CD98 activation increases surface expression and clustering of β₁ integrins in MCF-7 cells through FAK/Src- and cytoskeleton-independent mechanisms

Abstract

CD98, a disulfide-linked 125-kDa heterodimeric type II transmembrane glycoprotein, regulates β₁ integrin-mediated cell adhesion. However, the molecular mechanisms underlying CD98-mediated activation of β₁ integrin are presently unclear. In this study, the effects of CD98 signaling on the expression and clustering of β₁ integrin were investigated. Activation of CD98 augmented surface expression of β₁ integrin on MCF-7 cells. Cross-linking CD98 induced clustering of β₁ integrins. Inhibition of phosphorylation of focal adhesion kimase (FAK) by PP2, an inhibitor of Src family kinase, reduced cell-extracellular matrix adhesion, but not surface expression and clustering of β₁ integrin on MCF-7 cells. This result was confirmed by over-expression of dominant negative forms of FAK. In addition, phalloidin or cytochalasin D inhibited CD98-mediated induction of cell-ECM adhesion, but not surface expression and clustering of β₁ integrins. The inhibitory effects of PP2, cytochalasin D or phalloidin on CD98-stimulated cell adhesion were diminished by pretreatment of cells with Mn²⁺, which is shown to induce conformational change of integrins. These results provide the first evidence that CD98 activation increases not only β₁ integrin affinity but also its surface expression and clustering and the latter is independent of FAK/Src and cytoskeleton.

Introduction

Regulation of integrin activity plays a key role in not only normal cellular processes such as cell migration and growth but also pathological processes including metastasis of cancer cells. The α and β chains of integrin heterodimers are each composed of an extracellular region, a transmembrane segment and a small cytoplasmic domain. Integrins form complexes with other transmembrane proteins that can modulate their functions by inducing intracellular signaling. Several receptor molecules such as T cell receptor (TCR), chemokine receptors and CD99 have been shown previously to generate "inside-out signaling" that provokes changes in the conformation of integrin extracellular domain or clustering of integrins (Hahn et al., 1997; Burbach et al., 2007). In addition, CD47 and the tetraspan family (TM4SF) of receptors, CD9, CD63, CD81, CD151 have been shown to associate with integrins and affect integrin functions (Lindberg et al., 1996; Berditchevski, 2001).

Previous studies indicate that CD98 is an important regulator of integrin-mediated cell adhesion (Fenczik et al., 2001; Rintoul et al., 2002). The human CD98 protein is a disulfide-linked 125-kDa heterodimeric type II transmembrane glycoprotein composed of a glycosylated 85-kDa heavy chain (designated CD98) and a nonglycosylated 40-kDa light chain. CD98 is broadly expressed on many cell types including monocytes, thymocytes, proximal convoluted tubules in kidney, epithelial cells in esophagus and breast, sarcolemma in skeletal muscle and melanoma cells (Bellone et al., 1989; Dixon et al., 1990). CD98 functions include regulation of β₁ integrin function, as well as amino acid transport, cell fusion, and proliferation (Deves and Boyd, 2000). CD98 activation increased β₁ integrin-mediated adhesion of the small cell lung-cancer cell line (SCLC) H345 and breast cancer cells to fibronectin and laminin by inducing the conformational changes of β₁ integrins (Fenczik et al., 1997; Chandrasekaran et al., 1999). β₁ integrin-blocking antibodies inhibited CD98 signaling in human peripheral T lymphocytes, which may reflect convergent signaling mechanisms between β₁ integrin and CD98 (Warren et al., 2000). This convergent signaling mechanism between integrins and CD98 is supported by recent findings that CD98 is associated with integrins (Zent et al., 2000; Fenczik et al., 2001), and that cross-linking CD98 induced the clustering of β₁-integrins and integrin-like signaling (Rintoul et al., 2002). Taken together, these results suggest that CD98 regulates β₁ integrin functions. However, its underlying molecular mechanisms remain elusive.

In this study, we investigated whether CD98 regulates the function of β₁ integrin by increasing the total number of β₁ integrin on the surface, or forming higher order multimers or clusters and whether the effect of CD98 stimulation on cell adhesion is mediated by FAK phosphorylation and cytoskeleton reorganization. The activation of CD98 molecules resulted in augmentation of cell adhesion and surface levels or clustering of β₁ integrins on MCF-7 cells. Inhibition of phosphorylation of FAK or cytoskeleton reorganization diminished CD98-induced adhesion of MCF-7 cells to fibronectin, but not CD98-induced surface expression and clustering of β₁ integrins.

Materials and Methods

Reagents

Culture reagents and Lipofectamine were purchased from Invitrogen (Carlsbad, CA). Mouse anti-human CD98 mAb (UM7F8) was purchased from BD Biosciences (San Jose, CA). Horseradish-conjugated goat anti-mouse IgG and purified goat anti-mouse IgG were from DiNona (Seoul, Korea). Anti-FAK rabbit polyclonal antibody (C-20) and anti-phophotyrosine mAb (PY99) were from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Mouse anti-CD29 (β₁ integrin) mAb (3S3) and FITC-conjugated mouse anti-β₁ integrin mAb (S3S) was from Serotec Inc. (Raleigh, NC). R-PE-conjugated mouse anti-β₁ integrin mAb (TDM29) was from Cymbus Biotechnology (Eastleigh, Hampshire, UK). Cytochalasin D and phalloidin were from Tocris bioscience (Avonmouth, UK). Mouse IgG, fibronectin and phalliodin-FITC were purchased from Sigma-Aldrich (St. Louis, MO). PP2 was from Biomol Research Laboratories Inc. (Plymouth Meeting, PA).

Cell culture

MCF-7 human breast adenocarcinoma cell line was obtained from Korean Cell Line Bank (Seoul, Korea) and cultured in RPMI1640 medium supplemented with 10% FBS, 100 U/ml penicillin, 10 µg/ml streptomycin and 100 mM sodium pyrubate, 5 mg/ml insulin. Cell cultures were maintained and incubated in 5% CO₂ at 37℃.

Flow cytometric analysis

Flow cytometric analysis used for detecting β₁ integrin expression was carried out as described previously (Choi et al., 2007).

Stable transfection

MCF-7 cells grown to 70-80% confluence were transfected with the pcDNA3 vector (Invitrogen) encoding FAK Y397F mutant or FAK related non-kinase (FRNK) by using Lipofectamine (Invitrogen) according to the manufacturer's instructions. The expression vector pcDNA3-FAK Y397F and pCDNA3-FRNK were kindly provided by Dr. Soo-Chul Park (Sookmyung Women's University). pcDNA3 vector was also transfected as a control. Neomycin-resistant clones were isolated by growth in DMEM containing 10% FBS and 0.8 mg/ml G418 (Invitrogen). Stable transfectant clones were selected by immunoblotting.

Adhesion assay

Nonspecific binding sites were blocked with 1% BSA in PBS for 1 h at room temperature. 2 µg of fibronectin in PBS was coated on 96-well microtiter plates overnight at 4℃. 5×10⁴ MCF-7 cells in the 100 µl of serum free medium were allowed to adhere to fibronectin-coated 96-well plates, after incubation with mouse IgG (3 µg) and goat anti-mouse IgG (30 µg), or anti-CD98 mAb (3 µg) and goat anti-mouse IgG (30 µg) for 1 h. After washing with PBS three times, cell attachment was determined by counting the number of attached cells. The effects of PP2, cytochalasin D, phalloidin or over-expression of dominant-negative FAK mutants on CD98-induced cell-ECM adhesion were examined in the presence or absence of 0.5 µM MnCl₂ as described above.

Immunoprecipitation and Western blot analysis

MCF-7 cells were lysed in protein lysis buffer (50 mM Hepes pH 7.5, 150 mM NaCl, 1.5 mM MgCl₂, 1% Triton X-100, 10 µg/ml aprotinin, 100 mM sodium orthovanadate, 1 mM PMSF). FAK was precipitated from the cell lysates with rabbit anti-FAK polyclonal antibody overnight at 4℃. Protein-antibody complexes were incubated with either protein A agarose beads for 3 h at 4℃. The precipitates were washed with Tris-buffered saline/Tween (20 mM Tris-HCl pH 7.4, 150 mM NaCl, and 0.05% [vol/vol] Tween 20) and boiled in SDS sample buffer for 10 min before loading on 10% SDS-polyacrylamide gel electrophoresis. The proteins were transferred to nitrocellulose membranes, blocked using 3% (wt/vol) albumin in Tris-buffered saline/Tween overnight at 4℃ and then incubated with anti-FAK or anti-phospho-tyrosine antibody. HRP-conjugated antibodies were used for secondary labeling. Immunoreactive bands were identified using Westzol (Intron, Seoul, Korea) according to the manufacturer's instruction.

Determination of integrin clustering

5×10⁵ MCF-7 cells were plated and grown onto glass cover-slips coated with fibronectin (20 µg/ml) for 24 h. After washing with SFM twice, cells were incubated with mouse IgG (30 µg/ml), or anti-CD98 mAb UM7F8 (30 µg/ml), or anti-β₁ integrin mAb 3S3 (30 µg/ml) for 1 h at room temperature. When cross-linking is necessary, goat anti-mouse IgG (300 µg/ml) was added as secondary antibody. After washing twice with PBS, cells were fixed with 2% paraformaldehyde in PBS for 30 min at room temperature. After washing with PBS, they were permeabilized in 0.1% Triton x-100. Permeabilized cells were incubated with 1 µg of R-PE-conjugated mouse anti-β₁ integrin mAb TDM29 and phalloidin-FITC for 1 h at room temperature. Finally, samples were washed, mounted onto slides, embedded with mounting medium (DakoCytomation, Carpinteria, CA), and analyzed under a confocal microscope Zeiss LSM510 META NLO (Carl Zeiss, Bernried, Germany). Images were processed using the Adobe Photoshop software. Signal intensities were measured by using LSM 510 Meta NLO software (Carl Zeiss, Bernried, Germany).

Statistical analysis

The data are expressed as the average of the mean values obtained ± SE. Statistical significance was determined as described previously (Lee et al., 2006).

Results

Cross-linking CD98 elevates the surface expression of β₁ integrin

Integrin affinity for ECM ligands was up-regulated by the activated CD98 (Fenczik et al., 1997; Chandrasekaran et al., 1999). Enhanced cell adhesion could result from the elevated expression of cellular adhesion molecules. For this reason, we examined the effect of CD98 activation on expression of β₁ integrin on MCF-7 cells. Flow cytometry analysis showed cell surface expression of β₁ integrin significantly increased in MCF-7 cells stimulated by anti-CD98 mAb (Figure 1A). However, Western blot analysis showed that CD98 activation did not have any effect on the total expression of β₁ integrin in MCF-7 cells, suggesting that CD98 activation may be involved in exocytosis rather than synthesis of β₁ integrin (Figure 1B).

Figure 1.

CD98 engagement increases the surface expression of β₁ integrin. (A) Freshly cultured MCF-7 cells were dispersed and then incubated with anti-CD98 mAb UM7F8 and secondary Ab (blue line) for 1 h. As negative controls, cells were treated with mouse IgG and secondary Ab (red line). FITC-conjugated anti-CD29 mAb was used to measure the expression level of β₁ integrin on the cells treated with mAbs as described above. An FITC-conjugated mouse anti-human lgG was used as the negative control (green line). (B) Cell extracts from MCF-7 cells treated as in (A) were analyzed by Western blotting using anti-β₁ integrin and anti-actin mAb.

Cross-linking of CD98 induces clustering of β₁ integrin

Previous studies demonstrated that cross-linking CD98 induced clustering of β₁ integrins (Kolesnikova et al., 2001). To determine the underlying mechanisms by which CD98 cross-linking induces clustering of β₁ integrins on the cell surface, we compared the cell surface distribution of β₁ integrins following CD98 or β₁ integrin engagement on MCF-7 cells. Confocal microscopy analysis revealed that cross-linking with anti-CD98 mAb induced the formation of multiple small patches of β₁ integrin on the cell surface much more than that with anti-β₁ integrin mAb (Figure 2A and B). Clustering of β₁ integrins was strictly dependent on cross-linking, because without a secondary Ab, β₁ integrin clustering was not observed on the cell surface. Cross-linking with mouse IgG and secondary Ab had no effect on clustering of β₁ integrins, supporting the specificity of the phenomenon further (data not shown). These observations suggest that cross-linking CD98 facilitates β₁ integrin clustering on the surface of MCF-7 cells by triggering signals to enhance clustering actively.

Figure 2.

CD98 cross-linking enhances β₁ integrin clustering. (A) MCF-7 cells were incubated with anti-CD98 mAb UM7F8 with or without secondary antibody. To determine whether cross-linking of β₁ integrins induces their clustering, cells were treated with anti-β₁ integrin mAb 3S3 in the presence of secondary antibody as well. Next, cells treated as described in "Materials and Methods" were analyzed by confocal microscopy. Actin cytoskeleton organization is visualized by staining with phalloidin-FITC whereas β₁ integrin clustering with R-PE-conjugated anti-β₁ integrin mAb. Images are from a single experiment representative of more than three so performed. Scale bar, 50 µm. Original magnification, × 400. (B) Relative intensities of β₁ integrin clusters in Figure 2A were measured with LSM5120 Meta NLO software. Results are values relative to the β₁ integrin signal intensity level of untreated controls, designated as 1.

Pharmacologic inhibition of FAK/Src kinase activity blocks CD98-induced cell adhesion

Previous study showed overexpression or cross-linking of CD98 led to an increase in FAK phosphorylation (Rintoul et al., 2002; Cai et al., 2005). We determined whether FAK phosphorylation is required for CD98-induced adhesion of MCF-7 cells to ECM. FAK phosphorylation in MCF-7 cell cultures after CD98 engagement was examined by immunoprecipitation and immunoblotting. Figure 3A showed crosslinking CD98 triggered FAK auto-phosphorylation in suspended MCF-7 cells before adhesion to matrix. This elevated phosphorylation of FAK was significantly reduced by the pretreatment with PP2 (0.2 µM), an inhibitor of Src family kinase (Salazar and Rozengurt, 2001), with no inhibitory effect due to the PP2 solvent DMSO (Figure 3A, lanes 2 and 3). To investigate effects of FAK phosphorylation on the CD98-mediated cell adhesion, we treated MCF-7 cells with anti-CD98 mAb in the presence or absence of PP2 (0.2 µM) and then the effect of PP2 on the adhesion of MCF-7 cells to fibronectin was examined. Figure 3B shows inhibition of CD98-induced cell adhesion to fibronectin by PP2. These results suggest that FAK/Src kinase activity may be required for CD98-stimulated cell adhesion. In contrast, inhibition of FAK phosphorylation with pretreatment of cells with PP2 (0.2 µM) has no effect on surface expression and clustering of β₁ integrins, suggesting that CD98 activation upregulates surface expression and clustering of β₁ integrin through FAK-independent pathway (Figure 3C and D).

Figure 3.

Inhibition of FAK/Src kinases with PP2 blocks CD98-induced cell adhesion, but not surface expression and clustering of β₁ integrins in MCF-7 cells. (A) The effect of PP2 and/or Mn²⁺ on FAK phosphorylation in MCF-7 cells treated with anti-CD98 mAb was determined by immunoprecipitaion assay. MCF-7 cells were incubated with anti-CD98 mAb and PP2 (0.2 µM) or a DMSO vehicle control in the presence or absence of 0.5 µM Mn²⁺ for 1 h and anti-FAK rabbit polyclonal antibody (C-20) was used to immunoprecipitate FAK from extracts of MCF-7 cells. Immunoprecipitates were blotted and probed with anti-phosphotyrosine mAb (clone PY99) and anti-FAK mAb. (B) The effect of PP2 treatment on cell adhesion rate was determined as described in Materials and Methods. In addition, whether addition of 0.5 µM Mn²⁺ interferes with the effect of PP2 on cell adhesion was determined by the same way. Results are expressed as mean ± SE of values relative to the adhesion rate of mouse IgG-treated controls, designated as 100%. Asterisks show a significant difference from control as follows: ^*P < 0.05. Additional statistical comparisons are indicated by lines. (C) MCF-7 cells were incubated with anti-CD98 mAb with or without PP2 (0.2 µM) and then analyzed by flow cytometry using FITC-conjugated anti-human β₁ integrin mAb. Data represent the mean ± SE of values relative to mean values of fluorescence intensity of mouse IgG-treated controls, designated as 100%. Asterisks show a significant difference from control as follows: ^*P < 0.05, ^***P < 0.001 (D) Confocal microscopy was performed as described in Figure 2 legend to investigate the effect of PP2 (0.2 µM) on CD98-induced clustering of β₁ integrins. Scale bar, 50 µm. Original magnification, × 400.

Next, we examined whether conformational change of β₁ integrin with MnCl₂ (0.5 µM) prevents the effect of PP2 on CD98-stimulated cell adhesion. Addition of MnCl₂ diminished the effect of PP2 on the CD98-induced cell adhesion (Figure 3B). However, conformational activation of β₁ integrin by Mn²⁺ is not associated with FAK (Figure 3A, bottom). These data are consistent with the previous report showing that addition of Mn²⁺ did not affect FAK or Src phosphorylation, but increased cell-ECM adhesion (Thamilselvan and Basson, 2004).

Over-expression of dominant-negative mutant FAK diminishes the effect of CD98 activation on cell adhesion

Because protein kinase inhibitors have the potential to target protein kinases nonspecifically, we confirmed the effect of FAK on the CD98-mediated cell adhesion by overexpression of dominant-negative forms of FAK in MCF-7 cells. We compared the effect of CD98 activation on adhesion of mock transfectants and cells transfected with dominant-negative forms of FAK that appeared to decrease integrin affinity (Thamilselvan and Basson, 2004). The cell lines that stably overexpressed the kinase-deficient FAK (FRNK or Y397F-FAK) were identified by Western blot analysis (Figure 4A). Transfected cells were treated with anti-CD98 mAb for 1 h, and then replated on fibronectin. As shown in Figure 4B, FRNK and Y397F-FAK-transfected cells did not increase adhesion in response to CD98 stimulus that increased adhesion of the untransfected cells as well as that of mock-transfected cells. Addition of Mn²⁺ to the dominant-negative mutant FAK clones restored cell adhesion. These results are consistent with the above results indicating that CD98 signaling increases cell adhesion through phosphorylation of FAK. In contrast, high-level expression of the FAK mutants did not have any effect on CD98-induced surface expression or clustering of β₁ integrins (Figure 4C and D). As a whole, these experiments indicate that FAK phosphorylation is required for CD98-induced adhesion of MCF-7 cells to extracellular matrix, but not for surface expression and clustering of β₁ integrins.

Figure 4.

The effects of dominant-negative variants of FAK on adhesiveness of MCF-7 cells, surface expression and clustering of β₁ integrins. (A) MCF-7 cells were stably transfected with dominant-negative mutant FAK constructs (FRNK, Y397F-FAK) or control vector, pcDNA3. The expression levels of endogenous FAK, FRNK and Y397F-FAK were determined by Western blot analysis using anti-FAK polyclonal antibody. (B) The effect of dominant-negative variants of FAK on CD98-induced cell adhesion rate was determined as described in Materials and Methods. In addition, whether addition of 0.5 µM Mn²⁺ interferes with the effect of PP2 on cell adhesion was determined by the same way. Results are expressed as in Figure 3 (B). ^*P < 0.05, ^**P < 0.01 (C) FAK variants- or mock-transfected MCF-7 cells were treatred with anti-CD98 mAb and secondary antibody and then analyzed for expression of β₁ integrin through flow cytometry using FITC-conjugated anti-human β₁ integrin. Results are expressed as in Figure 3 (C). Statistical comparisons are indicated by lines. ^**P < 0.01 (D) Confocal microscopy was performed as described above to investigate the effect of dominant-negative variants of FAK on clustering of β₁ integrins. Scale bar, 50 µm. Original magnification, × 400.

CD98-stimulated adhesion is blocked by actin cytoskeletal inhibition

Because the activation of FAK and Src leads to the reorganization of cytoskeleton, we investigated the role of actin cytoskeleton for CD98-stimulated cell adhesion. To investigate whether disrupting actin polymerization or reorganization could inhibit CD98-induced cell adhesion, CD98 were cross-linked in the presence of 4 µM cytochalasin D or 10 µM phalloidin. Pretreatment of MCF-7cells with cytochalasin D or phalloidin inhibited CD98-stimulated MCF-7 cell adhesion to fibronectin (Figure 5A). However, cytochalasin D or phalloidin-treated cells strongly adhered to fibronectin when treated with 0.5 µM MnCl₂, indicating that β₁ integrins on those cells are potentially functional, but cytochalasin D or phalloidin treatment inhibits inside-out activation of β₁ integrins or stabilization of adhesion complex through actin cytoskeleton.

Figure 5.

Cytochalasin D or phalloidin treatment inhibits CD98-induced adhesion of MCF-7 cells to fibronectin, but not surface expression and clustering of β₁ integrins in MCF-7 cells. (A) MCF-7 cells were incubated with anti-CD98 mAb in the presence or absence of cytochalasin D (4 µM) or phalloidin (10 µM) for 1 h. The effects of DMSO vehicle control are also shown. The effect of cytochalasin D treatment on cell adhesion was determined as described in Materials and Methods. In addition, whether addition of 0.5 µM Mn²⁺ interferes with the effect of cytochalasin D or phalloidin on cell adhesion was determined by the same way. Results are expressed as Figure 3B. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 (B) MCF-7 cells treated with anti-CD98 mAb in the presence or absence of cytochalasin D or phalloidin were analyzed for expression of β₁ integrin using flow cytometry. Results are expressed as in Figure 3 (C). ^**P < 0.01 (C) Confocal microscopy was performed as described above to investigate the effect cytochalasin D or phalloidin of on clustering of β₁ integrins. Scale bar, 50 µm. Original magnification, × 400.

Next, we investigated whether cytoskeletal reorganization is required for CD98-mediated upregulation of clustering and surface expression of β₁ integrins. Flow cytometric analyses revealed that cytochalasin D and phalloidin had no effects on the upregulation in expression levels of β₁ integrins by CD98 activation (Figure 5B). Confocal analysis showed that blocking actin rearrangement had no effect on clustering of β₁ integrins induced by cross-linking CD98 (Figure 5C). Taken together, these results indicate that actin polymerization is necessary for the CD98-induced adhesion of MCF-7 cells to extracellular matrix proteins, but may not be important for the surface expression and clustering of β₁ integrins.

Discussion

In this study, we have examined the effect of CD98 activation on adhesion of MCF-7 human breast carcinoma cells to ECM. We report that: 1) cross-linking CD98 increases surface expression and clustering of β₁ integrins as well as their affinity in MCF-7 cells. 2) inhibition of CD98-induced FAK phosphorylation or blocking actin rearrangement does not prevent surface expression and clustering of β₁ integrins but their conformational changes, which leads to reduction of cell adhesion.

Previous studies have demonstrated that CD98 activation leads to up-regulation of β₁ integrin activity (Fenczik et al., 1997; Cai et al., 2005). The underlying mechanisms could include up-regulation in the cell surface expression levels of β₁ integrin by exocytosis, its avidity by clustering, or its affinity by conformational changes. Our results suggest that some combination of mechanisms described above may be involved in CD98-mediated regulation of β₁ integrin activity, although it is difficult to determine the relative contributions of each (Hato et al., 1998).

Our flow cytometry analysis showed that cell surface expression of β₁ integrin was enhanced by CD98 activation, suggesting that CD98 upregulates the activity of β₁ integrin not only by inducing conformational changes (Chandrasekaran et al., 1999; Cai et al., 2005), but also by augmenting cell surface expression of active β₁ integrin quantitatively. CD98 stimulation increased the expression of β₁ subunit at cell surface in a relatively short time. In addition, Western blot analysis showed that CD98 activation did not have any effect on the total expression of β₁ integrin in MCF-7 cells, suggesting that CD98 activation may be involved in transport from endocytic vesicles to cell surface rather than de novo synthesis of β₁ integrin. This result is consistent with the observation that surface expression of CD98 increases in a human placental trophoblast cell line (BeWo) in the presence of CD98 mAb (Dalton et al., 2007).

According to our confocal microscopy analysis, cross-linking CD98 induces clustering of β₁ integrins on MCF-7 cells. This is supported by the results showing that CD98 engagement led to formation of round clusters of β₁ integrins (Kolesnikova et al., 2001; Rintoul et al., 2002). Clustering of integrins has been known to be a well-established mechanism to enhance integrin mediated adhesion (Jaakkola et al., 2003). Since CD98 specifically associates with β₁ integrins (Zent et al., 2000; Fenzik et al., 2001; Miyamoto et al., 2003), clustering might occur passively as a consequence of physical associations of integrins with CD98 without regard to cytoskeleton. However, this study showed that cross-linking CD98 induces more clustering of β₁ integrin than cross-linking β₁ integrin on MCF-7 cells, suggesting that CD98-induced clustering of β₁ integrins could not result from simple secondary antibody-mediated cross-linking of CD98 and its subsequent clustering of β₁ integrins. Alternatively, CD98 activation could induce actin reorganization and the associated distribution of integrins through inside-out signaling. Cross-linking of Ly6, a hemopoietic cell differentiation antigen found on a subset of CD8 T cells in the periphery, could trigger a signal for cytoskeletal reorganization and clustering of LFA-1 (Jaakkola et al., 2003). Phalloidin or cytochalasin D treatment did not inhibit CD98-induced clustering of β₁ integrins, suggesting that cross-linking of CD98 causes clustering of β₁ integrins on the surface of MCF-7 cells via a mechanism independent of reorganization of actin cytoskeleton. Interestingly, the same treatment inhibited the effects of CD98 on cell adhesion, but not surface expression and clustering of β₁ integrins. These results indicate that increased surface expression and clustering of β₁ integrins is not sufficient for CD98-induced cell adhesion.

Our data demonstrate that FAK, Src and actin cytoskeleton are required for CD98-induced cell adhesion, but not for surface expression and clustering of β₁ integrins. Previous studies showed that cross-linking CD98 increased phosphorylation of FAK dependent on β₁ integrin-mediated signaling pathway (Rintoul et al., 2002; Cai et al., 2005). Our study showed that CD98-induced cell adhesion was significantly reduced by the pretreatment with PP2. These results were confirmed by over-expression of dominant negative forms of FAK in this study and consistent with previous report (Rintoul et al., 2002). Furthermore, cytochalasin D or phalloidin inhibited CD98-induced cell adhesion. PP2-, cytochalasin D- or phalloidin-treated cells strongly adhered to fibronectin when treated with 0.5 µM MnCl₂, indicating that β₁ integrins on those cells are potentially functional. It has been proposed that FAK-, Src-, and actin cytoskeleton-dependent signaling could play a role in direct induction of a conformational change of β₁ subunit (Gomez-Rodriquez et al., 2007; Thamilselvan et al., 2007) or maintenance of talin activity (Cram and Schwarzbauer, 2004). Thus, CD98 signals may induce conformational changes in β₁ integrin through phosphorylation of FAK and reorganization of cytoskeleton. Alternatively, CD98 activation might modulate the affinity of β₁ integrin by directly inducing conformational changes, because CD98 physically associates with β₁ integrins. Previously, it was shown that thrombospondin-bound integrin associated protein (CD47) physically and functionally modifies integrin α_IIbβ₃ by its extracellular domain rather than traditional inside-out signaling (Fujimoto et al., 2003). In this case, FAK phosphorylation and subsequent actin reorganization may not be involved in the activation of β₁ integrins, but in the stabilization of adhesion structure (Mitra et al., 2005; Alon and Dustin, 2007). It remains to be addressed whether CD98-induced FAK phosphorylation and actin cytoskeletal reorganization could directly increase β₁ integrin affinity or simply the stability of adhesion complex.

In summary, we demonstrate that CD98 activation leads to an increase in surface expression and clustering of β₁ integrins, and that FAK, Src, and a functional actin cytoskeleton are required for CD98-induced cell adhesion to matrix. It remains elusive whether increase in surface expression of β₁ integrins causes enhancement of cell adhesion, and whether binding of cognate ligands for CD98 actually will induce clustering of β₁ integrins, which will increase binding avidity. In addition, it is not clear whether or not CD98 stimulates β₁ integrin affinity by an inside-out signaling pathway that requires FAK, Src, and a functional actin cytoskeleton. These issues should be addressed for further understanding the molecular mechanism of CD98 functions.

Abbreviations

ECM

extracellular matrix

FAK

focal adhesion kinase

FRNK

FAK related non-kinase

R-PE

R-phycoerythrin

SFM

serum-free medium

TCR

T cell receptor