Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Mammary-derived growth inhibitor (MDGI) interacts with integrin α-subunits and suppresses integrin activity and invasion


The majority of mortality associated with cancer is due to formation of metastases from the primary tumor. Adhesion mediated by different integrin heterodimers has an important role during cell migration and invasion. Protein interactions with the β1-integrin cytoplasmic tail are known to influence integrin affinity for extracellular ligands, but regulating binding partners for the α-subunit cytoplasmic tails have remained elusive. In this study, we show that mammary-derived growth inhibitor (MDGI) (also known as FABP-3 or H-FABP) binds directly to the cytoplasmic tail of integrin α-subunits and its expression inhibits integrin activity. In breast cancer cell lines, MDGI expression correlates with suppression of the active conformation of integrins. This results in reduced integrin adhesion to type I collagen and fibronectin and inhibition of cell migration and invasion. In tissue microarray of 1331 breast cancer patients, patients with MDGI-positive tumors had more favorable 10-year distant disease-free survival compared with patients with MDGI-negative tumors. Our data indicate that MDGI is a novel interacting partner for integrin α-subunits, and its expression modulates integrin activity and suppresses cell invasion in breast cancer patients. Retained MDGI expression is associated with favorable prognosis.


Integrins are heterodimeric transmembrane receptors capable of anchoring cells to extracellular matrix and mediating bidirectional signals between the cytosol and extracellular space (Hynes, 2002). Altogether 18 α- and 8 β-subunits in mammals are known to form 24 specific dimers, each with a different ligand-binding preference. On the cell surface, integrins can be expressed in a bent low-affinity conformation, in which they are inactive and unable to bind extracellular ligands. Either ligand binding to the extracellular domain of the integrin or regulator binding to the intracellular tail of the integrin can activate integrin and shift the conformation to the extended high-affinity state (reviewed in Legate et al., 2009). Separation of cytoplasmic α- and β-tails is associated with the final step in integrin activation, which is supported by binding of the talin head domain to the cytoplasmic β-tail (Anthis et al., 2009). However, talin binding alone may not be sufficient as members of the kindlin family interact with β-tails and cooperate with talin in integrin activation (reviewed in Kloeker et al., 2004; Shi et al., 2007; Böttcher et al., 2009). Different integrin conformations are important for cell motility, and activated integrins are present at the leading edge of migrating cells (Kiosses et al., 2001).

Mammary-derived growth inhibitor (MDGI) (also known as FABP-3 or H-FABP) is a small 15-kDa cytoplasmic protein that belongs to a family of fatty acid-binding proteins (FABPs). FABPs have a unique tissue-specific expression pattern and they bind hydrophobic ligands in a reversible manner (Hertzel and Bernlohr, 2000). The in vivo functions of FABPs have remained elusive, but they are proposed to function in solubilization, transportation and energy homeostasis of fatty acids. Some FABP functions may be independent of their fatty acid ligand. A C-terminal-derived 11 amino acids peptide of MDGI reduced colony formation of breast cancer cell lines in vitro and tumor formation in nude mice (Wang and Kurtz, 2000). However, MDGI lacks an N-terminal signal peptide and it is not clear whether it or its fragments are secreted in vivo (Clark et al., 2000). The expression of MDGI mRNA is hormonally regulated in breast epithelial cells during development and is highly expressed at the time of maximal differentiated function (just before lactation) (Kurtz et al., 1990). Interestingly, antiestrogen therapy can stabilize MDGI mRNA in vitro (Huynh and Pollak, 1997) suggesting that MDGI may also be involved in anticancer activity of antiestrogens. MDGI gene is silenced by hypermethylation in human breast cancer cell lines and in some primary breast tumors (Huynh et al., 1996). Taken together, these data suggest that MDGI may have a role as a tumor suppressor.

In recent years, there have been significant advances in our understanding of how proteins interacting with different regions of the integrin β-subunits positively or negatively regulate integrin activity. Conversely, the possible involvement of integrin α-subunits in integrin activity regulation has been poorly characterized apart from the role of positively charged residue(s) in integrin α-subunit, which contribute to a salt bridge involved in maintaining integrins in an inactive conformation (Anthis et al., 2009). Here, we present data of a novel integrin α-subunit-binding protein, MDGI, which interacts directly with several α-integrins and inhibits integrin activity in cells. We further demonstrate that MDGI expression correlates with increased distant disease-free survival (DDFS) in a large cohort of breast cancer patients.


MDGI interacts with several integrin α-subunits

To date, the mechanisms underlying the biological significance of the integrin α-subunit cytoplasmic tails are poorly understood. To identify novel interactors of integrin α-tails, the cytoplasmic domain of α2-integrin was used in a yeast two-hybrid screen as bait. The bait sequence contained the conserved membrane-proximal WKLGFFKR sequence shared by most integrin α-subunits followed by the α2-specific tail sequence. The screen was done against mouse E17 Matchmaker complementary DNA library (Clontech, Mountain View, CA, USA). One of the interacting candidate molecules was MDGI. The specificity of interaction was evaluated in yeast mating experiments, in which pGADT7–MDGI was mated with different pGBKT7 fusion proteins expressing cytoplasmic domains of human α1, α2, α10 or α11 (Figure 1a). Surprisingly, MDGI was able to interact with all the cytoplasmic α-domains tested, suggesting that MDGI is capable of interacting with several α/β1-heterodimers. This would be likely to occur through MDGI interaction with the highly conserved WKLGFFKR sequence of most α-subunits (WKLGFFAH sequence in α10, WKLGFFRS in α11) (Figure 1b). Surface plasmon resonance (Biacore, GE Healthcare, Little Chalfont, UK) was used next to test binding of purified recombinant MDGI to integrin cytoplasmic domains. The peptides α1-cyt, α2-cyt and β1-cyt were coupled to individual flow cells of the Biacore sensor chips. Different concentrations of MDGI protein were added in the solutions and binding to the peptides was monitored relative to an uncoupled control flow cell. MDGI interacted in a similar manner with both α1-cyt and α2-cyt peptides (Figure 1c). However, no binding to the cytoplasmic tail peptide of β1 (β1-cyt) was detected (Figure 1c).

Figure 1

MDGI interacts with integrin cytoplasmic α-domains. (a) MDGI (pGADT7–MDGI) and different collagen-binding integrin cytoplasmic domains (pGBKT7-tails) were transformed into yeast strains AH109 or Y187, respectively, and mated to confirm protein–protein interaction. Empty expression vectors or interaction between p53 and large T antigen were included as negative and positive controls. (b) Integrin cytoplasmic α-domains used in the yeast two-hybrid screen and mating experiments. (c) Surface plasmon resonance sensorgrams of MDGI binding to α1-, α2- or β1-integrin peptides. MDGI concentrations used were 0.1, 1, 3 or 5, 10, 20 and 40 μM. AU, arbitrary units. (d) Stably GFP–MDGI-expressing MDA-MB-468 cells were harvested and subjected to immunoprecipitations (IP) with the indicated antibodies followed by blotting with anti-β1-integrin or anti-GFP.

To study the interaction of integrins and MDGI in intact cells, stably green fluorescent protein (GFP) or GFP–MDGI-expressing MDA-MB-468 and MDA-MB-231 polyclonal breast cancer cell lines were generated. Both parental cell lines are endogenous MDGI-negative. In an immunoprecipitation assay, GFP–MDGI was co-precipitated with an anti-β1-antibody but not with control immunoglobulin G (Figure 1d). In cells, β1-subunit pairs with many different α-subunits all sharing the conserved WKLGFFKR sequence. Therefore, the ability of MDGI to associate with those non-collagen-binding integrins, which are expressed in MDA-MB-231 cells (Supplementary Figure 1a) was assessed by a technique called in situ proximity ligation assay (PLA) (Supplementary Figure 2a), which generates an antibody-based signal when two proteins in an intact cell are in close proximity (in the 20–100 nm range) to each other (Söderberg et al., 2006). As expected from the above data, a proximity ligation signal between anti-β1 and anti-GFP antibodies was detected at the plasma membrane in cells expressing GFP–MDGI but not in non-transfected or GFP-transfected cells (Figure 2a and Supplementary Figure 2b). Significant PLA signal was also detected between GFP–MDGI and α5-integrin, a fibronectin-binding integrin subunit, and MDGI and β4-integrin (which heterodimerizes with α6 to form a laminin receptor) (Figures 2b–c and Supplementary Figure 2b). Therefore, as would be expected for a protein that binds to the conserved sequence shared between integrin α-subunits, the ability of MDGI to associate with integrins in cells was not restricted to collagen-binding β1-integrins. These data demonstrate that MDGI can associate with several integrins also in intact cells. As the expression of endogenous MDGI is lost in all cultured cells because of hypermethylation (Huynh et al., 1996; Nevo et al., 2009), complex formation between endogenous MDGI and integrin cannot be tested. Together these data indicate that MDGI is a novel integrin α-tail-binding protein.

Figure 2

MDGI associates with various integrin subunits. MDA-MB-231 cells transfected with GFP or GFP–MDGI were plated on acid-prewashed glass coverslips, fixed and incubated overnight with primary antibodies against GFP and β1-integrin (a), GFP and α5-integrin (b) or GFP and β4-integrin (c) followed by incubation with oligo-linked secondary antibodies and counterstain with 4,6-diamidino-2-phenylindole (DAPI). PLA signal (red), which is produced when two proteins are closely associated, was detected and quantified from transfected and non-transfected cells (GFP and β1-integrin n=72 cells; GFP and α5-integrin n=64 cells; GFP and β4-integrin n=104 cells). Shown are mean fluorescent intensities of the PLA signal (mean±s.e.m.; ***P<0.001; two-tailed Student's t-test). Scale bar 10 μm. High-resolution images of individual cells are provided in Supplementary Figure 2b.

MDGI expression inhibits adhesion, integrin activity and invasion

Several cytoplasmic β-tail-interacting proteins are known to regulate integrin affinity for the extracellular ligands. The best studied examples are talin and kindlins, which are known to interact directly with cytoplasmic β1- and β3-domains and thus activate integrins through conformational changes (Tadokoro et al., 2003; Kloeker et al., 2004; Shi et al., 2007; Bouaouina et al., 2008). To test whether MDGI modulates integrin-mediated functions, we analyzed adhesion of multiple breast cancer cell lines (MDA-MB-468, MDA-MB-231, MCF-7 and SKBr-3) transiently or stably (polyclonal cell lines) expressing GFP and GFP–MDGI. Cell adhesion to type I collagen was significantly decreased in all GFP–MDGI-expressing cell lines (Figure 3a). Similar inhibition in adhesion of GFP–MDGI-expressing cells to fibronectin was also observed (Supplementary Figure 3). Activity of β1-integrins has been widely studied using antibodies against conformation sensitive epitopes (Byron et al., 2009). Integrin 9EG7 antibody detects the extended and talin-bound β1-integrin conformations (Bazzoni et al., 1995; Czuchra et al., 2006; Zhang et al., 2008) and HUTS-21 antibody recognizes active ligand-bound β1-integrin and also stimulates adhesion (Luque et al., 1996). Using these antibodies in fluorescence-activated cell sorting analyses, we found that GFP–MDGI-expressing MDA-MB-468 and MDA-MB-231 cells expressed significantly less active β1-integrin on their cell surface compared with control cells (Figure 3b). This reduction was specific to the active conformer of β1-integrin as the amount of total β1-integrin on the cell surface (detected with P5D2 pan-β1-integrin antibody) was similar between the GFP and GFP–MDGI-expressing cells in both cell lines (Figure 3c). Furthermore, the total expression levels of different integrin subunits on the cell surface were not significantly altered by MDGI expression (Supplementary Figure 1). Manganese (Mn2+) binding to the integrin ectodomain is sufficient to activate integrins through an outside-in signaling mechanism (Hynes, 2002). When stimulated with manganese, a substantial increase in the amount of integrin-bound Alexa-647-labeled fibronectin fragment, which harbors the integrin-binding RGD motif, was seen in the control cells compared with GFP–MDGI-expressing cells (Figure 3d). These data are indicative of an inhibitory role for MDGI in regulating integrin activity. The different cell lines used here have distinct integrin expression profiles on their cell surface (Supplementary Figure 4). The fact, that MDGI inhibited adhesion and integrin activity in these cell lines further underscores that MDGI expression can influence the function of several α/β1-integrins mediating adhesion to different matrixes. The correlation between MDGI expression and reduced integrin activity would imply that integrin binding to MDGI would be predominantly in the inactive conformation. Indeed, PLA signal detected between active β1-integrin-specific antibody (12G10) and MDGI was greater than threefold lower (Figure 3e) than PLA signal detected between total β1-integrin antibody and MDGI (Figure 2a). In line with this, expression of GFP–MDGI significantly reduced association between integrin-activating protein kindlin and active β1-integrin whereas expression of GFP alone had no effect (Figure 3f). Thus, it is possible that MDGI induces integrin inactivation, at least in part, by interfering with the ability of kindlin to bind to β1-integrin.

Figure 3

MDGI expression correlates with reduced cell adhesion and altered β1-integrin activity. (a) Adhesion of either stably (breast cancer cell lines MDA-MB-468 and MDA-MB-231) or transiently (breast cancer cell lines MCF-7 and SKBr-3) GFP and GFP–MDGI-expressing cells on type I collagen was measured after 60-min incubation. Shown are representative graphs of experiments (for MDA-MB-468 and MDA-MB-231 cells n>7; MCF-7 n=3; SKBr-3 n=2) with similar results (mean±s.e.m. of 8–12 parallel wells; **P<0.01; ***P<0.001; two-tailed Student's t-test). (b) The amount of active β1-integrin epitope on the cell surface in stably GFP or GFP–MDGI-expressing MDA-MB-468 and MDA-MB-231 cell lines was determined by fluorescence-activated cell sorting (FACS) analysis using 9EG7 and HUTS-21 antibodies. Active integrin was measured relative to P5D2 antibody, which recognizes total β1-integrin. Shown are representative graphs (mean±s.e.m.; n=3–4 parallel samples; **P<0.01; ***P<0.001; two-tailed Student's t-test). (c) Relative total β1-integrin amounts (pan-β1-integrin antibody P5D2) from experiments described in (b). Graphs represent (mean±s.e.m.; n=2 parallel samples). (d) The binding of Alexa-647-labeled fibronectin repeat 7–10 (FN) fragment to stably GFP or GFP–MDGI-expressing MDA-MB-468 and MDA-MB-231 cell lines was determined by FACS. In all, 5 mM MnCl2 was added to reaction mixture to stimulate FN binding. Alexa-647-FN binding was measured relative to non-stimulated cells. Shown are representative graphs (mean±s.d.; n=2 parallel samples). (e, f) MDA-MB-231 cells transfected with GFP or GFP–MDGI were plated on acid-prewashed glass coverslips, fixed and incubated overnight with primary antibodies against GFP and active β1-integrin (e) or kindlin and active β1 epitope 12G10 (f) followed by incubation with oligo-linked secondary antibodies and counterstain with 4,6-diamidino-2-phenylindole (DAPI). PLA signal (red), which is produced when two proteins are closely associated, was detected and quantified from transfected and non-transfected cells. Shown are mean fluorescent intensities of the PLA signal (mean±s.e.m.; GFP and active β1-integrin n=55 cells; active β1-integrin and kindlin n=92 cells; *P<0.05; ***P<0.001; NS, not significant; two-tailed Student's t-test). Scale bar 10 μm.

In line with the above data, GFP–MDGI-expressing cells displayed reduced spreading and less vinculin-positive focal adhesions when compared with untransfected cells adhering to collagen (Figure 4a). Adhesion to the extracellular matrix modulates cell migration. Migration speed is highest when adhesiveness between the substratum and cell is intermediate, because strong physical interaction exceeds the capacity of the cell to contract for moving, and relatively low adhesiveness prevents formation of proper adhesive sites (Huttenlocher et al., 1995). Using time-lapse imaging, we found that transiently transfected GFP–MDGI cells migrated significantly shorter distances than control cells (Figure 4b). To test invasiveness of stably GFP or GFP–MDGI-expressing MDA-MB-468 cells, both cell lines were allowed to invade into a mixture of fibronectin and matrigel in an inverted three-dimensional (3D) assay (the cells migrate upward, arrows in Figure 4c). GFP–MDGI cells invaded significantly less than control cells as seen in the number of green cells above the filter (dotted line) relative to all cells in side views of invading cells in z-stack images (Figure 4c).

Figure 4

MDGI inhibits both cell migration and invasion. (a) Stably GFP–MDGI-expressing MDA-MB-468 cells were allowed to adhere to type I collagen for 60 min and stained for vinculin and phalloidin, scale bar 10 μm. Below the image is shown quantification of cell area (n=9 per group; mean±s.e.m.; ***P<0.001; two-tailed Student's t-test). (b) Cumulative length of cell migration path measured from MDA-MB-231 GFP and GFP–MDGI cells by tracking over time (n=24 and n=35, respectively; mean±s.e.m.; **P<0.01; two-tailed Student's t-test). Shown are eight representative trajectories from individual cells of both cell lines on plastic. (c) Stably GFP or GFP–MDGI-expressing MDA-MB-468 cells were allowed to invade in inverted 3D invasion assay into mixture of matrigel and fibronectin. Shown are representative side views of invading cells in a confocal z-stack, in which transwell membrane is illustrated as a white dashed line. Black arrow indicates direction for migration. Quantification of area of invaded cells relative to the total area of cells is shown below (n=14 fields per group; mean±s.e.m.; ***P<0.001; two-tailed Student's t-test).

MDGI-expressing cells display phenotypic reversal in matrigel

The finding that MDGI-expressing cells are less invasive prompted us to investigate their phenotype in a 3D culture in more detail. In a reconstituted basement membrane (matrigel), normal human breast epithelial cells are able to form a structurally and functionally differentiated acinar phenotype, while breast carcinoma cells grow in disorganized colonies (Petersen et al., 1992), which can be reverted to a more normal phenotype by treating breast cancer cells with function blocking β1-integrin antibodies (Weaver et al., 1997). In line with the reduced β1-integrin-mediated adhesion, GFP–MDGI-expressing malignant breast cancer cells showed a shift toward acinar-like phenotype in matrigel when compared with the disorganized colonies formed by GFP-expressing cells (Figure 5). The proportion of cells growing in acinar-like structures was 42±8% higher in GFP–MDGI-expressing cells compared with GFP control cells. These results indicate that re-expression of MDGI correlates with phenotype reversion, which is likely to be connected to the reduced integrin activity detected in MDGI-expressing cells.

Figure 5

MDGI expression correlates with phenotypic reversal in 3D. Stably GFP or GFP–MDGI-expressing MDA-MB-468 cells were embedded in matrigel and incubated for 7 days before cell morphology was observed and scored. Arrowhead points to disorganized colony and arrow to acinus. Scale bar 50 μm. Data are represented below as (mean±s.e.m.; n=18–22 fields of view; ***P<0.001).

Retained expression of MDGI is associated with increased DDFS in breast cancer

MDGI expression correlated with inhibition of β1-integrin activity, cell adhesion and invasion in breast cancer cells in vitro. Expression of MDGI also induced phenotypic reversal in these cancer cells in a 3D matrix environment. All these alterations in cell behavior are highly relevant to progression of clinical cancers. Therefore, to examine whether MDGI expression correlates with clinical parameters in vivo, we stained a tissue microarray consisting of 1331 unilateral female invasive breast cancers at the time of diagnosis with anti-MDGI antibody. The series has been described in detail earlier (Lundin et al., 2001). In brief, the median age at the time of the diagnosis was 58.0 years, and 28% of the women were aged 50 or less. A total of 750 (59%) tumors were classified as pT1 (2 cm in diameter), 471 (37%) pT2 (2.1–5.0 cm) and 49 (3.9%) pT3-4 (>5.0 cm) according to the Tumor-Node-Metastasis classification. In total, 809 (63%) subjects had axillary node-negative and 471 (37%) node-positive cancer, and none of the patients had distant metastases at diagnosis (M0). Most cancers were estrogen receptor (ER)-positive (60%) and progesterone receptor-positive (47.3%), while only 16.8% were HER2-positive. Patients with ER-positive and node-positive cancers were usually treated with antiestrogen therapy (62.0%); whereas most (90.5%) patients with node-negative cancer did not receive any adjuvant systemic therapy. None of the patients received adjuvant therapy targeted to HER2 or EGFR. Adjuvant chemotherapy (consisting usually of six cycles of cyclophosphamide, methotrexate and fluorouracil, cyclophosphamide, epirubicin and fluorouracil or cyclophosphamide, doxorubicin and fluorouracil) was administered to 20.3, 12.7 and 36.3% of patients with triple-negative cancer (ER-, progesterone receptor- and HER2-negative), HER2-negative cancer and node-positive cancer, respectively, and adjuvant antiestrogen therapy (consisting usually of tamoxifen) was administered to 24.4, 24.8 and 55.2%, respectively. The median follow-up time was 9.5 years after the diagnosis.

Expression of MDGI was evaluated by two of the investigators (PH and PB). Both investigators were blinded to the clinicopathological data at the time of scoring. MDGI staining was scored either negative (no expression in cancer cells=0) or positive (either faint=1, moderate=2 or strong=3 expression in any proportion of the cancer cells). Specificity of the staining has been described earlier (Nevo et al., 2009). MDGI immunoreactivity was absent in 5.7% (76 of 1331), faint in 40.6% (540 of 1331), moderate in 36.5% (486 of 1331) and strong in 17.2% (229 of 1331) of the invasive breast carcinoma specimens. It is noteworthy that absence of MDGI expression was significantly more frequent in triple-negative (ER-, progesterone receptor- and HER2-negative) breast carcinomas (P=0.001; Supplementary Table 1).

When analyzing the data, we observed that all tumors with MDGI expression (even faint) showed a trend toward increased DDFS (Figure 6a; calculated from the date of diagnosis to the first diagnosis of distant metastasis or death, whichever occurred first). Therefore, we chose to analyze the series divided into MDGI-negative and -positive groups. In this series, women whose cancer expressed MDGI had more favorable DDFS compared with patients whose cancer was MDGI-negative (10-year DDFS 72 vs 61%, respectively; hazard ratio 0.68, 95% confidence interval 0.46–1.00 in a univariate Cox analysis; log-rank test two-sided P=0.047; Figure 6b and Table 1). In exploratory subgroup analyses, MDGI expression was significantly associated with favorable DDFS in axillary node-positive disease (P=0.035; Figure 6c), HER2-negative cancer (P=0.008; Figure 6d), and triple-negative disease (P=0.031; Figure 6e), but not in ER-positive axillary node-negative or HER2-positive cancer. MDGI expression was not an independent prognostic factor in a Cox multivariate analysis unlike the axillary nodal status, tumor size or the HER2 status. Taken together, these data indicate that in breast cancer subtypes with poorer prognosis or limited benefit from targeted therapies retained expression of MDGI is associated with increased survival.

Figure 6

Patients with retained MDGI expression have increased DDFS. (a) Survival of all patients with different MDGI levels. MDGI expression was scored according to the intensity of staining in immunohistochemistry with MDGI antibody. No MDGI expression scored 0; increasing MDGI expression levels, scored 1–3. (b) Combined data from the graph presented in (a). Solid line indicates MDGI-negative patients (scored 0; n=76) and dashed line indicates MDGI-positive patients (scored 1–3; n=1255), P=0.047. (c) Survival of the patients with axillary node-positive disease. Solid line indicates MDGI-negative patients (n=32) and dashed line indicates MDGI-positive patients (n=439), P=0.035. (d) Survival of the patients with HER2-negative disease. Solid line indicates MDGI-negative patients (n=53) and dashed line indicates MDGI-positive patients (n=895), P=0.008. (e) Survival of patients with triple-negative breast cancer (HER2-, ER- and progesterone receptor-negative). Solid line indicates MDGI-negative patients (n=21) and dashed line indicates MDGI-positive patients (n=179), P=0.031.

Table 1 Ten-year DDFS according to MDGI expression


Here, we show that MDGI is a novel binding partner for integrin cytoplasmic α-tail. In breast cancer cells, expression of MDGI correlates with inactive integrin conformation, reduced adhesion, migration and invasion. In addition, re-expression of MDGI in human MDA-MB-468 breast cancer cell line correlated with partial phenotype reversion in 3D culture suggesting a more normal behavior. Furthermore, classification of breast cancer patients according to MDGI expression clearly indicated positive correlation between the retained MDGI levels and DDFS. Taken together, these data suggest that MDGI can function as a modulator of integrin activity retarding cell invasion and metastasis.

Several regulatory proteins have been shown to bind directly to integrin cytoplasmic β-tail. Proteins binding to the β-subunit like talin and kindlins function to activate integrins, whereas negative regulators include proteins like ICAP-1, Filamin and Dok1, which can compete with talin for overlapping binding site on integrin cytoplasmic β-tail (Bouvard et al., 2003; Wegener et al., 2007; Millon-Frémillon et al., 2008; Takala et al., 2008). Thus, significant advances have been made in our understanding of the role of β-tail in regulation of integrin activation. In contrast, the possible role for cytoplasmic α-tail in this process is less clear.

Here, we have identified MDGI as a novel binding protein for integrin α-subunits. At present, we do not have the full molecular details of how MDGI binding to the cytoplasmic α-tails could interfere with integrin activity. However, as MDGI binds directly to several integrin α-tails and MDGI expression in cells correlates with reduced integrin activity, it is likely that MDGI may influence integrin activity by a direct mechanism. Integrins have been shown to be held in an inactive conformation by an α/β-salt bridge, which includes the arginine of this cytoplasmic α-sequence (Luo et al., 2004; Vinogradova et al., 2004). Binding of MDGI to this sequence could be expected to disrupt the salt bridge, which would lead to release of the integrin-inactivating clasp. However, we found that binding of MDGI to integrin correlates with reduced activity. Recent models display the α-subunit in relatively close proximity of the bulky talin head binding to the β-tail (Anthis et al., 2009). Using molecular modeling, we investigated how MDGI could influence talin binding to the β-tail. Although the model is based on the NMR-structure of MDGI, it must be considered fully hypothetical since at present we lack information of the orientation in which MDGI interacts with the integrin α-tail. However, on the basis of the model it is possible that MDGI interaction with integrin α-tail could interfere with the ability of talin to interact with integrins and the membrane and induce the active integrin conformation in cells (Figure 7). The ability of kindlin to activate integrins is talin-dependent (Böttcher et al., 2009; Harburger et al., 2009). We find that expression of MDGI reduces kindlin association with active β1-integrin in cells. This is also supportive of our model. However, the exact mechanism how MDGI inhibits integrin activity requires further investigation.

Figure 7

A hypothetical model of how MDGI could hinder indirectly the binding of talin into the cytoplasmic tail of integrin β-subunit based on the published structures (Anthis et al., 2009; Lau et al., 2009). (a) Talin (blue) binding into the integrin β-tail (orange) and the simultaneous interaction of the positively charged residues of talin (yellow arrows; amino acids shown as a space fill model) with partially negatively charged head-groups (red minus-signs) of membrane lipids (gray rectangular box) interrupts the position of the transmembrane and intracellular parts of the integrin α-subunit (silver; gray arrows), thus separating subunits from each other. (b) The hypothetical binding model of MDGI (brown) into integrin α-chain sterically blocks the talin (transparent blue) binding into the integrin β-chain (orange).

Increasing evidence points to an important role for β1-integrins in tumor progression in breast cancer. Data from cell culture models indicate that β1-integrins regulate phenotypic reversal of breast cancer cells in 3D cultures (Weaver et al., 1997; Wang et al., 2002). In addition, increased β1-expression is associated with an invasive signature in primary mammary tumors and with decreased survival in breast cancer (Wang et al., 2004; Yao et al., 2007). Interestingly, recent described tumor suppressor (SCAI) negatively regulates expression of β1-integrin and loss of SCAI in several cancer types results in increased β1-integrin expression and invasion (Brandt et al., 2009). Integrins also show potential as therapeutic targets because inhibition of β1-integrin with antibodies results in decreased proliferation and increased apoptosis in vitro and in vivo in xenograft models, and genetic ablation of β1-integrin expression inhibits mammary tumor formation in a mouse model of breast cancer (White et al., 2004). Currently, it is unclear how integrin activity correlates with clinical outcome of cancer. Here, we show that expression of MDGI correlates with reduced β1-integrin activity and invasion in vitro and with increased DDFS in breast cancer patients. We propose that MDGI-induced inhibition of β1-integrin function is linked to the putative tumor-suppressor functions assigned to MDGI in earlier studies (Wang and Kurtz, 2000). Hypermethylation has been shown to result in silencing of MDGI expression in breast cancer (Huynh et al., 1996), thus it is possible that loss of MDGI and increased integrin activity are the result of epigenetic changes occurring during cancer progression. This suggests an intriguing link between epigenetic changes and modulation of integrin activity relevant in human cancer.

Materials and methods

Antibodies and reagents

For the flow cytometry, antibodies against active β1-integrin epitope (9EG7 or HUTS-21; BD Biosciences, San Jose, CA, USA) and pan-β1-integrin (P5D2; Abcam, Cambridge, UK) were used. Antibodies against α1 (MAB1973), α3 (MAB1952), α4 (MAB16983), α5 (MAB1999) and β4 (MAB2060) were from Chemicon (Millipore, Billerica, MA, USA) and α2 (MCA2025), α6 (MCA699) and β3 (MCA728) from Serotec (Dusseldorf, Germany). Anti-αV-integrin antibody (L230) has been described earlier (Koistinen et al., 1999). The following antibodies were used in immunoblotting: GFP (A11122; Molecular Probes, Carlsbad, CA, USA) and β1-integrin (mixture of MAB2252, Chemicon and 610468, BD Transduction Laboratories, San Jose, CA, USA). For the immunofluorescence, anti-vinculin (V9131; Sigma-Aldrich, St Louis, MO, USA) and Alexa Fluor-546 phalloidin (Invitrogen, Carlsbad, CA, USA) were used. The following antibodies were used in PLA:GFP (A11122, Molecular Probes), pan-β1-integrin (P5D2, Abcam), active β1-integrin (12G10, Abcam), kindlin (ab68041, Abcam), α5 (MCA1187, Serotec) and β4 (MAB2060, Chemicon).

Cell cultures

The origin and propagation of breast cancer cell lines have been described elsewhere (Nevo et al., 2009). For phenotype reversal assay, cells were embedded in Matrigel (BD Biosciences) according to manufacturer's protocol followed by 30-min polymerization at 37 °C. Normal growth medium was added on top of the gel and cells were grown for 7 days before analysis.

Expression vectors, transfection and selection of stabile cell lines

Full-length human MDGI cloning to pEGFP-C1 vector and generation of stably GFP or GFP–MDGI-expressing polyclonal MDA-MB-468 cells have been described earlier (Nevo et al., 2009). All transfections were done with Lipofectamine 2000 (Invitrogen) according to manufacturer's protocol. MDA-MB-231 cells were selected with 1000 μg/ml G418 to establish stable expressing polyclonal GFP and GFP–MDGI cell lines.

Yeast two-hybrid screen and mating experiments

The yeast two-hybrid screen was done as described earlier (Mattila et al., 2005). For mating experiments, mouse MDGI (IMAGE 3485639; amino acids 9–134) lacking eight amino acids from the N-terminus fused to GAL4 activation domain and different collagen-binding integrin cytoplasmic tails cloned to GAL4 DNA-binding domain (pGBKT7-tails; see tail sequences in Figure 1b) were transformed into yeast strains AH109 or Y187, respectively (Gietz and Woods, 2002). pGADT7–MDGI was originally isolated from the yeast two-hybrid screen done in the mouse E17 Matchmaker complementary DNA library (Clontech). For mating, five single colonies of both transformants were picked from fresh selective plates and resuspended in 0.4 ml of YPDA medium for overnight incubation at 30 °C shaking with 250 r.p.m. followed by plating on the selective SD agar plates for growth scoring.

Quantitative interaction analysis

The real-time protein–protein binding was analyzed by surface plasmon resonance (Biacore System, GE Healthcare), in which the binding of full-length human MDGI protein injected over the sensor surface to immobilized integrin tail peptides was measured. The α1-peptide (α1-cyt) contained the conserved α-integrin-specific sequence and a nine amino acid α1-integrin-specific sequence in double repeat (WKIGFFKRPLKKKMEKRPLKKKMEK; underlining indicates the conserved sequence). The α2-peptide (α2-cyt) represented the conserved cytoplasmic tail sequence of α2-integrin (WKLGFFKRKYEKM; underlining indicates the conserved sequence), and the β1-peptide (β1-cyt) enclosed the specific β1-integrin cytoplasmic sequence (WKLLMIIHDRREFAKFEKEKMNAKWDTGENPIYKSAVTTVVNPKYEGK). Binding reactions were performed in HBS-P running buffer (0.01 M HEPES pH 7.4, 0.15 M NaCl, 0.0005% Surfectant P20). Biotinylated α-peptides were captured on a streptavidin surface (Biacore sensor SA). One flow channel 1 was used as a negative control for nonspecific interactions between analyte and sensor surface, the other flow channel 2 was coated with 500RU of either α1-, α2- or β1-peptide for the qualitative analysis of peptide-binding reactions. The response data were analyzed for a series of MDGI concentrations ranging from 0.1 to 40 μM at a flow rate of 30 μl/min. Signal changes on the flow channel 1 were subtracted from the peptide-MDGI-binding interactions as well as nonspecific binding reactions caused by the running buffer itself.

Adhesion assay

Adhesion assay to collagen has been described earlier (Pellinen et al., 2006). In all, 1 and 5 μg/ml type I collagen precoated plates were used for transiently and stably transfected cell lines, respectively. For fibronectin studies, the plates were precoated with 5 μg/ml fibronectin. Cells were incubated on 96-well plates for 30 or 60 min, washed with phosphate-buffered saline, fixed and the amount of attached green cells per well was counted with Acumen Explorer laser scanner at 488 nm (TTP LabTech Ltd, Royston, UK).

In situ proximity ligation (PLA)

This technique visualizes two proteins in close proximity by detecting a fluorescent signal produced by the rolling-circle amplification reaction (Jarvius et al., 2007). Cells are stained with primary antibodies, followed by incubation with oligo-linked secondary antibodies. A PLA signal is produced when the two antibodies bind their antigens in close proximity to each other (20–100 nm). MDA-MB-231 cells transfected with GFP–MDGI or GFP alone were plated on acid-prewashed glass coverslips for overnight followed by fixation with 4% parafolmaldehyde and permeabilization with 0.1% Triton-X/1% bovine serum albumin/phosphate-buffered saline. Parallel samples were incubated overnight with GFP and β1, active β1, α5 or β4 antibodies, or with active β1 and kindlin antibodies (a positive control for PLA). Proximity ligation was done according to manufacturer's protocol (Duolink in situ PLA, Olink Bioscience, Uppsala, Sweden). Proximity signal (which appear as bright fluorescent dots) was detected with a Zeiss inverted wide-field microscope from several fields of view (Carl Zeiss MicroImaging, Inc., Thornwood, NY, USA). The background fluorescence was subtracted from the images and the signal intensities of individual cells were quantified by using ImageJ analysis software (

Inverted 3D invasion assay

This assay has been described earlier (Tuomi et al., 2009).


Staining protocol has been described earlier (Nevo et al., 2009).

Molecular modeling

Model of integrin's transmembrane-cytoplasmic part with bound talin was built by combining the experimentally solved structures of integrin αIIbβ3 transmembrane-cytoplasmic tail (pdb-code: 2k9j (Lau et al., 2009)) and talin F2-F3–integrinβ1D cytoplasmic tail (pdb-code: 3g9w (Anthis et al., 2009)). The continuous model of transmembrane-cytoplasmic part of the integrin β-subunit from above mentioned two structures with integrin α-subunits was based on the sequence alignment of integrin β3 and β1 combined with the RMSD-superpositioning option in BODIL (Lehtonen et al., 2004). The hypothetical model of MDGI-integrin α-chain cytoplasmic tail was built by positioning cluster of positively charged amino acids at the surface of the MDGI (pdb-code: 1g5w (Lücke et al., 2001)) into close proximity to the partially negatively charged head groups of the membrane that was placed to cover the helical region of the integrin α-chain, and simultaneously allowing the MDGI to interact with the conserved region of the cytoplasmic tail of the integrin α-chain.

Statistical analysis

When indicated, Student's t-test or Cox multivariate analysis was used for analysis.


  1. Anthis NJ, Wegener KL, Ye F, Kim C, Goult BT, Lowe ED et al. (2009). The structure of an integrin/talin complex reveals the basis of inside-out signal transduction. EMBO J 28: 3623–3632.

    CAS  Article  Google Scholar 

  2. Bazzoni G, Shih DT, Buck CA, Hemler ME . (1995). Monoclonal antibody 9EG7 defines a novel beta 1 integrin epitope induced by soluble ligand and manganese, but inhibited by calcium. J Biol Chem 270: 25570–25577.

    CAS  Article  Google Scholar 

  3. Bouaouina M, Lad Y, Calderwood DA . (2008). The N-terminal domains of talin cooperate with the phosphotyrosine binding-like domain to activate beta1 and beta3 integrins. J Biol Chem 283: 6118–6125.

    CAS  Article  Google Scholar 

  4. Bouvard D, Vignoud L, Dupé-Manet S, Abed N, Fournier HN, Vincent-Monegat C et al. (2003). Disruption of focal adhesions by integrin cytoplasmic domain-associated protein-1 alpha. J Biol Chem 278: 6567–6574.

    CAS  Article  Google Scholar 

  5. Brandt DT, Baarlink C, Kitzing TM, Kremmer E, Ivaska J, Nollau P et al. (2009). SCAI acts as a suppressor of cancer cell invasion through the transcriptional control of beta1-integrin. Nat Cell Biol 11: 557–568.

    CAS  Article  Google Scholar 

  6. Byron A, Humphries JD, Askari JA, Craig SE, Mould AP, Humphries MJ . (2009). Anti-integrin monoclonal antibodies. J Cell Sci 122: 4009–4011.

    CAS  Article  Google Scholar 

  7. Böttcher RT, Lange A, Fässler R . (2009). How ILK and kindlins cooperate to orchestrate integrin signaling. Curr Opin Cell Biol 21: 670–675.

    Article  Google Scholar 

  8. Clark AJ, Neil C, Gusterson B, McWhir J, Binas B . (2000). Deletion of the gene encoding H-FABP/MDGI has no overt effects in the mammary gland. Transgenic Res 9: 439–444.

    CAS  Article  Google Scholar 

  9. Czuchra A, Meyer H, Legate KR, Brakebusch C, Fässler R . (2006). Genetic analysis of beta1 integrin ″activation motifs″ in mice. J Cell Biol 174: 889–899.

    CAS  Article  Google Scholar 

  10. Gietz RD, Woods RA . (2002). Transformation of yeast by lithium acetate/single-stranded carrier DNA/polyethylene glycol method. Methods Enzymol 350: 87–96.

    CAS  Article  Google Scholar 

  11. Harburger DS, Bouaouina M, Calderwood DA . (2009). Kindlin-1 and −2 directly bind the C-terminal region of beta integrin cytoplasmic tails and exert integrin-specific activation effects. J Biol Chem 284: 11485–11497.

    CAS  Article  Google Scholar 

  12. Hertzel AV, Bernlohr DA . (2000). The mammalian fatty acid-binding protein multigene family: molecular and genetic insights into function. Trends Endocrinol Metab 11: 175–180.

    CAS  Article  Google Scholar 

  13. Huttenlocher A, Sandborg RR, Horwitz AF . (1995). Adhesion in cell migration. Curr Opin Cell Biol 7: 697–706.

    CAS  Article  Google Scholar 

  14. Huynh H, Alpert L, Pollak M . (1996). Silencing of the mammary-derived growth inhibitor (MDGI) gene in breast neoplasms is associated with epigenetic changes. Cancer Res 56: 4865–4870.

    CAS  PubMed  Google Scholar 

  15. Huynh H, Pollak M . (1997). Stabilization of mammary-derived growth inhibitor messenger RNA by antiestrogens. Clin Cancer Res 3: 2151–2156.

    CAS  PubMed  Google Scholar 

  16. Hynes RO . (2002). Integrins: bidirectional, allosteric signaling machines. Cell 110: 673–687.

    CAS  Article  Google Scholar 

  17. Jarvius M, Paulsson J, Weibrecht I, Leuchowius KJ, Andersson AC, Wählby C et al. (2007). In situ detection of phosphorylated platelet-derived growth factor receptor beta using a generalized proximity ligation method. Mol Cell Proteomics 6: 1500–1509.

    CAS  Article  Google Scholar 

  18. Kiosses WB, Shattil SJ, Pampori N, Schwartz MA . (2001). Rac recruits high-affinity integrin alphavbeta3 to lamellipodia in endothelial cell migration. Nat Cell Biol 3: 316–320.

    CAS  Article  Google Scholar 

  19. Kloeker S, Major MB, Calderwood DA, Ginsberg MH, Jones DA, Beckerle MC . (2004). The kindler syndrome protein is regulated by transforming growth factor-beta and involved in integrin-mediated adhesion. J Biol Chem 279: 6824–6833.

    CAS  Article  Google Scholar 

  20. Koistinen P, Pulli T, Uitto VJ, Nissinen L, Hyypi T, Heino J . (1999). Depletion of alphaV integrins from osteosarcoma cells by intracellular antibody expression induces bone differentiation marker genes and suppresses gelatinase (MMP-2) synthesis. Matrix Biol 18: 239–251.

    CAS  Article  Google Scholar 

  21. Kurtz A, Vogel F, Funa K, Heldin CH, Grosse R . (1990). Developmental regulation of mammary-derived growth inhibitor expression in bovine mammary tissue. J Cell Biol 110: 1779–1789.

    CAS  Article  Google Scholar 

  22. Lau TL, Kim C, Ginsberg MH, Ulmer TS . (2009). The structure of the integrin alphaIIbbeta3 transmembrane complex explains integrin transmembrane signalling. EMBO J 28: 1351–1361.

    CAS  Article  Google Scholar 

  23. Legate KR, Wickström SA, Fässler R . (2009). Genetic and cell biological analysis of integrin outside-in signaling. Genes Dev 23: 397–418.

    CAS  Article  Google Scholar 

  24. Lehtonen JV, Still DJ, Rantanen VV, Ekholm J, Björklund D, Iftikhar Z et al. (2004). BODIL: a molecular modeling environment for structure-function analysis and drug design. J Comput Aided Mol Des 18: 401–419.

    CAS  Article  Google Scholar 

  25. Lundin J, Lundin M, Holli K, Kataja V, Elomaa L, Pylkkänen L et al. (2001). Omission of histologic grading from clinical decision making may result in overuse of adjuvant therapies in breast cancer: results from a nationwide study. J Clin Oncol 19: 28–36.

    CAS  Article  Google Scholar 

  26. Luo BH, Springer TA, Takagi J . (2004). A specific interface between integrin transmembrane helices and affinity for ligand. PLoS Biol 2: e153.

    Article  Google Scholar 

  27. Luque A, Gomez M, Puzon W, Takada Y, Sanchez-Madrid F, Cabanas C . (1996). Activated conformations of very late activation integrins detected by a group of antibodies (HUTS) specific for a novel regulatory region (355-425) of the common beta 1 chain. J Biol Chem 271: 11067–11075.

    CAS  Article  Google Scholar 

  28. Lücke C, Rademacher M, Zimmerman AW, van Moerkerk HT, Veerkamp JH, Rüterjans H . (2001). Spin-system heterogeneities indicate a selected-fit mechanism in fatty acid binding to heart-type fatty acid-binding protein (H-FABP). Biochem J 354: 259–266.

    Article  Google Scholar 

  29. Mattila E, Pellinen T, Nevo J, Vuoriluoto K, Arjonen A, Ivaska J . (2005). Negative regulation of EGFR signalling through integrin-alpha1beta1-mediated activation of protein tyrosine phosphatase TCPTP. Nat Cell Biol 7: 78–85.

    CAS  Article  Google Scholar 

  30. Millon-Frémillon A, Bouvard D, Grichine A, Manet-Dupé S, Block MR, Albiges-Rizo C . (2008). Cell adaptive response to extracellular matrix density is controlled by ICAP-1-dependent beta1-integrin affinity. J Cell Biol 180: 427–441.

    Article  Google Scholar 

  31. Nevo J, Mattila E, Pellinen T, Yamamoto DL, Sara H, Iljin K et al. (2009). Mammary-derived growth inhibitor alters traffic of EGFR and induces a novel form of cetuximab resistance. Clin Cancer Res 15: 6570–6581.

    CAS  Article  Google Scholar 

  32. Pellinen T, Arjonen A, Vuoriluoto K, Kallio K, Fransen JA, Ivaska J . (2006). Small GTPase Rab21 regulates cell adhesion and controls endosomal traffic of beta1-integrins. J Cell Biol 173: 767–780.

    CAS  Article  Google Scholar 

  33. Petersen OW, Ronnov-Jessen L, Howlett AR, Bissell MJ . (1992). Interaction with basement membrane serves to rapidly distinguish growth and differentiation pattern of normal and malignant human breast epithelial cells. Proc Natl Acad Sci USA 89: 9064–9068.

    CAS  Article  Google Scholar 

  34. Shi X, Ma YQ, Tu Y, Chen K, Wu S, Fukuda K et al. (2007). The MIG-2/integrin interaction strengthens cell-matrix adhesion and modulates cell motility. J Biol Chem 282: 20455–20466.

    CAS  Article  Google Scholar 

  35. Söderberg O, Gullberg M, Jarvius M, Ridderstrale K, Leuchowius KJ, Jarvius J et al. (2006). Direct observation of individual endogenous protein complexes in situ by proximity ligation. Nat Methods 3: 995–1000.

    Article  Google Scholar 

  36. Tadokoro S, Shattil SJ, Eto K, Tai V, Liddington RC, de Pereda JM et al. (2003). Talin binding to integrin beta tails: a final common step in integrin activation. Science 302: 103–106.

    CAS  Article  Google Scholar 

  37. Takala H, Nurminen E, Nurmi SM, Aatonen M, Strandin T, Takatalo M et al. (2008). Beta2 integrin phosphorylation on Thr758 acts as a molecular switch to regulate 14-3-3 and filamin binding. Blood 112: 1853–1862.

    CAS  Article  Google Scholar 

  38. Tuomi S, Mai A, Nevo J, Laine JO, Vilkki V, hman TJ et al. (2009). PKCepsilon regulation of an alpha5 integrin-ZO-1 complex controls lamellae formation in migrating cancer cells. Sci Signal 2: ra32.

    Article  Google Scholar 

  39. Wang F, Hansen RK, Radisky D, Yoneda T, Barcellos-Hoff MH, Petersen OW et al. (2002). Phenotypic reversion or death of cancer cells by altering signaling pathways in three-dimensional contexts. J Natl Cancer Inst 94: 1494–1503.

    CAS  Article  Google Scholar 

  40. Wang HL, Kurtz A . (2000). Breast cancer growth inhibition by delivery of the MDGI-derived peptide P108. Oncogene 19: 2455–2460.

    CAS  Article  Google Scholar 

  41. Wang W, Goswami S, Lapidus K, Wells AL, Wyckoff JB, Sahai E et al. (2004). Identification and testing of a gene expression signature of invasive carcinoma cells within primary mammary tumors. Cancer Res 64: 8585–8594.

    CAS  Article  Google Scholar 

  42. Weaver VM, Petersen OW, Wang F, Larabell CA, Briand P, Damsky C et al. (1997). Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies. J Cell Biol 137: 231–245.

    CAS  Article  Google Scholar 

  43. Wegener KL, Partridge AW, Han J, Pickford AR, Liddington RC, Ginsberg MH et al. (2007). Structural basis of integrin activation by talin. Cell 128: 171–182.

    CAS  Article  Google Scholar 

  44. White DE, Kurpios NA, Zuo D, Hassell JA, Blaess S, Mueller U et al. (2004). Targeted disruption of beta1-integrin in a transgenic mouse model of human breast cancer reveals an essential role in mammary tumor induction. Cancer Cell 6: 159–170.

    CAS  Article  Google Scholar 

  45. Vinogradova O, Vaynberg J, Kong X, Haas TA, Plow EF, Qin J . (2004). Membrane-mediated structural transitions at the cytoplasmic face during integrin activation. Proc Natl Acad Sci USA 101: 4094–4099.

    CAS  Article  Google Scholar 

  46. Yao ES, Zhang H, Chen YY, Lee B, Chew K, Moore D et al. (2007). Increased beta1 integrin is associated with decreased survival in invasive breast cancer. Cancer Res 67: 659–664.

    CAS  Article  Google Scholar 

  47. Zhang X, Jiang G, Cai Y, Monkley SJ, Critchley DR, Sheetz MP . (2008). Talin depletion reveals independence of initial cell spreading from integrin activation and traction. Nat Cell Biol 10: 1062–1068.

    CAS  Article  Google Scholar 

Download references


We thank Professor J Ylänne and Professor M Salmi for critically reviewing the paper; J Siivonen, H Marttila, O Levälampi and P Terho for the excellent technical assistance. This study has been supported by Academy of Finland, ERC Starting Grant, Sigrid Juselius Foundation, EMBO YIP and Finnish Cancer Organizations. JN has been supported by Maud Kuistila Memorial Foundation and Turku University Foundation. JN, ST and TP were supported by Turku Graduate School of Biomedical Sciences and AM by Drug Discovery Graduate School.

Author information



Corresponding author

Correspondence to J Ivaska.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Nevo, J., Mai, A., Tuomi, S. et al. Mammary-derived growth inhibitor (MDGI) interacts with integrin α-subunits and suppresses integrin activity and invasion. Oncogene 29, 6452–6463 (2010).

Download citation


  • integrin
  • MDGI
  • breast cancer
  • adhesion
  • migration
  • invasion

Further reading


Quick links