An increased expression of focal adhesion kinase (FAK) in a variety of cancers is associated with a poor disease prognosis. To study the role of FAK in breast tumor growth and metastasis formation, we used conditional doxycycline-regulated expression of a dominant-negative acting splice variant of FAK, FAK-related non-kinase (FRNK), in MTLn3 mammary adenocarcinoma cells in a syngeneic Fischer 344 rat tumor and metastasis model. In cell culture, doxycycline-mediated expression of FRNK inhibited MTLn3 cell spreading and migration in association with reduced formation of focal adhesions and phosphorylation of FAK on Tyr397, but FRNK did not cause apoptosis. Continuous expression of FRNK decreased the primary tumor growth in the mammary fat pad by 60%, which was not due to induction of apoptosis. Lung metastasis formation was almost completely prevented when FRNK was already expressed 1 day before tumor cell injection, whereas expression of FRNK 11 days after injection did not affect lung metastasis formation. FRNK expression during the first 5 days was sufficient to block metastasis formation, excluding the possibility of FRNK-induced dormancy of tumor cells. Together, these data fit with a model wherein FAK is required for breast tumor cell invasion/migration processes that take place in the early phase of metastasis formation. Our findings suggest that FAK is a good candidate for therapeutic intervention of metastasis formation.

Breast cancer is the most frequently occurring type of cancer in women. Primary breast cancer tumors can be removed or irradiated relatively easily, but the distant metastases are hard to treat. Improved insights into the mechanisms involved in the tumor-metastasis process are necessary to define potential novel drug targets that can be used to combat metastasis formation. The formation of a metastasis involves different steps, including detachment, migration, invasion, extravasation, and proliferation of the cancer cells, which require the appropriate extracellular signals. Focal adhesions, the closest contacts between cells and the extracellular matrix (ECM), are important sites for these signaling events. Focal adhesion kinase (FAK) is a non–receptor protein tyrosine kinase that is important in cell-ECM-mediated signaling. Immunohistochemical staining of different cervical and breast carcinomas for FAK revealed that increased expression marks the malignant transformation of squamous cells of the uterine cervix and epithelial cells of breast ducts (1). Increased expression of FAK has been shown in various tumors, including tumors derived from the breast (1), head and neck (2), and ovary (3). Furthermore, the involvement of FAK in cancer is emphasized by the observation that amplification of chromosome 8q (the FAK encoding region) is correlated to a poor disease prognosis (4). Moreover, FAK heterozygous mice develop less papillomas in a chemical skin carcinogenesis model (5).

FAK plays a central role in controlling cell migration, proliferation, and survival. Several domains of FAK are important for these functions. The NH2-terminal region of FAK harbors a FERM domain that mediates binding of FAK to integrins and growth factor receptors and thereby FAK activation (6). The COOH-terminal focal adhesion targeting domain is required for focal adhesion localization through binding to paxillin and talin (7). FAK activity is controlled by autophosphorylation of Tyr397, providing a binding site for Src family kinases as well as phosphatidylinositol 3-kinase (8). Consequently, Src activity mediates phosphorylation of Tyr576/Tyr577 in the kinase domain and Tyr861 and Tyr925 in the COOH-terminal region (6). The FAK-Src complex mediates phosphorylation of several focal adhesion–associated adapter proteins, including paxillin (9) and p130CAS (10). The focal adhesion targeting domain as well as a splice variant of FAK, the FAK-related non-kinase (FRNK), which lacks the entire NH2 terminus and the kinase domain, are often used to inhibit FAK localization at focal adhesions and thereby many of the cell biological functions of FAK (11).

Several studies suggest a direct role for FAK in tumor formation and progression. Increased expression of active CD2-FAK in Madin-Darby canine kidney cells enables s.c. tumor formation in nude mice (12). Similarly, overexpression of FAK in U-215MG human malignant astrocytoma cells results in increased tumor cell proliferation in a xenograft model (13). Oppositely, expression of FRNK in human epidermoid carcinoma cells reduces cellular outgrowth when these cells are inoculated onto chorioallantoic membranes of chick embryos (14). Stable overexpression of FRNK in NIH 3T3 fibroblasts transformed with the dominant oncogene v-Src, inhibits cell invasion in vitro and reduces experimental metastases formation in nude mice (15). Constitutive expression of FRNK in B16-F10 melanoma resulted in a 50% reduction in the number of lung metastases (16). Because FAK mediates survival signaling, it cannot be excluded that, in both v-Src NIH 3T3 cells and B16-F10 cells, selection occurred in vitro during stable transfection, favoring the selection of cells that were relatively resistant to FRNK-induced apoptosis. Such a selection may affect the characteristics and thereby the in vivo behavior of the tumor cells. Moreover, because FAK has been linked to both migration/invasion (17, 18) and survival/proliferation (19, 20) processes, thus far, it remains unclear whether the in vivo effects of FRNK are related to early inhibition of migration/invasion processes and/or to inhibition of tumor cell survival/proliferation in the metastases. Alternatively, FRNK expression may cause dormancy of tumor cells in the distant target organs that eventually might grow out into full metastases when FAK is no longer inhibited. To study these important issues, conditional models to modulate the function of FAK are required.

Here, we examined the importance of FAK signaling during primary tumor growth as well as during the different steps in metastasis formation of breast tumor cells. For this purpose, we made use of the metastatic mammary adenocarcinoma cell line MTLn3 (21), which has been used both in vitro and in vivo to study the mechanisms of metastasis formation (22). We generated a cell line with a doxycycline-regulated expression of hemagglutinin (HA)–tagged FRNK: tetFRNK-MTLn3 cells. This cell line enabled us to study, to our knowledge for the first time, the importance of FAK at different phases in the process of metastasis formation. Conditional expression of HA-FRNK in vitro resulted in a reduced attachment, spreading, and migration of the tetFRNK-MTLn3 cells but did not affect cell death. In vivo, HA-FRNK inhibited primary tumor growth, whereas HA-FRNK almost completely prevented experimental lung metastasis formation; expression of HA-FRNK during the first 5 days after tumor cell injection was sufficient for this effect. Expression of HA-FRNK starting 11 days after injection of the tumor cells did not significantly affect the number of lung metastases. Altogether, we conclude that FAK signaling is essential during the initial steps of metastasis formation, most likely by interfering with tumor cell invasion/migration processes.

Chemicals and antibodies. α-Modified MEM with ribonucleosides and deoxyribonucleosides (α-MEM), fetal bovine serum, PBS, trypsin, LipofectAMINE Plus, and geneticin (G418 sulfate) were from Life Technologies (Rockville, MD). Collagen (type I, rat tail) was from Upstate Biotechnology (Lake Placid, NY). Propidium iodide (PI), RNase A, and doxycycline were from Sigma (St. Louis, MO). Annexin V was from Boehringer Mannheim (Basel, Switzerland). Hoechst 33258 and the Alexa protein labeling kit were from Molecular Probes (Leiden, the Netherlands). Hygromycin was from Roche (Penzberg, Germany). All other chemicals were of analytic grade. Primary antibodies were anti-HA (clone 3F10, Roche, Indianapolis, IN) and anti-HA (clone 12CA5, Roche, Indianapolis, IN), anti-paxillin (clone 349; BD Biosciences, Rockville, MD), anti-FAK (Upstate Biotechnology), anti–active caspase-3 rabbit polyclonal (New England Biolabs, Beverly, MA), anti-paxillin-PY118, and anti-phospho-PY397FAK (Biosource Europe S.A., Nivelles, Belgium).

Cell culture and generation of stable cell lines. MTLn3 rat mammary adenocarcinoma cells (21) were cultured as described previously (23). For creating tetFRNK-MTLn3 cells, cells were first transfected with the pTet-on plasmid (Clontech, Mountain View, CA) using LipofectAMINE Plus. Colonies resistant to G418 were selected and tested for inducibility under the Tet-on system by transient transfection with a tetracycline response element promoter (pTRE) luciferase reporter. Next, proper clones were transfected with the pTRE-HA3-FRNK construct (FRNK cDNA was kindly provided by Dr. T.J. Parsons of Dr. T.J. Parsons (University of Virginia Health Science Center, Charlottesville, VA) and inserted between two EcoRI sites in a pTRE-HA3 vector) and a pTK-Hyg plasmid followed by selection with hygromycin. Hygromycin-resistant colonies were evaluated for doxycycline (1 μg/mL) induced expression of HA-FRNK. Selected tetFRNK-MTLn3 cells were cultured in complete medium supplemented with 50 μg/mL hygromycin and 100 μg/mL G418 and used for up to 10 passages during which >90% of the cells maintained doxycycline-induced expression of HA-FRNK.

Cell proliferation, apoptosis, and anoikis assays. tetFRNK cells were cultured for 24 hours in the absence or presence of doxycycline and subsequently trypsinized and replated in 24-well plates (1 × 104 cells per well). Cells were washed twice with PBS, 200 μL milliQ was added, and plates were frozen (−80°C). Lysed cells were incubated with Hoechst 33258 (2 μg/mL) and DNA concentration was measured in a fluorescence plate reader (HTS 7000 Bio assay reader, Perkin-Elmer, Norwalk, CT) with calf thymus DNA as a standard. Cell death was determined by staining the pooled attached and detached cells for Annexin V/PI or by cell cycle analysis as described previously (23). For anoikis experiments, cells were cultured in the absence or presence of doxycycline, serum was withdrawn for 1 hour, and cells were trypsinized. Cells were counted, resuspended to a final concentration of 1 × 106 cells/mL, and incubated in a 15 mL on a roller bank in the absence of serum at 37°C, and cell death was determined by the Annexin V/PI.

Immunoblotting. Frozen lung or tumor tissue was cryosectioned into 10 μm slices followed by lysis in TSE [10 mmol/L Tris-HCl, 250 mmol/L sucrose, 1 mmol/L EGTA (pH 7.4)] supplemented with inhibitors (1 mmol/L DTT, 10 μg/mL leupeptin, 10 μg/mL aprotinin, 1 mmol/L sodium vanadate, 50 mmol/L NaF, 1 mmol/L phenylmethylsulfonyl fluoride). Cells were scraped in ice-cold TSE supplemented with inhibitors. After sonication of either cells or tissue, protein concentrations were determined by the Bio-Rad (Hercules, CA) protein assay using IgG as a standard. Equal amounts of total cellular protein were separated by 7.5% SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA). Blots were blocked with 5% (w/v) bovine serum albumin in TBST [0.5 mol/L NaCl, 20 mmol/L Tris-HCl, 0.05% (v/v) Tween 20 (pH 7.4)] and probed with primary antibody (overnight, 4°C) followed by incubation with secondary horseradish peroxidase–coupled antibody and visualized with Enhanced Chemiluminescence Plus reagent (Amersham Biosciences, Uppsala, Sweden) by scanning on a multilabel Typhoon imager 9400 (Amersham Biosciences).

Immunofluorescence. Culturing and immunostaining of cells were essentially the same as described (23). For staining of tissue sections, paraffin-embedded tissue was cut into 4 μm sections. After deparaffinization, sections were blocked in 1.5% normal goat serum for 1 hour and incubated with primary antibodies (overnight, 4°C) followed by incubation with fluorescently labeled secondary antibodies for 1 hour at room temperature. After washing, slides were incubated with 2 μg/mL Hoechst 33258 and mounted in Aqua PolyMount. Cells and tissue were visualized using a Bio-Rad Radiance 2100 MP confocal laser scanning system equipped with a Nikon (Melville, NY) Eclipse TE2000-U inverted fluorescence microscope and a 60× Nikon objective.

Cell attachment and wound healing assays. Cells incubated for 1 hour in the absence of fetal bovine serum were trypsinized and resuspended to a final concentration of 4 × 105 cells/mL in α-MEM (± doxycycline). Next, cells were replated into six-well plates (2.5 mL) coated with rat tail collagen type I. Cells were incubated for the indicated times at 37°C. At each time point, unattached cells were removed, phase-contrast pictures were taken, and attached cells were trypsinized and counted. For wound healing, confluent cells were serum starved and subsequently wounded with a pipette tip. Photographs were taken at indicated times followed by measuring the wound width.

In vivo tumor growth and metastasis formation. Cells were injected into the fat pad or into the lateral tail vein of Fischer 344 rats (7-8 weeks old, Charles River, Maastricht, the Netherlands). After the fat pad injection, the size of the primary tumors was measured by using calipers starting from day 20. Horizontal (h) and vertical (v) diameters were determined and tumor volume (V) was calculated: V = 4/3π(1/2[√(h × v)]3). For the tail vein injection experiment, rats were pretreated with the NK-depleting antibody (mAb3.2.3) on days −3, −2, and −1 (150 μg in 0.5 mL PBS, i.p.) before the injection of the cells. tetFRNK cells (passage +1) were cultured in the absence or presence of 1 μg/mL doxycycline for 24 hours and cells were trypsinized, counted, washed twice, and resuspended in PBS. Viable cells (1 × 105) in 0.2 mL PBS were injected into the lateral tail vein, and cells (1 × 106) in 0.5 mL PBS were injected into the fat pad. Where indicated, doxycycline [400 μg/mL in 2.5% (w/v) sucrose] was added to the drinking water; control animals received 1.5% (w/v) sucrose in their drinking water, which resulted in equal drinking volumes. At the indicated time points, animals were anesthetized with pentobarbital and the lungs were excised and rinsed in ice-cold PBS. The right lung was cut into three pieces and fixated in isopentane (used to prepare tissue homogenates for immunoblot analysis), 4% paraformaldehyde, or Carnoy's [10% (v/v) acetic acid, 30% (v/v) chloroform in 60% ethanol]. The left lung was injected with ink solution and thereafter destained in water and fixated in Fekete's [4.3% (v/v) acetic acid, 0.35% (v/v) formaldehyde in 70% ethanol]. Pictures of the stained lungs were taken and lung metastases (white dots) were counted.

Statistical analysis. Student's t test was used to determine significant differences between two means (P < 0.05). For the in vivo lung metastasis experiments, statistical significance was determined by the Mann-Whitney rank sum test.

Doxycycline-mediated expression of HA-FRNK in MTLn3 cells inhibits proliferation but does not affect cell survival. To elucidate the role of FAK in the migratory, proliferative, and metastatic behavior of breast tumor cells in vitro and in vivo, we created a MTLn3 cell line that conditionally expresses the COOH-terminal splice variant of FAK, FRNK that has an inhibitory effect on FAK. The obtained tetFRNK-MTLn3 clone had a tightly concentration-dependent regulation of HA-FRNK expression and >90% of the cells stained positive for HA-FRNK after doxycycline treatment as indicated by antibody staining directed against the HA tag followed by flow cytometric analysis and immunofluorescence (Supplementary Fig. S1A and S1B). Immunostaining for active FAK (i.e., tyrosine phosphorylation on Tyr397) indicated that HA-FRNK expression strongly decreased PY397FAK levels at focal adhesions (Supplementary Fig. S1C). Importantly, tetFRNK-MTLn3 cells kept their potential for doxycycline-induced HA-FRNK expression for at least 10 passages (data not shown). These data show that we created a doxycycline-regulated HA-FRNK-expressing MTLn3 cell line in which HA-FRNK diminishes FAK phosphorylation at focal adhesions.

Several studies show that inhibition of FAK induces apoptosis. Transient transfection of MTLn3 cells with enhanced green fluorescent protein (GFP)-FRNK indeed induced apoptosis.1

1

van Nimwegen et al., in preparation.

If HA-FRNK itself induces apoptosis in the tetFRNK-MTLn3 cells, this could be mistaken for a reduction in tumor growth or a decrease in metastasis formation in vivo. HA-FRNK slightly reduced cell growth (Fig. 1A). Doxycycline exposure resulted in a 27% growth reduction after 6 days. Immunoblotting of the doxycycline-exposed cells revealed that the expression of HA-FRNK was not altered between day 2 (i.e., steady-state levels of HA-FRNK) and day 6 (Fig. 1B), indicating that HA-FRNK expression does not result in selection of HA-FRNK-negative cells during this time. The reduced cell growth was not caused by an increase in cell death, because no significant difference in the level of apoptosis between the control and the HA-FRNK-expressing cells could be detected in log-phase growing cells (Fig. 1C). When metastases are formed, tumor cells have to survive in the circulation. It is known that constitutively active FAK can rescue cells from anoikis (12). HA-FRNK did not increase anoikis in cells that were kept in suspension for 10 hours (Supplementary Fig. S2). Altogether, these in vitro results show that in our inducible cell line HA-FRNK slightly inhibits proliferation but does not induce apoptosis by itself or promote the onset of anoikis when cells are kept in suspension.

Figure 1.

HA-FRNK inhibits proliferation without affecting apoptosis. MTLn3 and tetFRNK-MTLn3 cells were seeded on day 0. Cells were grown in the absence or presence of doxycycline (doxy) and medium was replaced daily. A, total amount of DNA was quantified daily as described in Materials and Methods. B, expression of FAK, HA-FRNK, and tubulin was determined by immunoblotting using anti-COOH-terminal FAK and anti-tubulin antibodies. C, after 3 days, attached and floating cells were pooled and the percentage of apoptotic cells was measured by flow cytometric cell cycle analysis. Representative of three independent experiments.

Figure 1.

HA-FRNK inhibits proliferation without affecting apoptosis. MTLn3 and tetFRNK-MTLn3 cells were seeded on day 0. Cells were grown in the absence or presence of doxycycline (doxy) and medium was replaced daily. A, total amount of DNA was quantified daily as described in Materials and Methods. B, expression of FAK, HA-FRNK, and tubulin was determined by immunoblotting using anti-COOH-terminal FAK and anti-tubulin antibodies. C, after 3 days, attached and floating cells were pooled and the percentage of apoptotic cells was measured by flow cytometric cell cycle analysis. Representative of three independent experiments.

Close modal

HA-FRNK attenuates attachment and spreading and inhibits motility of cells. Given the role of FAK in focal adhesion formation and turnover, next we investigated the effect of HA-FRNK expression on MTLn3 cell attachment, spreading, and migration. Four hours after replating the tetFRNK-MTLn3 cells, HA-FRNK-expressing cells were poorly attached to the collagen-coated support, whereas the control cells were almost completely spread (Fig. 2A). Quantification of the number of attached cells showed that after 8 hours >80% of the control cells were attached, whereas only 50% of the HA-FRNK-expressing cells were attached (Fig. 2B). At 4 hours after replating, hardly any Tyr397 phosphorylated FAK was present at focal adhesions when HA-FRNK was expressed (Fig. 2A). Western blot analysis of the tetFRNK-MTLn3 cells displayed an ∼50% reduction in overall PY397FAK on HA-FRNK expression (data not shown). To determine the effect of HA-FRNK on cell motility, we studied the migration of tetFRNK-MTLn3 cells using a wound healing assay. To prevent growth of the cells, which could be mistaken for migration, we did the experiments in the absence of serum. HA-FRNK-expressing cells were not able to initiate wound closure within 8 hours. In contrast, in the control cells, 30% of the wound were closed within 8 hours (Fig. 2C and D). Because MTLn3 cells die in the absence of serum, we were not able to study migration after 24 hours under these conditions. However, in the presence of low concentrations of serum, both control and HA-FRNK-expressing cells closed the wound in 24 hours (data not shown).

Figure 2.

HA-FRNK inhibits attachment, spreading, and migration of MTLn3 cells. tetFRNK-MTLn3 cells (incubated for 24 hours in the absence or presence of doxycycline) were trypsinized and kept in suspension for 1 hour. Thereafter, cells were replated on collagen-coated plastic culture dishes. A, at the indicated time points after replating, the number of attached cells was counted at different time points. B, immunofluorescence analysis by confocal laser scanning microscope (CLSM) with anti-PY397FAK, anti-paxillin, and anti-HA antibodies of the attached tetFRNK-MTLn3 cells at 4 hours after replating. Confluent tetFRNK-MTLn3 cell layers (± doxycycline) were wounded as described in Materials and Methods. C, at the indicated time points, phase-contrast pictures of the wounds were taken. D, quantification of cell motility by measuring wound width at three different locations in the wound. Columns, mean of three independent experiments; bars, SE (A and D).

Figure 2.

HA-FRNK inhibits attachment, spreading, and migration of MTLn3 cells. tetFRNK-MTLn3 cells (incubated for 24 hours in the absence or presence of doxycycline) were trypsinized and kept in suspension for 1 hour. Thereafter, cells were replated on collagen-coated plastic culture dishes. A, at the indicated time points after replating, the number of attached cells was counted at different time points. B, immunofluorescence analysis by confocal laser scanning microscope (CLSM) with anti-PY397FAK, anti-paxillin, and anti-HA antibodies of the attached tetFRNK-MTLn3 cells at 4 hours after replating. Confluent tetFRNK-MTLn3 cell layers (± doxycycline) were wounded as described in Materials and Methods. C, at the indicated time points, phase-contrast pictures of the wounds were taken. D, quantification of cell motility by measuring wound width at three different locations in the wound. Columns, mean of three independent experiments; bars, SE (A and D).

Close modal

HA-FRNK reduces primary tumor growth. Next, the effect of HA-FRNK on the in vivo behavior of MTLn3 cells was studied. We first tested the effect of inhibition of FAK function on primary tumor formation. For this purpose, tetFRNK-MTLn3 cells were injected into the fat pad of female Fischer 344 rats and doxycycline was added to the drinking water of the experimental animals. The tumors of the doxycycline-exposed animals showed a significant reduction in volume compared with tumors of animals that did not receive doxycycline. After 30 days, this reduction in volume corresponded to a reduction in tumor weight (Fig. 3A). The expression of HA-FRNK in the tumors of the doxycycline-exposed animals was confirmed by Western blot analysis; control animals did not express HA-FRNK (Fig. 3B). Doxycycline itself did not affect tumor formation or the size of the formed MTLn3 tumors (Supplementary Fig. S3A). To check the effect of HA-FRNK on cell survival in vivo, tumor sections were stained with antibodies directed against active caspase-3. Only few caspase-3-positive cells were observed and the percentage of active caspase-3-positive cells in the HA-FRNK-expressing tumors was comparable with the expression in the control tumors (Fig. 3C). This suggests that the HA-FRNK-induced tumor reduction is most likely due to effects on proliferation rather than apoptosis.

Figure 3.

HA-FRNK decreases primary breast tumor growth without affecting apoptosis. Female Fischer 344 rats were injected with 106 tetFRNK-MTLn3 cells [either pretreated or left untreated with doxycycline in vitro (1 μg/mL)] into the mammary fat pad and tumor growth was followed by measuring tumor size as described in Materials and Methods (A). At day 30, the tumors were isolated and the weight of the tumors was determined (inset). Points/columns, mean of five animals; bars, SE. B, cryosections of primary tumor material of animals that were exposed to doxycycline for 7, 20, or 30 days or left untreated were analyzed by immunoblotting for FAK and HA-FRNK expression using anti-HA and anti-FAK antibodies. C, immunofluorescence analysis of Carnoy's fixed breast tumor sections of control and doxycycline-exposed animals with antibodies directed against HA (green) and active caspase-3 (red). Hoechst 33258 (blue) was used for nuclear staining. White arrows, apoptotic cells. Note that no increase in apoptosis in HA-positive cells is seen.

Figure 3.

HA-FRNK decreases primary breast tumor growth without affecting apoptosis. Female Fischer 344 rats were injected with 106 tetFRNK-MTLn3 cells [either pretreated or left untreated with doxycycline in vitro (1 μg/mL)] into the mammary fat pad and tumor growth was followed by measuring tumor size as described in Materials and Methods (A). At day 30, the tumors were isolated and the weight of the tumors was determined (inset). Points/columns, mean of five animals; bars, SE. B, cryosections of primary tumor material of animals that were exposed to doxycycline for 7, 20, or 30 days or left untreated were analyzed by immunoblotting for FAK and HA-FRNK expression using anti-HA and anti-FAK antibodies. C, immunofluorescence analysis of Carnoy's fixed breast tumor sections of control and doxycycline-exposed animals with antibodies directed against HA (green) and active caspase-3 (red). Hoechst 33258 (blue) was used for nuclear staining. White arrows, apoptotic cells. Note that no increase in apoptosis in HA-positive cells is seen.

Close modal

HA-FRNK inhibits experimental lung metastasis formation. Next, we investigated the effect of HA-FRNK on metastasis formation. tetFRNK-MTLn3 cells were injected into the tail vein of rats to induce experimental lung metastases. To improve the efficiency of in vivo experimental metastasis formation, we depleted NK cells from the Fischer 344 rats, thereby preventing NK cell–mediated killing of circulating MTLn3 cells.1 An advantage of the HA-FRNK-inducible cell line is that HA-FRNK can be expressed after the tumor cells have been injected. This allowed us to discriminate between the effect of inhibition of FAK on attachment, invasion, and secondary tumor formation by exposing three groups of rats to doxycycline at different time points (i.e., days 0, 1, and 11) after injection of the tetFRNK-MTLn3 cells as well as a control group that did not receive doxycycline. Lungs of the animals that were continuously exposed to doxycycline, both in vitro and in vivo, contained 14% of surface tumors compared with the control animals (Fig. 4A). Induction of the expression of HA-FRNK starting in vivo with no doxycycline pretreatment in vitro resulted in a decreased lung tumor metastasis burden of ∼41%. When doxycycline was added to the drinking water from day 11, a nonsignificant decrease in the number of lung metastases was seen: 69% of the control (Fig. 4A). H&E staining of lung tissue showed that the effect of HA-FRNK on the number of surface metastases was a reflection of the metastasis burden inside the lungs (Fig. 4B). In a control experiment, no statistically significant difference in lung tumor metastasis burden of the founder cell line Tet-on-MTLn3 clone 25 was found between animals that were treated or left untreated with doxycycline (Supplementary Fig. S3B). This indicates that the effect observed in tetFRNK-MTLn3 cells is mediated by HA-FRNK. Next, we verified the expression of HA-FRNK in the metastases. Lungs of two of the six animals that received doxycycline both in vitro and in vivo, with the highest number of surface metastases (for median and range, see Table 1), showed HA-FRNK expression as determined by Western blotting of cryosections (Fig. 4C). Animals that were exposed to doxycycline at later time points all expressed HA-FRNK in the lungs (Fig. 4C).

Figure 4.

Inhibition of FAK strongly reduces the number of experimental lung metastases. tetFRNK-MTLn3 cells (105) either pretreated overnight or left untreated in vitro with doxycycline (1 μg/mL) were injected into the tail vein of female Fischer 344 rats. When indicated, drinking water of the animals was supplemented with doxycycline to induce expression of HA-FRNK. A, after 28 days, lungs were injected with ink and the number of surface tumors was counted and is displayed as percentage of the lung tumor burden in the control animals (no doxycycline exposure). B, representative images of lungs of control animals and animals that were continuously exposed to doxycycline. Top, surface lung metastases; bottom, lung sections stained with H&E. C, immunoblotting of lung tissue cryosections for HA-FRNK expression using antibodies directed against HA and COOH-terminal FAK. For each treatment group, different animals were chosen that possessed lung tumor metastases. D, HA-FRNK expression in lung metastasis was verified by CLSM after immunofluorescence staining for FRNK and HA; nuclear counter staining was with Hoechst 33258. Columns, mean; bars, SE.

Figure 4.

Inhibition of FAK strongly reduces the number of experimental lung metastases. tetFRNK-MTLn3 cells (105) either pretreated overnight or left untreated in vitro with doxycycline (1 μg/mL) were injected into the tail vein of female Fischer 344 rats. When indicated, drinking water of the animals was supplemented with doxycycline to induce expression of HA-FRNK. A, after 28 days, lungs were injected with ink and the number of surface tumors was counted and is displayed as percentage of the lung tumor burden in the control animals (no doxycycline exposure). B, representative images of lungs of control animals and animals that were continuously exposed to doxycycline. Top, surface lung metastases; bottom, lung sections stained with H&E. C, immunoblotting of lung tissue cryosections for HA-FRNK expression using antibodies directed against HA and COOH-terminal FAK. For each treatment group, different animals were chosen that possessed lung tumor metastases. D, HA-FRNK expression in lung metastasis was verified by CLSM after immunofluorescence staining for FRNK and HA; nuclear counter staining was with Hoechst 33258. Columns, mean; bars, SE.

Close modal
Table 1.

Effect of HA-FRNK expression on tetFRNK-MTLn3 experimental metastasis formation

GroupIncidence*No. metastases, median (range)SignificanceMetastasis size (size class)Significance§% Range HA-FRNK-positive metastasis (5)
A: No doxycycline 11/11 130 (57-180)  2.7 ± 0.3  
B: Doxycycline in vitro 5/6 17 (0-42) P < 0.0005 0.4 ± 0.2 P < 0.001 50-70 
C: Doxycycline day 0 13/13 63 (3-128) P < 0.05 1.5 ± 0.2 P < 0.01 70-90 
D: Doxycycline day 11 12/13 100 (0-220) NSD 2.5 ± 0.2 NSD 80-90 
GroupIncidence*No. metastases, median (range)SignificanceMetastasis size (size class)Significance§% Range HA-FRNK-positive metastasis (5)
A: No doxycycline 11/11 130 (57-180)  2.7 ± 0.3  
B: Doxycycline in vitro 5/6 17 (0-42) P < 0.0005 0.4 ± 0.2 P < 0.001 50-70 
C: Doxycycline day 0 13/13 63 (3-128) P < 0.05 1.5 ± 0.2 P < 0.01 70-90 
D: Doxycycline day 11 12/13 100 (0-220) NSD 2.5 ± 0.2 NSD 80-90 
*

Number of animals with visible lung surface metastases at day 28/total number of animals in the group.

Statistical significance was determined by the Mann-Whitney rank sum test; NSD, no statistical difference for group D lung tumor burden compared with the lung burden in the control group A. P indicates the confidence level for the statistical difference between group C or B and control group A.

Metastasis size is determined by classification of the size of the metastases in H&E-stained lung sections in four groups: ranging from 0 (very small) to 3 (very large).

§

Statistical significance of the difference in metastasis size was determined by the Student's t test; NSD, no statistical difference in metastasis size between groups D and A. P indicates the confidence level for the statistical difference between group C or B and control group A.

Range of the percentage HA-FRNK-positive cells per metastasis.

HA-FRNK reduces cell growth both in vitro (Fig. 1A) and in vivo in the primary tumor model (Fig. 3A). The average size of the HA-FRNK-positive lung tumor metastases was significantly decreased when the animals were directly exposed to doxycycline regardless of doxycycline pretreatment in vitro. This was not seen when HA-FRNK was expressed 11 days after injection of the cells (Table 1). This further supports the in vitro and primary tumor in vivo observations. Immunofluorescence analysis of the lungs revealed that in all animals exposed to doxycycline HA-FRNK was expressed in the lung metastases (Fig. 4D). Fifty percent to 70% of the lung metastases possessed HA-FRNK-positive cells when doxycycline was provided in vitro and in vivo from day 0 (Table 1). For doxycycline treatment that started only in vivo from day 0, 70% to 90% of the metastases had HA-FRNK-positive cells. This number was 80% to 90% when animals were treated from day 11 with doxycycline (Table 1). No evident changes in the structure of the tumors were observed between all the different groups.

Inhibition of lung metastasis formation by HA-FRNK is not caused by tumor cell dormancy. Inhibition of FAK is known to induce dormancy, which can be overcome by activating the downstream extracellular signal-regulated kinase (ERK)/mitogen-activated protein kinase-ERK kinase pathway (14). To determine whether HA-FRNK expression resulted in dormant MTLn3 cells that are able to grow out into metastases when FAK is no longer inhibited, we did another experimental metastasis formation experiment. Again, tetFRNK-MTLn3 cells (either preincubated with doxycycline in vitro or left untreated) were injected into the tail vein of the female Fischer 344 rats. One group was exposed to doxycycline for the entire duration of the experiment, whereas the other group was kept on doxycycline-containing drinking water only for the first 5 days after tumor cell injection. The animals that were continuously exposed to doxycycline had only 5% of the tumor burden compared with the control animals (Fig. 5A). In the lungs where HA-FRNK was only expressed for the first 5 days of the experiment, the tumor burden was 13%. Immunohistochemical analysis of the lungs showed HA-FRNK-positive tumor cells in animals continuously exposed to doxycycline, whereas no HA-FRNK could be detected in either control animals or animals that were exposed to doxycycline during the first 5 days of the experiment (Fig. 5B).

Figure 5.

Inhibition of tumor metastasis formation by HA-FRNK is not related to MTLn3 cell dormancy. tetFRNK cells were pretreated in vitro or left untreated followed by injection in the tail vein. Rats were either continuously treated with doxycycline in the drinking water or only for the first 5 days. Control animals did not receive doxycycline. A, the number of surface tumors was counted after 28 days and is displayed as percentage of the lung tumors in the control animals (no doxycycline). Columns, mean (n = 8); bars, SE. B, to verify HA-FRNK expression, tissue sections were immunostained with anti-COOH-terminal FAK and anti-HA followed by CLSM. Representative of the different treatment groups. Note that the metastases in the animals that were exposed to doxycycline from days −1 to 4 do no longer express HA-FRNK.

Figure 5.

Inhibition of tumor metastasis formation by HA-FRNK is not related to MTLn3 cell dormancy. tetFRNK cells were pretreated in vitro or left untreated followed by injection in the tail vein. Rats were either continuously treated with doxycycline in the drinking water or only for the first 5 days. Control animals did not receive doxycycline. A, the number of surface tumors was counted after 28 days and is displayed as percentage of the lung tumors in the control animals (no doxycycline). Columns, mean (n = 8); bars, SE. B, to verify HA-FRNK expression, tissue sections were immunostained with anti-COOH-terminal FAK and anti-HA followed by CLSM. Representative of the different treatment groups. Note that the metastases in the animals that were exposed to doxycycline from days −1 to 4 do no longer express HA-FRNK.

Close modal

In this study, we inhibited FAK function in breast tumor cells by conditional expression of a dominant-negative acting splice variant of FAK, FRNK. Using a syngeneic rat MTLn3 adenocarcinoma tumor/metastasis model, we obtained results that indicate that (a) FAK is important in primary breast tumor growth, (b) FAK is essential in the breast tumor lung metastasis formation, (c) FAK-mediated control of metastasis occurs in the initial phase of the lung metastasis process, and (d) inhibition of FAK does not cause dormancy of tumor cells. These effects of FAK are not dependent on the control of cell survival; FAK rather interferes with MTLn3 breast tumor cell migration and invasion. Our observations suggest that FAK is a potential target for therapeutic intervention in the attachment and invasion steps of metastasis formation.

We found that FAK is essential in the early phase of metastasis formation, involving invasion and migration processes, leading to colonization of the lung by MTLn3 breast tumor cells: inhibition of FAK during the first 5 days of metastasis formation dramatically reduced the number of experimental lung metastasis (Figs. 4 and 5). The importance of FAK in the control of adhesion and migration of MTLn3 cells was supported by our in vitro observations, because expression of HA-FRNK reduced attachment, spreading, and migration of the tumor cells. These observations agree with the effect of FRNK in various cell types in vitro (11, 2426). Localization of FRNK at the focal adhesions seems crucial in the inhibition of in vivo metastasis formation. Wild-type FRNK reduced invasion and metastasis formation of v-Src-transformed 3T3 cells through a mechanism that is independent of effects on cell motility. The FRNK Leu1034Ser mutant, which is impaired in focal adhesion localization and therefore cannot compete with FAK at the focal adhesions, did not affect the metastatic phenotype of v-Src 3T3 cells (15). In the MTLn3 cells, HA-FRNK inhibits FAK localization at focal adhesions; prevention of this localization likely impaired the turnover of focal adhesions. In contrast to the v-Src 3T3 cells, in which FRNK does not reduce motility, HA-FRNK-expressing MTLn3 cells showed a reduced cell spreading and attenuated focal adhesion formation as well as a reduced cell migration. This distinct effect of FRNK on motility in the v-Src 3T3 versus MTLn3 cells is probably due to the v-Src transformation; v-Src can phosphorylate other targets (i.e., cdc42, p190Rho-GAP, Vav1/2, and ezrin) that (in)directly regulate actin cytoskeletal organization and migration (27). The role of FRNK in impairment, but not complete prevention, of FAK-mediated focal adhesion formation/turnover in our cells explains why MTLn3 experimental metastasis formation was not fully inhibited. We found that despite HA-FRNK expression at the time of the tail vein injection of the tumor cells, after 28 days still some HA-FRNK-positive tetFRNK-MTLn3 colonies could be found in a subset of metastases-bearing animals (Table 1). This indicates that also under in vivo conditions HA-FRNK inhibits but does not completely block focal adhesion turnover, a requirement for cell migration. Alternatively, given the fact that the levels of HA-FRNK expression after doxycycline exposure vary within the population of tetFRNK-MTLn3 cells (Supplementary Fig. S1B), it is possible that the formed HA-FRNK-positive metastases were derived from cells that initially expressed low levels of HA-FRNK and, as a consequence, had a higher change of invading the lungs. Based on our data, we think that FAK is essential in the metastasis process. Because we used experimental metastasis as a model, we can only be certain that FAK is important in the later steps of the entire metastasis process: extravasation and tissue invasion. Because HA-FRNK inhibited tumor growth in the mammary fat pad, we were unable to study a role for FAK in the early steps of metastasis process: detachment of the primary tumor and intravasation into the bloodstream.

Although our combined in vitro and in vivo data support a role for FAK in the migration and invasion processes involved in metastasis formation, the exact downstream substrates that mediate these effects remain to be identified. FAK interacts with c-Src to form a FAK/Src tyrosine kinase complex to phosphorylate downstream signaling (adapter) proteins, including paxillin and p130CAS. Indeed, HA-FRNK expression inhibited the early tyrosine phosphorylation of paxillin (data not shown). Alternatively, FAK can function as a docking protein to target other kinases (or kinase linkers) to the focal adhesions. In v-Src-transformed 3T3 cells, FRNK inhibited c-Jun NH2-terminal kinase (JNK) activation in association with a reduced expression of matrix metalloproteinase-2; pharmacologic inhibition of JNK with SP600125 decreased the invasive phenotype of v-Src 3T3 cells (15). In our tetFRNK-MTLn3 cells, no overall difference in the phosphorylation of JNK could be detected in the presence or absence of HA-FRNK (data not shown). However, because FAK binds to, for instance, mitogen-activated protein kinase-ERK kinase kinase 1 and can thereby target other kinases, including JNK, to focal adhesions, loss of FAK at focal adhesions may indirectly affect the phosphorylation of proteins that are crucial in focal adhesion turnover. In this respect, we should note that paxillin can be phosphorylated on Ser178 by JNK; this is required for the turnover of focal adhesions (28). MTLn3 cells that stably overexpress S178A mutant paxillin show impaired migration in vitro.2

2

Verkoeijen and Van de Water, in preparation.

Possibly, the decreased cell motility caused by inhibition of FAK localization at focal adhesions by HA-FRNK is related to a decreased FAK-mediated targeting of JNK to the focal adhesions with reduced paxillin Ser178 phosphorylation and focal adhesion turnover as a consequence. By using genome-wide gene expression analysis of HA-FRNK-MTLn3 cells, we are currently also trying to identify downstream targets that are involved in the HA-FRNK-induced inhibition of experimental metastasis formation.

Adhesion of cells to the ECM is an important cell survival signaling route, which is mediated in part by the activity of FAK. Various studies showed that inhibition of FAK, by expression of either FRNK or focal adhesion targeting or by down-regulation of FAK levels using small interfering RNA, resulted in an increase in apoptosis of both attached and suspended cells, including breast tumor cells (2932). Transient transfection of MTLn3 cells with GFP-FRNK resulted in an increase in the percentage of apoptotic cells.1 These apoptosis-inducing effects are probably the result of high levels of FRNK: increasing expression levels of GFP-FRNK resulted in increasing percentages of apoptotic cells, whereas cells with relatively low levels of GFP-FRNK were not susceptible to apoptosis. In our tetFRNK-MTLn3 cells, the expression level of HA-FRNK is lower than the expression level of transient transfected GFP-FRNK; therefore, HA-FRNK was not able to induce apoptosis. This allowed us to study more subtle biological effects that are rather related to focal adhesion organization and turnover than to interference with cell survival signaling. Indeed, no effects of HA-FRNK expression were observed on the levels of apoptosis in vitro or in the primary tumor and lung metastasis in vivo. In vivo, neither prolonged expression of HA-FRNK (Fig. 3B) nor expression of HA-FRNK for only 3 days resulted in an increase in apoptosis of HA-FRNK-positive cells (data not shown). Because low levels of HA-FRNK affect migration, whereas high levels affect cell survival, a divergence in the signaling mechanism involved in both processes is likely. Therefore, although not all breast tumor cell lines are sensitive toward FRNK-mediated apoptosis (29), inhibition of FAK-mediated focal adhesion turnover may be a general strategy to prevent the attachment and invasion steps involved in the formation of metastases irrespective of apoptosis susceptibility.

In conclusion, the data indicate that proper FAK signaling at the focal adhesions is essential in the early steps of metastasis formation. Therefore, FAK should be recognized as an important therapeutic target for the development of anticancer drugs. Such a pharmacologic modulation of FAK function would be sufficient at the level of the focal adhesions (i.e., by preventing FAK localization at the focal adhesions). Preferably, this would be mediated by small molecules that prevent FAK from binding to its partners, such as paxillin and talin. A selective inhibition of FAK localization would likely result in a reduced number of unwanted side effects compared with the alternative approach: inhibition of overall FAK activity by blocking kinase activity.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Grant support: Dutch Cancer Society grants 2001/2477 and 2002/2741 and Royal Academy for Arts and Sciences fellowship (B. van de Water).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Peter Kuppen for providing the mAb3.2.3 antibody, Dymphy Huntjens and Chantal Pont for assistance during the in vivo experiments, and Dr. Jurjen Velthuis for assistance, helpful suggestions, and discussions.

1
Oktay MH, Oktay K, Hamele-Bena D, Buyuk A, Koss LG. Focal adhesion kinase as a marker of malignant phenotype in breast and cervical carcinomas.
Hum Pathol
2003
;
34
:
240
–5.
2
Kornberg LJ. Focal adhesion kinase and its potential involvement in tumor invasion and metastasis.
Head Neck
1998
;
20
:
745
–52.
3
Sood AK, Coffin JE, Schneider GB, et al. Biological significance of focal adhesion kinase in ovarian cancer: role in migration and invasion.
Am J Pathol
2004
;
165
:
1087
–95.
4
Agochiya M, Brunton VG, Owens DW, et al. Increased dosage and amplification of the focal adhesion kinase gene in human cancer cells.
Oncogene
1999
;
18
:
5646
–53.
5
McLean GW, Brown K, Arbuckle MI, et al. Decreased focal adhesion kinase suppresses papilloma formation during experimental mouse skin carcinogenesis.
Cancer Res
2001
;
61
:
8385
–9.
6
Parsons JT. Focal adhesion kinase: the first ten years.
J Cell Sci
2003
;
116
:
1409
–16.
7
Schlaepfer DD, Hauck CR, Sieg DJ. Signaling through focal adhesion kinase.
Prog Biophys Mol Biol
1999
;
71
:
435
–78.
8
Schaller MD, Hildebrand JD, Shannon JD, Fox JW, Vines RR, Parsons JT. Autophosphorylation of the focal adhesion kinase, pp125FAK, directs SH2-dependent binding of pp60src.
Mol Cell Biol
1994
;
14
:
1680
–8.
9
Thomas JW, Cooley MA, Broome JM, et al. The role of focal adhesion kinase binding in the regulation of tyrosine phosphorylation of paxillin.
J Biol Chem
1999
;
274
:
36684
–92.
10
Tachibana K, Urano T, Fujita H, et al. Tyrosine phosphorylation of Crk-associated substrates by focal adhesion kinase. A putative mechanism for the integrin-mediated tyrosine phosphorylation of Crk-associated substrates.
J Biol Chem
1997
;
272
:
29083
–90.
11
Sieg DJ, Hauck CR, Schlaepfer DD. Required role of focal adhesion kinase (FAK) for integrin-stimulated cell migration.
J Cell Sci
1999
;
112
:
2677
–91.
12
Frisch SM, Vuori K, Ruoslahti E, Chan-Hui PY. Control of adhesion-dependent cell survival by focal adhesion kinase.
J Cell Biol
1996
;
134
:
793
–9.
13
Wang D, Grammer JR, Cobbs CS, et al. p125 focal adhesion kinase promotes malignant astrocytoma cell proliferation in vivo.
J Cell Sci
2000
;
113
Pt 23:
4221
–30.
14
Aguirre Ghiso JA. Inhibition of FAK signaling activated by urokinase receptor induces dormancy in human carcinoma cells in vivo.
Oncogene
2002
;
21
:
2513
–24.
15
Hauck CR, Hsia DA, Puente XS, Cheresh DA, Schlaepfer DD. FRNK blocks v-Src-stimulated invasion and experimental metastases without effects on cell motility or growth.
EMBO J
2002
;
21
:
6289
–302.
16
Abdel-Ghany M, Cheng HC, Elble RC, Pauli BU. Focal adhesion kinase activated by β4 integrin ligation to mCLCA1 mediates early metastatic growth.
J Biol Chem
2002
;
277
:
34391
–400.
17
Owen JD, Ruest PJ, Fry DW, Hanks SK. Induced focal adhesion kinase (FAK) expression in FAK-null cells enhances cell spreading and migration requiring both auto- and activation loop phosphorylation sites and inhibits adhesion-dependent tyrosine phosphorylation of Pyk2.
Mol Cell Biol
1999
;
19
:
4806
–18.
18
Ilic D, Kovacic B, Johkura K, et al. FAK promotes organization of fibronectin matrix and fibrillar adhesions.
J Cell Sci
2004
;
117
:
177
–87.
19
Zhao J, Bian ZC, Yee K, Chen BP, Chien S, Guan JL. Identification of transcription factor KLF8 as a downstream target of focal adhesion kinase in its regulation of cyclin D1 and cell cycle progression.
Mol Cell
2003
;
11
:
1503
–15.
20
Oktay M, Wary KK, Dans M, Birge RB, Giancotti FG. Integrin-mediated activation of focal adhesion kinase is required for signaling to Jun NH2-terminal kinase and progression through the G1 phase of the cell cycle.
J Cell Biol
1999
;
145
:
1461
–9.
21
Neri A, Welch D, Kawaguchi T, Nicolson GL. Development and biologic properties of malignant cell sublines and clones of a spontaneously metastasizing rat mammary adenocarcinoma.
J Natl Cancer Inst
1982
;
68
:
507
–17.
22
Kiley SC, Clark KJ, Goodnough M, Welch DR, Jaken S. Protein kinase Cδ involvement in mammary tumor cell metastasis.
Cancer Res
1999
;
59
:
3230
–8.
23
Huigsloot M, Tijdens IB, Mulder GJ, van de Water B. Differential regulation of doxorubicin-induced mitochondrial dysfunction and apoptosis by Bcl-2 in mammary adenocarcinoma (MTLn3) cells.
J Biol Chem
2002
;
277
:
35869
–79.
24
Taylor JM, Mack CP, Nolan K, Regan CP, Owens GK, Parsons JT. Selective expression of an endogenous inhibitor of FAK regulates proliferation and migration of vascular smooth muscle cells.
Mol Cell Biol
2001
;
21
:
1565
–72.
25
Richardson A, Malik RK, Hildebrand JD, Parsons JT. Inhibition of cell spreading by expression of the C-terminal domain of focal adhesion kinase (FAK) is rescued by coexpression of Src or catalytically inactive FAK: a role for paxillin tyrosine phosphorylation.
Mol Cell Biol
1997
;
17
:
6906
–14.
26
Sieg DJ, Hauck CR, Ilic D, et al. FAK integrates growth-factor and integrin signals to promote cell migration.
Nat Cell Biol
2000
;
2
:
249
–56.
27
Frame MC. Newest findings on the oldest oncogene; how activated src does it.
J Cell Sci
2004
;
117
:
989
–98.
28
Huang C, Rajfur Z, Borchers C, Schaller MD, Jacobson K. JNK phosphorylates paxillin and regulates cell migration.
Nature
2003
;
424
:
219
–23.
29
Xu LH, Yang X, Bradham CA, et al. The focal adhesion kinase suppresses transformation-associated, anchorage-independent apoptosis in human breast cancer cells. Involvement of death receptor-related signaling pathways.
J Biol Chem
2000
;
275
:
30597
–604.
30
Zhang Y, Lu H, Dazin P, Kapila YL. Squamous cell carcinoma cell aggregates escape suspension-induced, p53-mediated anoikis: fibronectin and integrin (α)v mediate survival signals through focal adhesion kinase.
J Biol Chem
2004
;
279
:
48342
–9.
31
Subauste MC, Pertz O, Adamson ED, Turner CE, Junger S, Hahn KM. Vinculin modulation of paxillin-FAK interactions regulates ERK to control survival and motility.
J Cell Biol
2004
;
165
:
371
–81.
32
Heidkamp MC, Bayer AL, Kalina JA, Eble DM, Samarel AM. GFP-FRNK disrupts focal adhesions and induces anoikis in neonatal rat ventricular myocytes.
Circ Res
2002
;
90
:
1282
–9.

Supplementary data