Main

About 97 million adults in the United States are overweight or obese, which substantially increases the risk of cancer and morbidity.1 The predominant cancers associated with obesity include endometrial, breast, prostate, and colon.2, 3, 4 Among these cancers, breast cancer is the most frequently diagnosed cancer in woman. High levels of triglycerides in the blood and free fatty acids (FFAs) are linked with obesity. Increased plasma concentrations of FFAs are associated with breast cancer.5, 6 Diverse effects of FFAs on breast cancer growth have been observed in cell culture and in epidemiological studies of mice xenograft models.7, 8, 9 However, the molecular mechanisms underlying the role of FFAs in breast cancer pathogenesis remain unknown.

Emerging clinical studies have implicated plasminogen activator inhibitor-1 (PAI-1) both in tumor tissues and in plasma as a key prognostic factor in breast cancer. PAI-1 is the primary inhibitor of urokinase-type (uPA) and tissue-type (tPA) plasminogen activators and belongs to the family of serine protease inhibitors. Recent studies showed that PAI-1 is dramatically upregulated in obesity, and increased FFAs concentrations in obesity are positively correlated with PAI-1 levels in plasma,10, 11 suggesting that PAI-1 may be important in mediating the effects of FFAs. PAI-1 is released into the circulation and extracellular spaces under physiological conditions by only a few cells: liver, smooth muscle, adipocytes, and platelets. However, under pathological conditions, tumor cells and endothelial cells secrete abundant PAI-1.12

In cohort study, several components of the plasminogen activator system (uPA, the receptor uPAR (CD87), PAI-1, and PAI-2) are potential predictors of relapse-free survival and overall survival in 2780 patients with primary invasive breast cancer.13 In another study, PAI-1 level, which was determined in tumor tissue extracts by immunoenzymatic methods, was reported as a significant prognostic marker both in the primary breast cancer and for survival after long-term median follow-up of 77 months.14 Although initial studies focused on PAI-1 levels in tumor lysates and cytosols, tissue findings were then extended to plasma studies and similar results were seen. Specifically, research study showed increased levels of PAI-1 above the normal cut-off in 24% of breast cancer plasmas, 30% of colon cancer plasmas, 40% of lung cancer plasmas, and 12% of prostate cancer plasmas.15 More recently, it was shown that metastatic breast cancer patients who had elevated pre-treatment plasma PAI-1 had a shorter overall survival than metastatic breast cancer patients with normal levels of PAI-1.16 Moreover, a common polymorphism (4G/5G) in the promoter region of the PAI-1 gene has been reported to influence transcription and plasma levels of PAI-1. Studies of different populations have shown consistently that subjects homozygous for the 4G allele have significantly higher plasma PAI-1 levels than those homozygous for the 5G allele. Patients homozygous for the 4G allele had significantly poorer disease-free survival and overall survival than those homozygous for the 5G allele, suggesting that implication of high plasma PAI-1 level in poor prognosis of breast cancer patients.17

We previously demonstrated that FFAs induced PAI-1 expression in human hepatoma HepG2 cells.18, 19 The data reported here represent the first effort to understand the link between elevation of FFAs and PAI-1 in the pathogenesis of breast cancer. PAI-1 expression is higher in invasive MDA-MB-231 cells compared with less invasive and non-metastatic Hs578T cells. FFAs induced the expression of PAI-1 in breast cancer cells, which, in turn, contributes to increased migration. Further, FFA-induced PAI-1 expression was mediated by the transcription factor SMAD4. These studies provide insight into understanding the mechanism underlying FFAs and PAI-1 in regulation of breast cancer progression, and will facilitate exploring the application of drugs that are known to lower FFAs and PAI-1 for breast cancer therapy.

MATERIALS AND METHODS

Cell Culture

The human breast cancer cell lines MDA-MB-231, Hs578T, and MDA-MB-468 were obtained from American Type Culture Collection (ATCC no. HTB-26, HTB-126, and HTB-132, respectively). MDA-MB-231 was cultured at 37°C with 5% CO2 in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with penicillin (50 U/ml), streptomycin (50 μg/ml), and 10% fetal bovine serum. Hs578T was maintained at 37°C with 10% CO2 in DMEM supplemented with penicillin, streptomycin, insulin (10 μg/ml), and 10% fetal bovine serum. MDA-MB-468 was kept at 37°C with 5% CO2 in RPMI 1640 supplemented with penicillin, streptomycin, insulin–transferrin–selenium (ITS; GIBCO-BRL, NY, USA), and 10% fetal bovine serum. Experiments were performed in Dulbecco's Modified Eagles media with Hams' nutrient mixture F12 (DME/F-12).

FFA Preparation and Treatment

FFA preparation and treatment of cells was performed as described previously by us.20 One percent fatty acid-free bovine serum albumin (FAF-BSA, fraction V; Sigma) was used as a vehicle for the FFA treatment. Sodium salts of FFAs (Sigma) were dissolved in water at certain temperatures (eg, linoleic acid at 55°C and stearic acid at 80°C) and then added dropwise to 1% FAF-BSA in DME/F12. After adjusting to pH 7.4 and filtering through a 0.22-μm filter, all FFAs were verified using Wako NEFA C kits (Biochemical Diagnostics). For FFA treatment, 80% confluent cells were serum-starved and subsequently exposed to control condition or BSA-bound FFAs.

Western Blot Analysis

Cells were seeded in six-well dishes at 1.5 × 105 cells/well and incubated for 24 h. After a 24-h period of serum starvation in medium containing 1% BSA, cells were incubated with or without BSA-bound fatty acid solution for 24 h. After treatment, cells were washed and lysed in protein extraction buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, 1 mM PMSF, and Complete Protease Inhibitor Cocktail (Roche, Mannheim, Germany). Lysates were cleared by centrifugation and protein concentrations of the supernatants were determined using the BCA protein assay (Pierce, IL, USA). Equal amounts of protein (20 μg per lane) were separated by SDS-PAGE and transferred to PVDF membrane. Immunoblotting was performed according to the manufacturer's instructions using monoclonal anti-PAI-1 antibodies (IMA-33H1F7; Molecular Innovations Inc., MI, USA) and a chemiluminescence system (Amersham Biosciences, Uppsala, Sweden). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Protein band densities were determined by densitometric analysis.

ELISA Analysis

PAI-1 protein concentration in cell culture media was determined using the PAI-1 enzyme-linked immunosorbent assay (ELISA) kit (TintElize PAI-1; Biopool, Umeå, Sweden) according to the manufacturer's recommendations.

RT-PCR Analysis

Total RNA was isolated from breast cancer cells using Trizol (Invitrogen, CA, USA). RNA content and purity were determined by photometry at 260 vs 280 nm. A 1 μg weight of total RNA was incubated with 20 pmol of oligo-dT primer for 2 min at 70°C and the mixture was subsequently cooled on ice for 5 min. Reverse transcription was carried out using the Advantage RT-for-PCR kit (Clontech, CA, USA) for 1 h at 42°C, with an additional 5 min at 94°C for inactivation of the reverse transcriptase. PCR was performed with the iCycler thermal cycler (Bio-Rad, CA, USA) using the Titanium Taq PCR kit (Clontech). The density of bands was analyzed with the use of densitometry and Kodak software (Eastman Kodak).

Quantitative Reverse Transcriptase Real-Time PCR

Relative quantification of PAI-1 mRNA was performed with the MyiQ Single-Color Real-Time PCR Detection System (Bio-Rad, CA, USA). The expression of PAI-1 and GAPDH mRNA was quantified by fluorescent SYBR green technology (Bio-Rad) as described previously.21 PCR were performed using specific primers for human PAI-1 (forward 5′-CATCCCCCATCCTACGTGG and reverse 5′-CCCCATAGGGTGAGAAAACCA) and human GAPDH (forward 5′-ATGGGGAAGGTGAAGGTCG and reverse 5′-GGGGT CATTGATGGCAACAATA). The mRNA level of PAI-1 was normalized by those of GAPDH used as an endogenous internal control. Thermal cycling conditions were 3 min at 94°C followed by 40 cycles of three-step PCR consisting of 15 s at 94°C, 15 s at 55°C, and 15 s at 68°C. All samples were amplified in triplicate.

In Vitro Invasion Assay

Matrigel matrix (Becton Dickinson Labware, NJ, USA) was polymerized in 24-well 6.5-mm Transwell inserts containing polycarbonate membranes with 8-μm pores to create invasion chambers as directed by the supplier. MDA-MB-231 cells and Hs578T cells were serum-starved for 24 h, harvested by trypsinization after 24-h FFA treatment, and subsequently washed twice in DMEM without added serum or proteinase inhibitors. The cells were suspended in DME/F-12 at 3.5 × 105/ml. Serum-free DME/F-12 (0.6 ml) was added to each plate well and 0.1 ml (3.5 × 104 cells) of the cell suspension was added to each insert. Assays were performed in triplicate wells for each condition. The plates of inserts were incubated for 8 h at 37°C. After incubation, the chambers were processed and stained with Diff-Quick staining kit as directed by the supplier (IMEB Inc., CA, USA). Cells were counted in four fields per chamber under × 100 magnification with the aid of a ruled grid.

Generation of Breast Cancer Cells with PAI-1 Knockdown

Adeno-associated virus (AAV) plasmid carrying antisense against PAI-1 mRNA was generated by subcloning cDNA encoding PAI-1 (provided by Dr Paul Declerck, University of Belgium) into pSub201 backbone. This plasmid carries a gene encoding green fluorescent protein (GFP) that has been used as a marker to monitor the cells after transfection. After confirmation by sequencing, the AAV–anti-PAI-1 plasmid was packaged into HEK 293 cells using pDG helper plasmid to produce recombinant AAV.22 MDA-MB-231 cells were then transduced with the AAV–anti-PAI-1 viruses or the control viruses, AAV–GFP. Transduced cells were sorted by their GFP expression with the use of flow cytometry and a stable clone with PAI-1 gene knocked down was identified by western blot analysis.

Electrophoretic Mobility-Shift Assay

Electrophoretic mobility-shift assay (EMSA) was performed by using 5 μg of nuclear extracts prepared from untreated cells or cells treated with linoleic acid for 24 h and a probe (5′-TCGAG AGCCAGACA AAA AGCCAGACA TTT AGCCAGACA C-3′), which contains the consensus SMAD-binding sites that are underlined.23 The oligonucleotide was radioactively end-labeled with [γ-32P] ATP and T4 polynucleotide kinase (New England Biolabs). Binding reactions were allowed to proceed at room temperature for 20 min. Protein–DNA complexes were resolved by electrophoresis on non-denaturing 5% polyacrylamide gels and were visualized by autoradiography. The specificity of binding interactions was assessed by competition with a 200 molar excess of unlabeled double-stranded oligonucleotide of identical sequence. For supershift analysis of SMAD4, 2 μg of SMAD4 (C-20-X) antibody from Santa Cruz Biotechnology Inc. was used.

Statistical Analysis

Data are presented as means±s.d. Differences in data between the groups were compared with Student's paired two-tailed t-test or one-way ANOVA where appropriate. A P-value less than 0.05 was considered statistically significant.

RESULTS

FFAs Increase the Expression of PAI-1 in Breast Cancer Cells

The level of PAI-1 as a prognostic marker in primary breast cancer has been well documented.13, 14, 15, 16 To determine whether the expression of PAI-1 may contribute to the invasiveness of breast cancer cells, we measured the secreted levels of PAI-1 in two breast cancer cell lines, Hs578T and MDA-MB-231. As determined by ELISA, the secreted levels of PAI-1 were higher in the highly metastatic MDA-MB-231 cells as compared with that in the less invasive Hs578T cells (fold=5.7±0.8 for MDA-MB-231 vs Hs578T, n=4, P=0.002; Figure 1).

Figure 1
figure 1

Secreted PAI-1 levels in MDA-MB-231 and Hs578T cells. Cells were seeded in six-well dishes at 1.5 × 105 cells/well and incubated for 24 h. After a 24-h period of serum starvation in medium containing 1% BSA, cells at 90% confluency were incubated with or without BSA-bound fatty acid solution for 24 h. Secreted PAI-1 expression levels in the media of MDA-MB-231 and Hs578T cells after 24-h serum starvation were analyzed by ELISA assay. ELISA results represent the means±s.d. of four independent experiments; *n=4, P=0.002 by Student's t-test.

Full size image

The effect of FFAs on the expression of PAI-1 in MDA-MB-231 cells was determined using FFAs with different saturation and chain lengths (Figure 2). We found that linoleic acid and oleic acid (to a lesser extent) induced the expression of intracellular PAI-1 protein in MDA-MB-231, whereas other FFAs had no effect on the expression of PAI-1. Treatment with linoleic acid for 24 h resulted in the greatest induction of intracellular PAI-1 in MDA-MB-231 cells (fold=1.8±0.2 for linoleic acid compared with control, n=4, P=0.004; Figure 2a). Secretion levels of PAI-1 were determined in MDA-MB-231 cells treated with FFAs for 24 h by ELISA assay. Linoleic acid induced the expression of secreted PAI-1, whereas other FFAs showed no effect on the secretion of PAI-1 (fold=2.2±0.3 for linoleic acid compared with control, n=3, P=0.0016; Figure 2b). We further determined that PAI-1 mRNA was increased in MDA-MB-231 cells after exposure to linoleic acid and oleic acid for 24 h as determined by RT-PCR (fold=1.4±0.3 for linoleic acid, n=3, P<0.01; Figure 2c). The induction of PAI-1 mRNA by linoleic acid was confirmed by quantitative reverse transcriptase real-time PCR (fold=1.5±0.5 compared with control, n=3, P=0.01; Figure 2d).

Figure 2
figure 2

FFAs increased the expression of PAI-1 in breast cancer cells. Cells were seeded in six-well dishes at 1.5 × 105 cells/well and incubated for 24 h. After a 24-h period of serum starvation in medium containing 1% BSA, cells at 90% confluency were incubated with or without BSA-bound fatty acid solution for 24 h. (a) Intracellular PAI-1 expression levels in cell lysates of MDA-MB-231 after 24-h exposure to 500 μM FFAs was analyzed by western blot. (lane 1: control; lane 2: palmitic acid; lane 3: palmitoleic acid; lane 4: stearic acid; lane 5: oleic acid; lane 6: linoleic acid). (b) Secreted PAI-1 expression levels in the media of MDA-MB-231 after 24-h exposure to 500 μM FFAs were determined by ELISA assay. Results represent the means±s.d. of three independent experiments; *n=3, P=0.002 by ANOVA. (c) PAI-1 mRNA expression levels in MDA-MB-231 after 24-h exposure to 500 μM FFAs were analyzed by RT-PCR (lane 1: control; lane 2: palmitic acid; lane 3: palmitoleic acid; lane 4: stearic acid; lane 5: oleic acid; lane 6: linoleic acid). (d) PAI-1 mRNA expression levels in MDA-MB-231 after 24-h exposure to 500 μM linoleic acid was analyzed by quantitative reverse transcriptase real-time PCR. Results represent the means±s.d. of three independent experiments; *n=3, P=0.01 by Student's t-test.

Full size image

FFA-Induced In Vitro Migration of MDA-MB-231 Cells

Differential effects of FFAs on breast cancer cell growth and invasion have been demonstrated in previous reports.24, 25 To investigate the effect of FFAs on the migration of MDA-MB-231 cells and Hs578T cells, an in vitro invasion assay using Matrigel-coated filters was performed. Results showed that linoleic acid induced the migration of MDA-MB-231 cells (fold=4.0±1.1 compared with control, n=3, P<0.001). Oleic acid showed less effect on cell migration (fold=1.9±0.1 compared with control, n=3, P<0.01) and other FFAs had no effect (Figure 3a). While, Hs578T cells showed no significant increase in migration by all FFAs treated, indicating that induced PAI-1 expression by FFAs might be required for breast cancer cell migration in vitro (Figure 3b).

Figure 3
figure 3

FFAs induced in vitro migration of MDA-MB-231. Matrigel matrix polymerized in 24-well 6.5-mm Transwell inserts containing polycarbonate membranes with 8-μm pores to create invasion chambers. Cells were serum-starved for 24 h, harvested by trypsinization after 24-h FFA treatment and subsequently washed twice in DMEM without added serum or proteinase inhibitors. Cells were suspended in DME/F-12 at 3.5 × 105/ml. Serum-free DME/F-12 (0.6 ml) was added to each plate well and 0.1 ml (3.5 × 104 cells) of the cell suspension was added to each insert. Induction of migration of MDA-MB-231 cells (a) and Hs578T cells (b) after 24-h exposure to 500 μM FFAs was analyzed by in vitro Matrigel assay. Results represent the means±s.d. of three independent experiments; *n=3, P<0.01, **n=3, P<0.009 and ***n=3, not significant by ANOVA.

Full size image

Inhibition of PAI-1 by Antisense mRNA Blocked FFA-Induced Migration of MDA-MB-231 Cells

The specificity of and requirement for PAI-1 in mediating FFA-regulated breast cancer cell migration was characterized with a MDA-MB-231 derivative in which PAI-1 expression was knocked down by AAV carrying anti-PAI-1.22 Immunoblotting analysis was performed to verify reduced expression of PAI-1 following AAV–anti-PAI-1 viral transduction into MDA-MB-231. Linoleic acid-induced expression of PAI-1 protein in cells infected with AAV–anti-PAI-1 viruses was decreased when compared with the cells infected with control viruses (Figure 4a).

Figure 4
figure 4

Inhibition of PAI-1 by antisense blocked FFA-induced migration of MDA-MB-231. (a) MDA-MB-231 cells transduced with control AAV–GFP or AAV–anti-PAI-1 virus were treated with 500 μM linoleic acid for 24 h. Intracellular expression of PAI-1 was analyzed by western blot. (b) In vitro Matrigel assay following AAV–anti-PAI-1 viral transduction into MDA-MB-231. Results represent the means±s.d. of three independent experiments; *n=3, not significant and **n=3, P=0.0013 by ANOVA.

Full size image

Further, we determined whether expression of PAI-1 contributes to the effect of linoleic acid on the migration of breast cancer cells. Linoleic acid-induced migration was blocked in MDA-MB-231 cells transduced with AAV–anti-PAI-1, indicating a critical role of PAI-1 in FFA-induced migration of MDA-MB-231 cells (Figure 4b).

SMAD4 Mediated FFA-Induced Expression of PAI-1 in Breast Cancer Cells

We previously identified an FFA-response element within the 5′ flanking region of the PAI-1 gene in HepG2 cells. Sequence analysis revealed that this region contains a SMAD4-binding element.20 To investigate the role of SMAD4 in FFA-induced PAI-1 expression in MDA-MB-231, we performed an EMSA. The SMAD-binding activity was markedly increased in linoleic acid-treated cells and it was inhibited when unlabeled SMAD4 oligonucleotide was added to the binding reaction (Figure 5a). Using an antibody specific for SMAD4, we were able to determine a higher-molecular-weight band, indicating the presence of SMAD4 in the binding complex. This result suggests that linoleic acid activates SMAD4 signaling in MDA-MB-231 cells that contribute to the increase in PAI-1 expression by linoleic acid.

Figure 5
figure 5

SMAD4 mediated FFA-induced expression of PAI-1in breast cancer cells. (a) Increased binding of SMAD4 in nuclear extracts of MDA-MB-231 treated with 500 μM linoleic acid for 24 h was identified by EMSA. (b) Intracellular PAI-1 expression levels in cell lysates of MDA-MB-468 after 24-h exposure to 500 μM FFAs were analyzed by western blot. Results indicate the representative image of three independent experiments (lane 1: control; lane 2: palmitic acid; lane 3: palmitoleic acid; lane 4: stearic acid; lane 5: oleic acid; lane 6: linoleic acid).

Full size image

Furthermore, we characterized the role of SMAD4 in FFA-induced expression of PAI-1 using SMAD4-deficient MDA-MB-468 cells, which are breast carcinoma cells with a homozygous deletion for SMAD4.26 Unlike MDA-MB-231 cells, we found that the increased expression of PAI-1 by any FFA was not detected in the SMAD4-deficient cells, which confirmed that FFAs induce the expression of PAI-1 through SMAD4 signaling in breast cancer cells (Figure 5b).

DISCUSSION

Increased PAI-1 and FFAs are both associated with poor prognosis of breast cancer. We hypothesized that there was a link between the elevation of FFAs and PAI-1 levels, which contributes to breast cancer pathogenesis. It has been postulated that higher consumption of n-6 PUFA, particularly linoleic acid, in the diet may contribute to the high breast cancer incidence in Western women.27 Although both FFA and PAI-1 levels are generally elevated in breast cancer patients, there is no clinical link between plasma FFAs and PAI-1 levels in breast cancer patients. Our results demonstrate that PAI-1 level is increased by FFAs, especially linoleic acid, in breast cancer cells in vitro, thus establishing a connection between increased FFA levels and PAI-1 in breast cancer cells.

To begin addressing this hypothesis, we investigated the effect of five of the most abundant FFAs in plasma on the expression of PAI-1 and breast cancer cell migration. As shown previously,28 linoleic acid had the greatest effect on the induction of PAI-1 in MDA-MB-231 cells, consistent with our previous observations that linoleic acid increased the expression of PAI-1 in HepG2 cells.18, 19 It was also reported that butyric acid inhibited both cell proliferation and uPA activity, but had no effect on PAI-1 level in MDA-MB-231 cells,29 again suggesting that FFAs with different saturation and chain lengths show different effects on the expression of PAI-1. A concentration of 500 μM was selected for individual FFAs because our previous studies demonstrated that this concentration of FFAs induces PAI-1 expression in HepG2 cells.18, 19 In addition, total serum FFA concentrations of at least 5 mM have been reported in humans30 which makes 500 μM a physiologically attainable concentration, if not normal.31

Previously, we have reported that FFAs induce PAI-1 expression in HepG2 cells by activating a transcription factor that binds with the sequence 5′-TG(G/C)1-2CTG-3′, which is repeated four times in the PAI-1 promoter (−528 and −599).20 One of the SMAD3/SMAD4-binding sequences termed CAGA boxes, which are essential and sufficient for the induction of PAI-1 by TGF-β1,23 has been identified in the PAI-1 promoter −580 within the FFA-responsive region. We showed by EMSA that SMAD4 is activated by treatment with linoleic acid of MDA-MB-231 cells (Figure 5a). Moreover, we found that linoleic acid failed to increase the expression of PAI-1 in SMAD4-deficient MDA-MB-468 cells (Figure 5b), suggesting that SMAD4 mediates linoleic acid-induced PAI-1 expression in breast cancer cells.

PAI-1 can potentially influence cell motility at several levels by competing with uPAR and integrins to bind vitronectin and/or by inducing the internalization of adhesion receptors.32 TGF-β1-induced PAI-1 has been reported to be associated with its stimulatory effect on invasion and metastasis of MDA-MB-231 cells.29 Incubation of endothelial cells with conditioned medium from two breast cancer cell lines (MCF7 and MDA-MB-231) also demonstrated that conditioned medium from highly invasive MDA-MB-231 cells increased the tube formation of endothelial cells, which was related to the high levels of uPA and PAI-1 secreted by MDA-MB-231 cells as compared with low levels in MCF7 cells.33 Furthermore, downregulation of the expression of the PAI-1 gene in MDA-MB-231 cells was associated with inhibition of invasion, indicating that the expression of PAI-1 may be associated with the invasion of MDA-MB-231 cells.34 Therefore, our results, demonstrating that the highly invasive MDA-MB-231 cells expressed greater amounts of secreted PAI-1 compared with less invasive Hs578T cells (Figure 1), are consistent with these previous observations that secreted PAI-1 might contribute to the invasion and metastasis of breast cancer. The effect of PAI-1 induced by FFAs on breast cancer cell migration was further investigated. The induction of PAI-1 by linoleic acid in MDA-MB-231 cells paralleled the linoleic acid-induced increase in cell migration (Figure 3a), suggesting that PAI-1 induced by linoleic acid in MDA-MB-231 might be responsible for inducing migration of the cancer cells in vitro. By inhibiting PAI-1 expression in MDA-MB-231 cells, we demonstrated that PAI-1 expression is necessary for linoleic acid-induced MDA-MB-231 breast cancer cell migration (Figure 4b). Moreover, less invasive Hs578T cells (Figure 3b) and SMAD4-deficient MDA-MB-468 cells (data not shown) showed no significant increase in migration, indicating the importance of SMAD4-mediated PAI-1 expression by linoleic acid on breast cancer cell migration.

A previous dietary study also demonstrated that a linoleic acid-enriched diet (12% compared with 2% normal) increases the incidence of microscopic metastases in MDA-MB-231 cell tumor-bearing nude mice (68 vs 42%). While linoleic acid did not affect the pro-migratory and pro-invasive factors, type IV collagenase and gelatinase in MDA-MB-231 cells,35 it has been shown to stimulate MDA-MB-231 cell growth36 and to increase MDA-MB-435 breast cancer cell invasion in vitro in a cyclooxygenase-dependent manner.37 Also, a direct correlation was demonstrated between COX-2 and PAI-1 immunohistochemical expression in a series of breast cancer cases,38 suggesting a putative link between expression of PAI-1 and linoleic acid-induced COX-2 signaling.

In summary, linoleic acid stimulates PAI-1 expression and increases the invasion of MDA-MB-231 breast cancer cells through the SMAD4 signaling pathway. As mentioned above, the effect of FFAs, particularly linoleic acid, on the growth and invasion of human breast cancer cells was examined in terms of eicosanoid synthesis catalyzed by cyclooxygenase and lipooxygenase.37 For the first time, we identified the importance of transcriptional activation of PAI-1 mediated by FFA–SMAD4 signaling in human breast cancer cell invasion in vitro. It was previously shown that PAI-1 deficiency in mice retarded fibrosarcoma development.39 It will be also interesting to investigate the effect of linoleic acid-enriched diet on implanted human breast cancer cell growth and invasion in PAI-1-deficient mice. Our study suggests that PAI-1 could be a target for treatment for breast cancer by novel dietary and/or pharmaceutical strategies to inhibit tumor invasion and metastasis.