Stained cells were visualized using appropriate filters on a Nikon Eclipse E800M Microscope and imaged using a Sony Cats Eye Digital Photo Camera and Imaging System.
Cell attachment assay
MCF-7 cell lines were plated in fibronectin-free medium at 50,000 cells/well in 96-well tissue culture plates pretreated with either 50 μg/ml of fibronectin (Sigma Aldrich, Poole, UK), vitronectin (InVitrogen), laminin (Sigma Aldrich), Matrigel (diluted 1:8 in medium; Becton Dickinson, Mountain View, CA) or 1% Bovine Serum Albumin Fraction V (Sigma Aldrich) for 1 hr at 37°C, followed by blocking with 1% BSA. Cells were allowed to adhere for 1 hr, after which nonattached cells were washed away with 1× PBS and shaking. The remaining attached cells were fixed in 1% glutaraldehyde, stained in 0.1% (w/v) crystal violet and lysed in 1% SDS. The intensity of stain, directly proportional to the number of adherent cells, was quantitated by absorbance at 540 nm using a Titertek Multiskan MCC/340 microplate reader (Huntsville, AL). Each cell line was tested in 3 separate wells over 3 independent experiments.
ECIS cell attachment assay
The electrical cell-substrate impedance sensor system (ECIS; Applied Biophysics, Troy, NY) was applied to analyze the ability of DAL-1 protein to alter the steady-state cell adhesive properties of MCF-7. Cells from MCF-7 clones were plated at 200,000 cells/well in ECIS chambers with 400 μl medium to give a confluent layer of cells across each electrode. For the inducible MCF-7 clone 27 cell line, DAL-1 protein expression was induced 48 hr prior to ECIS analysis by exposure to 1 μM muristerone. Cell attachment behavior was monitored for up to 3 hr and graphed as increasing resistance.22–24 Both the rate of initial attachment as well as the level of steady-state adhesion over time can be visualized. Cells are plated in vast excess to ensure complete coverage of the 250 μm electrode surface. Complete single-layer electrode coverage was confirmed microscopically prior to data analysis.
To analyze ECIS results, trace values are adjusted by subtraction of each individual electrode's resistance values for medium-only conditions. Values are then averaged together to achieve an overall ECIS trace within a single experiment. To compare data between ECIS runs, adjusted averaged traces are normalized by the resistance value of the MCF-7 parental cell line at equivalent time points. These normalized traces can then themselves be combined to give an overall adhesion profile for each cell line.
For growth suppression assays, significance was determined by normalizing data for each cell line relative to the control plasmid group and performing a Student's t-test to calculate p-values resulting from comparing the control group means with the DAL-1-transfected group means. The alpha level used for these comparisons was α = 0.05. A Student's t-test was also used to determine significance values in the apoptosis-related annexinV and TUNEL assays. For cell attachment assays, using both DAL-1 constitutive- and inducible-expressing cell lines, significance (p-value) was calculated using Wilcoxon's 2-sample nonparametric exact tests, with Hochberg's adjustment. Family-wise alpha levels were set at α = 0.05. With all cases, calculated p-values are reported in the appropriate results section.
DAL-1 suppresses growth in several breast cancer cell lines
The expression profile of DAL-1 was analyzed in 6 breast cancer cell lines by Western blot analysis. The lung cancer cell line NCI-H460, already shown to express high levels of DAL-1 protein,1 was used as a positive control. These studies revealed that MCF-7, SkBr-3, Hs578t, MDA-MB-231 and BT-20 were negative for DAL-1 protein, whereas T-47D expressed endogenous protein at levels comparable to those of NCI-H460 (Fig. 1a). Immunocytochemistry on T-47D cells using the polyclonal anti-DAL-1 antibody showed that this protein concentrates in regions of cell-cell contact (Fig. 1b), as previously described for NCI-H460.1
To determine its ability to suppress growth in a variety of breast cancer cells, the DAL-1 cDNA1 was transfected into these cell lines in a minimum of 2 independent experiments, and cells were selected for 10–14 days. Relative to vector-only control transfections, cell yield was significantly affected by DAL-1 expression in MCF-7 (53.9 ± 4.5%; p < 0.0001), MDA-MB-231 (46.7 ± 6.1%; p < 0.0001) and SkBr-3 (27.6 ± 5.1%; p < 0.0001; Fig. 1c) but not in T-47D (89 ± 7.9%; p = 0.1919), which expresses endogenous DAL-1. Given that MCF-7 has been extensively studied with respect to the molecular pathways of breast cancer pathogenesis, we chose to continue to use this model cell line to begin to characterize the role of DAL-1 in breast cancer.
Generation of DAL-1-expressing MCF-7 cell lines
MCF-7 cells were transfected with the DAL-1 cDNA, and 2 constitutively expressing cell lines, clones 10 and 20, as well as a pcDNA3 vector only cell line (clone 1), were established. Western analysis using the anti-DAL-1 polyclonal antibody confirmed the presence of the transfected DAL-1 protein (Fig. 2a). Immunocytochemical analysis detected DAL-1 protein in clones 10 and 20 at the expected location, although the relatively small cytoplasm in MCF-7 cells makes this staining less defined than in T-47D cells (Fig. 2b). Clone 10 represented a heterogeneous population, with 40–50% of cells exhibiting intense staining for DAL-1 protein; clone 20 cells were more homogeneous, with >90% of cells expressing a moderate level of DAL-1 protein.
Given that DAL-1 suppresses growth in populations of MCF-7 cells, it is possible that these constitutively-expressing cell lines may have adapted to tolerate expression of this tumor suppressor gene by acquiring secondary changes and so may not represent the best model for analysis of its function. To address this issue, we generated an inducible DAL-1-expressing cell line. Inducible-expression clones offer the advantage of allowing the selection of clones in the absence of protein as well as control over the timing of protein expression. Importantly, uninduced cells of the same clone represent the best control for cells that have been induced, as they are clonally identical. Clonally unrelated cells harboring an empty inducible cassette are potentially confounding as they are not clonally identical. A single clone, clone 27, exhibiting tightly controlled, inducible and homogeneous DAL-1 expression, was isolated. This clone exhibited maximum levels of induced DAL-1 protein after 24–48 hr of exposure to 1 μM muristerone (Fig. 2c).
Effect of DAL-1 expression on cell growth and programmed cell death
The mechanism(s) by which DAL-1 suppresses growth is currently unknown. One possibility is that DAL-1 could reduce the rate of cell proliferation. Analysis of growth curves constructed for MCF-7, clone 1 vector control and DAL-1 constitutively expressing clones 10 and 20 showed no such alteration (data not shown). This finding suggests that cells that have been selected for their ability to survive in the presence of exogenous DAL-1 protein are not altered in their ability to proliferate. However, growth curves utilizing the inducible DAL-1 clone 27 cell line showed reduced cell proliferation (Fig. 2d) concomitant with induced expression of the DAL-1 protein (Fig. 2c). This growth suppression was observed over the 6-day course of the experiment. Together with the data from Figure 1c, our studies support the hypothesis that DAL-1 protein expression suppresses growth in breast cancer cells.
One mechanism by which DAL-1 transfection could reduce cell yield is through the induction of cell death. To investigate this possibility, cells grown in the presence or absence of 1 μM muristerone for 1–4 days were stained with an annexin V antibody to detect phosphatidylserine on the surface of apoptotic cells as well as by TUNEL assay to measure DNA fragmentation (Fig. 3). No significant alteration in the percentage of annexin V-positive cells was observed in parental MCF-7 cells with or without muristerone treatment over the 4-day experiment (p = 0.2508 on day 4), demonstrating that muristerone does not itself affect the basal level of apoptosis in these cells. However, a significant increase in the number of apoptotic cells was detected upon DAL-1 induction, with levels exceeding 5% annexin V-positive (p < 0.0001) and 7% TUNEL-positive (p < 0.0055) cells by day 1 compared with only 0.64 and 0.5% for the noninduced cells, respectively. This DAL-1 expression-associated increase in apoptotic cells remained through day 4 [7.25% by Annexin staining (p = 0.0022) and 8.1% by TUNEL assay (p < 0.0001)], suggesting an ongoing apoptotic process. These findings indicate that the DAL-1 protein is capable of triggering an apoptotic response in MCF-7 in the absence of exposure to additional recognized apoptotic stimuli.
Effect of DAL-1 expression on cell attachment
Although DAL-1 is not itself an adhesion molecule in the classical sense, we hypothesized that because of its localization to the sub-plasma membrane and its similarity to other family members that function as linker proteins between the cytoskeleton and the plasma membrane, it might modulate a tumor cell's capacity to adhere to various extracellular matrices, potentially through interaction with transmembrane proteins and organization of the underlying cytoskeleton. To test this hypothesis, parental MCF-7 and vector control cells, as well as constitutive clones 10 and 20 and inducible clone 27, were assayed for their ability to attach to BSA, fibronectin, laminin, vitronectin and Matrigel. BSA was used as a control for general cell attachment behavior in this assay. As shown in Figure 4a, parental MCF-7 cells showed low affinity to BSA and fibronectin but stronger attachment to laminin, Matrigel and vitronectin. Both constitutively-expressing DAL-1 clones (10 and 20) showed a significantly increased ability over MCF-7 to attach to fibronectin (both clones p < 0.0001) and Matrigel (clone 10, p = 0.0114; clone 20, p = 0.0315) while maintaining low nonspecific binding to BSA. Only clone 20 showed a significantly increased ability to attach to laminin (p = 0.0153).
Attachment assays using the inducible cell line clone 27 after 24, 48 and 72 hr of DAL-1 expression induced by 1 μM muristerone confirmed these results (Fig. 4b). Noninduced clone 27 cells showed attachment profiles similar to those of parental MCF-7 cells, whereas attachment increased after only 24 hr of DAL-1 expression an average of 2.2-fold on Matrigel (p < 0.0002) and 1.97-fold on fibronectin (p < 0.0001). Furthermore, by varying the concentration of fibronectin, it was shown that these attachment increases were directly proportional to ECM concentration and not nonspecific protein interactions (data not shown).
Measurement of steady-state cell adhesion using the ECIS system
As the effects of DAL-1 on cell adhesion are likely to be indirect, we used a new technology more suited to measuring cell attachment. ECIS allows quantitative measurement of both the rate of attachment of cells and their steady-state adhesive behavior. As cells adhere to a 250 μm diameter electrode and to each other, both cell-to-substrate and cell-to-cell contacts impede the flow of current across this electrode, measured as an increase in resistance.22–24 Cells are plated in excess to ensure coverage of the entire electrode surface. This coverage was confirmed microscopically before data were accepted for analysis. As the ECIS traces reveal (Fig. 5a), resistance values at all timepoints after the first 30 min for clones 10 and 20 were higher than those for the MCF-7 and vector control cells, indicating greater steady-state adhesion for MCF-7 cells expressing DAL-1. However, the rate at which these cell lines initially attached to the electrode, as viewed by the slope of the trace in the initial 30 min, did not differ, indicating that DAL-1 modulates the strength of these contacts but not the rate at which they form.
The ability to study the rate as well as the overall steady state of adhesion is a significant advantage of this new technology over traditional assays. Data combined from 4 separate ECIS experiments (Fig. 5b) clearly revealed an approximate 2-fold increase in the steady-state attachment behavior for both MCF-7 clones 10 and 20 compared with vector control. ECIS traces, comparing induced and noninduced clone 27 cells (Fig. 5c), also confirmed that the presence of DAL-1 protein consistently increases the adhesive properties of MCF-7 cells but not the initial rate of attachment, which is more likely to be modulated by extracellular portions of classical adhesion proteins. This is further evidence that DAL-1 affects the behavior of protein complexes that regulate the ability of MCF-7 cells to adhere.
Maintenance of cell-cell and cell-matrix adhesion interactions are important in controlling cellular growth and differentiation in breast tumors and cell lines.5, 7, 8 Studies on protein 4.1 family members such as the tumor suppressor NF2, which link proteins in the membrane to cytoskeletal components, show the importance of such connections and the consequence of maintaining their associated signaling pathways in cancer.13, 17 DAL-1, a new member of the protein 4.1 superfamily, had been previously shown to localize at the cell membrane, suppress growth in a number of lung cancer cell lines and undergo LOH.1 The region of chromosomal band 18p11.3 to which DAL-1 has been localized was recently found to undergo LOH in >60% of primary breast tumors of all stages.3 Based on these results and the absence of DAL-1 expression in a number of breast cancer cell lines, we began to examine whether this protein has an important role in breast tumorigenesis.
Our studies revealed that expression of the DAL-1 protein could significantly reduce cell yield when it was introduced into DAL-1 null breast cancer cells. Further analysis in MCF-7 cells showed that DAL-1 expression induced apoptosis and also modulated cell attachment. These studies were initiated using the original 65 kDa DAL-1 protein isolate (cDNA clone 2A3) identified as a tumor suppressor gene.1 Clone 2A3 has recently been shown to represent a splice variant of the proposed full-length protein KIAA0987,25 yet it contains all the identified functional domains for DAL-1. Studies in other cell lines have shown that this 65 kDa DAL-1 protein and the full-length KIAA0987 protein equivalently suppress tumor cell growth and exhibit the same profile of interactions with NF2 protein interactors,25 indicating that KIAA0987 and the DAL-1 protein used in this report25 are functionally comparable.
Given DAL-1's position at the cell membrane, its ability to suppress growth and its homology with protein 4.1 family members, we hypothesized that it may function by indirect modulation of cellular adhesive protein complexes. Interestingly, immunocytochemical analyses have demonstrated DAL-1 colocalization with the adherens junction proteins E-cadherin and β-catenin.1 E-cadherin and β-catenin proteins have been reportedly detected in ezrin immunoprecipitates, and ezrin was found in both E-cadherin and β-catenin immunoprecipitates,26 suggesting that DAL-1 could also interact with these proteins.
Recent studies confirm that members of the protein 4.1/ERM/NF2 superfamily can affect adhesion, albeit in a complex manner.27–29 For example, in rat schwannoma cells, NF2 expression transiently reduces cell attachment to fibronectin, with levels returning to normal after 3 hr.27 In the present report, reintroduction of DAL-1 into MCF-7 cells increased cell attachment on all extracellular matrix proteins as measured at 1 hr using similar short-term adhesion assays. This modulation was modest but significant, with the increase ranging from 1.6- to 2-fold and correlating with the presence of specific matrix protein coatings. The observation that NF2 and DAL-1 differentially affect adhesion is consistent with other studies identifying biologic differences between the 2 proteins. Such differences include the recently reported lack of DAL-1 binding to actin,25 whereas NF2 binding has been measured.30 In addition, NF2 suppresses the growth of schwannoma cells, whereas DAL-1 does not,25 further arguing that these 2 protein 4.1 molecules have different functional properties and can differentially regulate cell growth when assayed in neural cell types.
Along with traditional attachment assays, we utilized a new quantitative technology, ECIS, to measure alterations in cell attachment as a result of DAL-1 expression. Techniques such as the traditional attachment assay simply inflict a shearing force to cells after a brief period of attachment, detaching those not tightly anchored and quantifying those cells remaining attached. This approach is likely to emphasize the initial stickiness of the detached, trypsinized MCF-7 cell and may predominantly measure the presence of extracellular attachment sites. ECIS, however, focuses on the kinetic aspects of initial cell attachment and prolonged steady-state adhesion, making it a sensitive and versatile approach for studying the multiple aspects of cell adhesion. This analysis includes steady-state cell-surface and cell-cell adhesion, processes more likely to be influenced by intracellular molecules such as the 4.1 superfamily. In addition, since it has been shown that cells anchor essentially in the same way to the gold-film electrodes as they do to normally treated polystrene culture dishes, results extracted from ECIS data are both relevant and significant.31
ECIS can also be used to analyze more complex properties of membrane proteins. For example, confluency as well as serum starvation have been shown to upregulate the level of merlin protein in NIH 3T3 cells, suggesting that this 4.1 family member's protein expression is regulated by responses to cell-cell contact and other growth-suppressive signals.32 Similar studies can be envisioned for DAL-1-expressing cells as well, now that initial attachment alterations have been characterized.
Changes in cellular interactions at the epithelial plasma membrane through alteration or loss of associated gene products may be early events in the development and progression of cancer. The process of anoikis, whereby disturbance in cell-matrix interactions and loosening of the degree of cellular adhesion, elicits an apoptotic response, has been well documented for epithelial cells.33 The discovery, therefore, that re-expression of DAL-1 protein can induce programmed cell death as well as increase cell attachment and adhesion was surprising. One possible explanation for this apparent discrepancy is that those experiments were performed in MCF-7 cells, an established cancer cell line with potentially altered adhesion and apoptotic properties. It is likely that these cells have escaped anoikis in the course of tumor evolution.
Studies on some members of the protein 4.1 family support the idea that modifications in cellular adhesion by these proteins can directly or indirectly influence apoptosis.20, 21 Although the exact mechanism is yet to be deciphered, 1 hypothesis might be that DAL-1 induces reorganization of signaling complexes near the membrane, which could enhance both adhesion and apoptosis. For example, DAL-1 might be involved in transducing signals from the environment through adhesion membrane receptors, thereby initiating signaling pathways involved in regulating apoptosis. Such pathways could include the proapoptotic JNK and antiapoptotic ERK cell signaling pathways, which can regulate the balance of expression and activity of bcl-2 family proteins.34 Studies aimed at placing DAL-1 in signaling pathways are ongoing.
Our report demonstrates that DAL-1's effects can now be extended to include suppression of breast cancer cell growth and that DAL-1 affects the behavior of MCF-7 cells via modulation of both cellular adhesion properties and apoptotic pathways. Although DAL-1 is not a classical transmembrane adhesion molecule, its effect on growth may be directly or indirectly related to its ability to regulate cell-cell and cell-substrate attachment by altering the physiologic or morphologic characteristics of the cell membrane, a property critical both early and late in the pathogenesis of breast cancer. Such modulations may in some circumstances signal an apoptotic response. Future experiments focusing on the cellular proteins interacting with the DAL-1 protein may help to clarify the important tissue-specific growth-suppressing effects of this novel tumor suppressor. Several proteins have recently been reported to interact with protein 4.1 family members. ERM proteins have been shown to bind to the integral membrane protein CD44 via their N terminus35 and to F-actin via their C terminus.36 NF2 has been shown to interact directly with ezrin, βII spectrin, a sodium hydrogen exchange regulatory factor (NHE-RF) and a zinc finger protein termed SCHIP-1 that specifically binds to spliced isoforms and mutant merlin proteins.37–39 More recently, DAL-1 has been shown to bind in vitro to some of the same proteins as NF2.25 Elucidation of these known and other novel interactions and their effect on the apoptotic machinery will define the role of DAL-1 in tumor suppression in lung and breast tumors.
The authors thank Dr. O. Bögler for his helpful discussions and review of this manuscript, Dr. S. Rempel and B. Golembeski for assistance with the attachment assays and Dr. C. Keese, Dr. G. Bowlin and J. Alexander for advice on the ECIS experiments. I.F.N. was the recipient of grant CA777330 from the NCI.