Protein-tyrosine Phosphatase PTPL1/FAP-1 Triggers Apoptosis in Human Breast Cancer Cells*

Studies in Jurkat leukemia cells have suggested that protein-tyrosine phosphatase PTPL1/FAP-1 rescues Fas-induced cell death. However, we have previously shown that this enzyme triggers 4-hydroxytamoxifen-induced growth inhibition in human breast cancer cells. The present study addresses the role of PTPL1/FAP-1 in an-tiestrogen-regulated apoptotic effect and insulin-like growth factor-I survival action in MCF7 cells and further identifies the impacted signaling pathway. By terminal deoxynucleotidyltransferase-mediated dUTP-bio-tin nick end labeling and cytoplasmic nucleosome enzyme-linked immunosorbent assay, we demonstrated that 4-hydroxytamoxifen-induced apoptosis was totally lost in PTPL1/FAP-1 antisense transfectants in which enzyme expression was abrogated, revealing the crucial role of this phosphatase in the apoptotic process in human breast cancer cells. Time-dependent expression of PTPL1/FAP-1 in MCF7 cells completely abolished the survival action of insulin-like growth factor-I. This effect occurred through a highly significant reduction in phosphatidylinositol 3-kinase/Akt pathway activation (80% reduction in phosphatidylinositol 3-kinase activity, 55% inhibition of Akt activation) accompanied by a 65% decrease in insulin receptor substrate-1 growth fac-tor-induced tyrosine phosphorylation.


Studies in Jurkat leukemia cells have suggested that protein-tyrosine phosphatase PTPL1/FAP-1 rescues Fasinduced cell death. However, we have previously shown that this enzyme triggers 4-hydroxytamoxifen-induced growth inhibition in human breast cancer cells. The present study addresses the role of PTPL1/FAP-1 in antiestrogen-regulated apoptotic effect and insulin-like growth factor-I survival action in MCF7 cells and further identifies the impacted signaling pathway. By terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling and cytoplasmic nucleosome enzyme-linked immunosorbent assay, we demonstrated that 4-hydroxytamoxifen-induced apoptosis was totally lost in PTPL1/FAP-1 antisense transfectants in which enzyme expression was abrogated, revealing the crucial role of this phosphatase in the apoptotic process in human breast cancer cells. Time-dependent expression of PTPL1/FAP-1 in MCF7 cells completely abolished the survival action of insulin-like growth factor-I.
This effect occurred through a highly significant reduction in phosphatidylinositol 3-kinase/Akt pathway activation (80% reduction in phosphatidylinositol 3-kinase activity, 55% inhibition of Akt activation) accompanied by a 65% decrease in insulin receptor substrate-1 growth factor-induced tyrosine phosphorylation. These results provide the first evidence that PTPL1/FAP-1 has a key role in the apoptotic process in human breast cancer cells independent of Fas but associated with an early inhibition of the insulin receptor substrate-1/phosphatidylinositol 3-kinase pathway. Our data therefore suggest new therapeutic routes and strengthen the importance of identifying endogenous regulators and substrates of this phosphatase in breast tumors.
Breast cancer is one of the most common malignancies affecting women in the Western countries. Although mortality and survival rates decreased in the UK (1), breast cancer incidence is still increasing. It is therefore crucial in the coming years to design new therapeutic strategies based on the updated knowledge of the mechanisms by which tumor growth is sustained.
Breast cancer proliferation is the result of the balance between cell division and cell apoptosis. It has been shown in vitro in human breast cancer cell models that steroid hormones (mostly estrogens) and growth factors (epidermal growth factor, transforming growth factor ␣, IGF-I 1 and II, etc.) are the major signals affecting proliferation (2,3). Most of these factors stimulate cell division and/or promote cell survival by mechanisms yet poorly understood, thus conferring growth advantage to tumor-responsive cells. On the other side, growth inhibitors and various antagonists have been shown to increase breast cancer cell apoptosis (4,5).
Estrogens and their antagonists mediate their action through their nuclear receptors (estrogen receptors ␣ and ␤), acting as transcriptional factors in coordination with numerous additional transcriptional cofactors. Peptide hormones and growth factor-signaling pathways involve the activation of several kinase cascades (mitogen-activated protein kinases, phosphatidylinositol (PI) 3-kinase, etc.). We and others provided evidence for the existence of multiple interactions between these seemingly distinct pathways (6 -10). We previously demonstrated that steroidal and non-steroidal antiestrogens such as tamoxifen, which is currently used in breast cancer adjuvant therapy, inhibit the growth of estrogen receptor-positive cells not only by acting as competitors of hormone agonists on the nuclear receptors but also by preventing growth factor mitogenic action in the total absence of estrogens (6). Cell growth inhibition, evidenced by the drastic decrease in the overall cell DNA measurement, correlated with the expression of some protein-tyrosine phosphatases (PTP) (11), among which is PTPL1/FAP-1 (12)(13)(14)(15), whose mRNA accumulation is time-and dose-dependently regulated by 4-hydroxytamoxifen (OH-Tam) and ICI 182,780 (16). Using stable transfectants expressing PTPL1/FAP-1 RNA antisense, we have shown that this particular enzyme was a key factor in the OH-Tam-induced antigrowth factor action.
The role of PTP in breast cancer cell apoptosis is poorly understood, and SHP-1 was the only enzyme yet clearly implicated in somatostatin-induced apoptosis (17). Besides, a study in Jurkat cells proposed an antiapoptotic role of PTPL1/FAP-1 by inhibition of the Fas-signaling pathway (15), whereas MCF7 cells like the majority of the breast cancer cell lines are resistant to Fas-induced apoptosis (18).
Our present study was designed to evaluate the role of PTPL1/FAP-1 in the apoptotic process and to determine PTPL1/FAP-1 action on the survival signaling pathway stimulated by IGF-I. Using PTPL1/FAP-1 antisense transfectants in which enzyme expression is abrogated, we here demonstrate that protein-tyrosine phosphatase PTPL1/FAP-1 abolishes the survival action of IGF-I by a drastic inhibition of the IRS-1/ PI3-kinase/Akt pathway. Altogether, these data bring the first evidence for a new biological role of PTPL1/FAP-1 independent of Fas, which might offer new breast cancer therapeutic strategies.

EXPERIMENTAL PROCEDURES
Materials-All biochemicals were from Sigma unless otherwise specified.
Cell Culture-MCF7 human breast cancer cells were obtained from the Michigan Cancer Foundation (Detroit, MI). Establishment and characterization of stable PTPL1 antisense-expressing clones were previously described (16), B3 and B10, respectively, corresponding to clones 3 and 10. MCF7 cells and derived clones were maintained in Ham's F-12/Dulbecco's modified Eagle's medium (1:1) supplemented with 10% fetal calf serum (FCS) (Invitrogen). HEK 293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% FCS (Invitrogen). Before all hormonal treatments, the cells were stripped of endogenous steroids by successive passages in medium without phenol red containing 10% (2 days), then 3% (5 days) charcoalstripped (DCC) FCS. They were finally treated with E2, IGF-I, OH-Tam, IGF-I plus OH-Tam, or ethanol alone (control) in the presence of 1% DCC/FCS. DNA Fragmentation-Attached cells were collected in a buffer containing 5 mM Tris (pH 8.0), 10 mM EDTA, and 0.5% Triton X-100 combined with unattached cells and incubated for 20 min at 4°C. The lysate was centrifuged for 30 min at 13,000 ϫ g (4°C). The pellet containing unfragmented DNA was assayed for DNA content by the diaminobenzoic acid fluorometric method (LS-5 spectrometer, 405 nm excitation, 495 nm emission, PerkinElmer Life Sciences) after 2 washes with methanol as in Vignon et al. (6). The supernatant was treated with 300 g of proteinase K/ml for 30 min at 60°C. DNA was extracted twice with equal volumes of phenol, chloroform, isoamyl alcohol (25:24:1) and precipitated with 1/10 volume of sodium acetate (3 N) and 2.5 volumes of 100% ethanol. The pellets were washed once with 70% ethanol and resuspended in buffer containing 10 mM Tris (pH 8.0), 1 mM EDTA, and 50 mM NaCl in the proportions of 1 l of buffer for 2 g of unfragmented DNA assayed in the first pellet. Cytoplasmic DNA was then treated with 60 g of RNase A/ml for 60 min at 37°C. 15 l of cytoplasmic DNA was run on a 1.5% agarose gel for 4 h at 50 V. The gel was stained with ethidium bromide and photographed under UV light.
Terminal Deoxynucleotidyltransferase-mediated dUTP-biotin Nick End Labeling (TUNEL) Assay-Apoptotic cells were detected by TUNEL assay by using the ApopTag direct in situ apoptosis detection kit with peroxidase (Oncor). Attached cells were harvested by trypsinization, combined with unattached cells, fixed for 20 min in 4% paraformaldehyde, permeabilized for 5 min in 0.1% Triton X-100, and resuspended in phosphate-buffered saline. Next 2.5 ϫ 10 5 cells were spread on glass slides by use of a Cytospin. Detection of apoptotic cells and counterstaining with hematoxylin were performed according to the instructions provided by the manufacturer. The slides were viewed and photographed, and the apoptotic index was calculated as the percentage of nuclei labeled by peroxidase using a Samba analyzer (Alcatel). At least 10 3 cells were examined and evaluated in each sample.
Quantitation of Apoptosis by the Cell Death ELISA-2 ϫ 10 5 cells were plated on 30-mm duplicate dishes. At the indicated time points, one dish was assayed for total DNA contents by the diaminobenzoic acid method (6), and cells from the other dish were washed once with phosphate-buffered saline and assayed for the presence of nucleosomal fragments in the cytoplasm by a cell death detection ELISA kit (Roche Molecular Biochemicals) according to the manufacturer's instructions. The ELISA results were normalized to the total DNA value from the duplicate dish.
PI3-K Assay-Phospholipid mixtures (2 mg/ml) containing PI and phosphatidylserine were dried under a stream of nitrogen and sonicated twice at 100 watts for 15 s in 10 mM Tris-HCl (pH 7.4). The resulting vesicles (PI/phosphatidylserine) were used as a substrate for PI3-K. The preparation of cell extracts and immunoprecipitation for PI3-K were performed at 4°C. The cultured cells were washed 3 times with ice-cold phosphate-buffered saline and then lysed in 500 l of lysis buffer (20 mM Tris-HCl (pH 7.4), 137 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2, 0.5% Nonidet P-40, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 10% glycerol, and 100 kallikrein units/ml Trasylol). After gentle shaking for 30 min, the cell extracts were obtained by centrifugation in a microcentrifuge at 13,000 ϫ g for 5 min. The PI3-K from extracts containing equal amounts of protein, evalu-ated by Bradford assay (19), was immunoprecipitated with antiserum against IRS-1 (Upstate Biotechnology) and retained on protein G-Sepharose. The immunoprecipitates were washed twice with phosphate-buffered saline, 0.5% Nonidet P-40, twice with 100 mM Tris-HCl (pH 7.4), 0.5 M LiCl, 1 mM dithiothreitol, and 0.2 mM Na 3 VO 4 , and twice with 10 mM Tris-HCl (pH 7.4), 100 mM NaCl, 1 mM dithiothreitol, and 0.2 mM Na 3 VO 4 . All washes were performed at 4°C. The reaction mixtures (50 l), containing the immunoprecipitates in 20 mM Tris-HCl (pH 7.4), 80 mM NaCl, 25 mM MgCl 2 , 0.5 mM EGTA, 10 M ATP, 1 Ci of [␥-32 P]ATP, and 2 g of PI/phosphatidylserine were incubated at 25°C for 30 min. The reactions were terminated, and the lipids were extracted by the addition of 10 volumes of chloroform/methanol/HCl (11.6 N) (50:100:1). The mixture was then vortexed, and the organic phase was extracted with 2 volumes of methanol, 4 N HCl (1:1). The extracted products were desiccated, dissolved in 10 l of chloroform, and separated by thin-layer chromatography in a developing solution composed of chloroform, methanol, 4 M NH 4 OH (9:7:2). The production of phosphatidylinositol 3-phosphate was detected and quantified by counting with a Fujix-Bas 1000.
Western Blot Analysis of p85/IRS-1 Association-Cells were lysed in lysis buffer, and IRS-1 from extracts containing equal amounts of protein was immunoprecipitated with antiserum against IRS-1 and retained on protein G-Sepharose (see PI3-K assay). After washes, the immunoprecipitates were separated on an 10% gel by SDS-PAGE. Proteins were electrotransferred to Hybond-P membrane (Amersham Biosciences) and incubated with antiserum against p85 (anti PI3-K p85, Upstate Biotechnology). Proteins were finally visualized as previously described.
Western Blot Analysis of Akt Activation-Cells were lysed in lysis buffer (see PI3-K assay) and separated on a 10% gel by SDS-PAGE. Proteins were electrotransferred to Hybond-P membrane (Amersham Biosciences) and incubated with anti-Akt or anti-phospho-Akt (New England Biolabs, Ser-473). Proteins were visualized as previously described.
Western Blot Analysis of IRS-1 Tyrosine Phosphorylation-Cells were lysed in lysis buffer, and IRS-1 from extracts containing equal amounts of protein was immunoprecipitated with antiserum against IRS-1 and retained on protein G-Sepharose (see PI3-K assay). After washes, the immunoprecipitates were separated on an 8% gel by SDS-PAGE. Proteins were electrotransferred to Hybond-P membrane (Amersham Biosciences) and incubated with anti-Tyr(P) (4G10) or anti-IRS-1 (Upstate Biotechnology). Proteins were visualized as mentioned above.

Biochemical Evaluations of OH-Tam-induced Apoptosis in Human Breast Cancer Cells-
Our previous studies about the respective roles of OH-Tam, IGF-I, and the protein-tyrosine phosphatase PTPL1 on MCF7 growth regulation were based upon the evaluation of total cell numbers assessed by DNA cell measurement after diverse treatments (11). However, this global estimation did not discriminate between the relative contributions of the mitotic and/or the apoptotic indexes to the net growth rate. The presence in MCF7 cells of widely accepted morphological modifications induced by the apoptotic process have been long recognized both in vitro (4,20) and in animal xenografts after steroid antagonist treatments or estrogen withdrawal (21). However, the presence of apoptosis-associated biochemical evidence such as DNA fragmentation remained a matter of debate in the literature (22,23). Therefore, we have first evaluated whether IGF-I and OH-Tam modulations of apoptosis were associated with DNA strand breaks in our present culture conditions. We have indeed shown that OH-Tam treatment induced a marked DNA fragmentation in MCF7 cells (Fig. 1A, fourth lane). With this technique, we are, thus, able to confirm the apoptotic effect of OH-Tam (4) and observe the survival action of IGF-I in MCF7 cells (Fig. 1A, third lane). Furthermore, by examining DNA laddering, we have shown that the antiestrogen was able to counteract the survival action of IGF-I. DNA degradation was indeed equivalent in control untreated cells and in cells submitted to a 4-day simultaneous treatment with IGF-I and OH-Tam (Fig. 1A, compare lanes two and five). We have quantified apoptosis with two alternative biochemical techniques, the TUNEL assay, which allows the determination of the percentage of labeled apoptotic nuclei after cytocentrifugation of the pooled cell fractions (attached, loosely attached, and floating; Fig. 1B, typical nuclear staining) or a quantitative ELISA of cytoplasmic histone-associated DNA fragments (mono-and oligonucleosomes). Using these two assays, we have first applied a long term treatment (7 days) to quantify apoptosis in wild-type (wt) MCF7 cells in conditions in which we previously demonstrated a drastic inhibition of IGF-I-stimulated overall growth by OH-Tam (11). The two methods gave very similar results and confirmed OH-Tam pro-apoptotic effect as well as its capacity to inhibit IGF-I survival activity in wt MCF7 cells (Fig. 2, A and B, black bars). In the TUNEL assay ( Fig. 2A), although less than 10% of the cells were apoptotic in control culture conditions, we observed that OH-Tam treatment raised the percentage of apoptosis to greater than 40%, whereas as high as 30% apoptosis was counted in the presence of IGF-I and OH-Tam. Similarly, ELISA provided evidence of a 3.9-fold increase of apoptosis with OH-Tam alone and a 2.6-fold augmentation in the presence of IGF-I and OH-Tam (Fig. 2B). The results obtained with both biochemical methods strongly provided evidence that the drastic increase in OH-Tam-induced apoptosis was responsible for its anti-growth factor activity.
The Protein-tyrosine Phosphatase PTPL1 Plays a Key Role in Apoptosis Induction-The present results associated with our previous data demonstrating that OH-Tam anti-growth factor activity required PTPL1 expression and correlated with phosphatase up-regulation (16) led us to propose that PTPL1 was a key factor in OH-Tam-induced cell death. To demonstrate the role of PTPL1/FAP-1 in apoptosis, we have compared by TUNEL assay or ELISA the effect of various combinations of 7-day treatments on the degree of apoptosis in wt MCF7 cells, the void plasmid mock-transfectant clone (P4), and two MCF7derived stable clones transfected with PTPL1 cDNA antisense plasmids in which OH-Tam was no longer able to block IGF-I action on growth (16). We showed an overall lower sensitivity of antisense clones to OH-Tam treatment, with a positive correlation between resistance to apoptosis and antisense expression (Fig. 2, A and B). After treatment with antiestrogen alone there was a gradual significant decrease in the level of apoptosis, from 3.5 to 4 times the control level in wt MCF7 and mock-transfected P4 to 2-3-fold control in low or high antisense expresser B10 or B3 clones (Fig. 2, A and B). After simultaneous treatment with IGF-I and OH-Tam, the apoptosis level reached 2-3 times the control level (60 -70% of OH-Tam alone level) in PTPL1-expressing cells and went down to 1.3-1.5 times the control level in the B10 clone to finally attain the control level or even slightly below the control level in the B3 antisense clone (Fig. 2, A and B). These data clearly point to the essential role of PTPL1 in the apoptotic process.
In the B3 clone, after 7 days of OH-Tam treatment we observed a residual induction of apoptosis and a remaining inhibition of IGF-I survival effect (Fig. 2, A and B). These effects could result either from an accumulation of PTPL1 mRNA bypassing antisense inhibition or from distinct mechanisms of apoptosis induction by OH-Tam that do not require PTPL1 expression. We had shown that a 5-day treatment with OH-Tam was sufficient to increase PTP activity in wt MCF7, whereas it remained inefficient in B3 and B10 clones (16). By direct Western blotting with an N-terminal PTPL1/FAP-1 antibody, we have now demonstrated that enzyme protein expression and antiestrogen regulation (5-day treatment) were se- verely reduced in B10 and completely abolished in B3 as compared with wt MCF7 or P4 mock transfectant (Fig. 3A). We then quantified apoptosis in wt MCF7 and on the three stable derived transfectants after a 5-day treatment with the various combinations of compounds. In these conditions, OH-Tam displayed the same apoptotic activity and capacity to reverse IGF-I survival effect in wt or mock-transfected MCF7 cells (Fig.  3B). On the opposite, the antiestrogen could not significantly induce apoptosis and led to partial (B10) or no (B3) inhibition of IGF-I survival effect in PTPL1-defective clones (Fig. 3B). After such a shorter period of treatment, we obtained a perfect correlation between PTPL1 expression and apoptosis induction or IGF-I survival effect inhibition. The next step was therefore to determine which signaling pathway was affected in PTPL1 apoptosis induction.
PTPL1-induced Apoptosis Is Mediated by Inhibition of the PI3-K/Akt Pathway-It is well established that IGF-I has a strong antiapoptotic function in many cell lines (for review, see Ref. 24), but the survival pathway used by IGF-I against apoptotic injury is uncertain. In many cell lines, the importance of the PI3-K/Akt pathway has been demonstrated (25), but it is also clear that the IGF-IR is connected to several alternative pathways, which include mitogen-activated protein kinase activation (26). We therefore tested in MCF7 cells the impact of specific inhibitors of these two major pathways on apoptosis and IGF-I survival effect. Representative data of Fig. 4 clearly demonstrated that the MEK-1 (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase) inhibitor PD98059 was without effect, whereas the PI3-K inhibitor LY294002 induced apoptosis and partially inhibited IGF-I protective effect in this breast cancer cell line. This partial remaining of IGF-I survival signaling indicated that, as in other cell lines (27), additional connecting pathways contributed to IGF-I antiapoptotic activity in MCF7 cells. However, the results obtained with LY294002 pointed to the fundamental function of the PI3-K/Akt pathway in the control of cell death in MCF7 cells, and furthermore, the similarity between LY294002 and OH-Tam action suggested an inhibitory action of the antiestrogen, and most probably of PTPL1, on the PI3-K/Akt pathway (Fig. 4).
This pathway originates with the phosphorylation of the major IGF-IR substrate, IRS-1 (28), which in turn activates PI3-K and downstream Akt (25). Among the concluding steps are the inhibitory phosphorylations by Akt of the caspase 9 (29) and of Bad (30), one of the pro-apoptotic members of the Bcl-2 family of proteins. To further document the involvement of this pathway in PTPL1 action, we have tested the effect of OH-Tam pretreatment, on the association of PI-3K p85 regulatory subunit with IRS-1, and on the PI-3K activity after IGF-I activation in wt MCF7 cells and transfectants. OH-Tam significantly reduced PI3-K activation by IGF-I in wild-type and P4 mocktransfected MCF7 cells (respectively, 75 and 85% inhibition) (Fig. 5, B and C; see Fig. 7) without any significant change in IRS-1 expression (Fig. 5A). Moreover, OH-Tam selectively inhibited to the same extent the p85/IRS-1 association in the P4 mock-transfected clone (90% inhibition) (Fig. 6A, top panel, and B) without any change in p85 expression (Fig. 6A, bottom  panel). These data indicate, as suggested previously when comparing LY294002 and OH-Tam effects on apoptosis (Fig. 4), that the antiestrogen has a strong inhibitory action on the PI3-K/Akt pathway. We did not know yet if the residual PI3-K activity observed after 4 days of OH-Tam treatment in MCF7 cells could be sufficient for full activation of the downstream effector in this survival pathway, the activation of Akt. We have therefore estimated Akt activation by probing Western blots with specific antibody to phosphorylated Akt and evidenced a drastic inhibition (55%) of its activation in wild-type and mock-transfected MCF7 cells (see Fig. 8, A and B). This indicates that the reduction in PI3-K activation by OH-Tam was sufficient for severely impeding this survival pathway. We have shown (Fig. 5A) that IRS-1 expression was not affected by OH-Tam treatment. Therefore, inhibition of PI3-K activity observed after 4 days of OH-Tam treatment in MCF7 cells could be due to a direct PI3-K inhibition or IRS-1 dephosphorylation. We have then evaluated IRS-1 tyrosine phosphorylation by IRS-1 immunoprecipitation and anti-phosphotyrosine Western blots. In mock-transfected MCF7 cells (P4) we observed a drastic inhibition of IRS-1 phosphorylation (65%) without any loss of IRS-1 expression (see Fig. 9, A and B). This indicates that OH-Tam is acting on IRS-1/PI3-K/Akt pathway.
Our previous experiments indicated that PTPL1 was a key mediator of OH-Tam-induced apoptosis (Figs. 2 and 3), and our last experiments demonstrated that these apoptotic effects are mediated by inhibition of the Akt/PI3-K pathway (Fig. 4), thus pointing to a role of PTPL1 on this pathway. To evaluate PTPL1 contribution, we have thus evaluated the effect of a 4-day OH-Tam treatment on IRS-1 phosphorylation, p85 association with IRS-1, PI3-K activity, and Akt activation in wt MCF7 or P4 mock transfectant by comparison with that ob-tained in the two MCF7-derived PTPL1-defective clones (B3 and B10) (Fig. 3B). As expected, OH-Tam no longer blocked IRS-1 phosphorylation, p85 association with IRS-1, PI3-K activity, and Akt activation by IGF-I in these clones (Figs. 5-9), thus confirming the inhibitory role of PTPL1 on PI3-K/Akt pathway. In previous studies (11,16), we indeed showed that significant induction of PTP activity and PTPL1 expression required 2-3 days of treatment with antiestrogen. In a timecourse experiment (Fig. 7), we confirmed in the mock transfectant P4 the perfect correlation between the time necessary for PTPL1 induction of expression and the neutralizing activity on PI3-K/Akt pathway, whereas no effect was seen in the PTPL1defective B3 clone. Altogether our data on apoptosis in wildtype and PTPL1 antisense transfectants are therefore concordant with the concept that PTPL1 is a pro-apoptotic enzyme in estrogen receptor-positive breast cancer cells by early inhibition of the PI3-K/Akt pathway. Cancer Cells-On the basis of data obtained in Jurkat leukemia cells, Sato et al. (15) proposes a model for an anti-apoptotic role of PTPL1/FAP-1 by inhibition of Fas death receptor transduction through interactions of PTPL1/FAP-1 PDZ domains with Fas cytoplasmic terminal domain. They further suggest that the uncontrolled growth of cells in some pathologies including cancer could result from a correlation between PTPL1/ FAP-1 expression and resistance to Fas-induced cell death. A series of investigations has confirmed this assumption (31)(32)(33), whereas other groups failed to find such a correlation (34 -36), which altogether seemed to highly depend on the cell line examined and therefore strongly suggested that it was a tissueor a cell-specific event. We have evaluated the ability of an anti-Fas antibody to induce apoptosis in three estrogen-sensitive breast cancer cell lines, which expressed varying levels of PTPL1/FAP-1, wild-type MCF7 cells (low expression), T47D cells (3.5-fold more expression than in MCF7 cells) (Fig. 10B), and the B3 clone (no expression). Fig. 10A displayed the absence of Fas-mediated apoptosis in MCF7 cells and a 20-fold induction of apoptosis by Fas antibody in the T47D cell line, clearly providing evidence for the absence of a positive correlation between PTPL1/FAP-1 expression and Fas resistance in these breast cancer cell lines. Furthermore, the absence of sensitivity to Fas antibody in B3 clone (which did not express PTPL1) demonstrated that the PTPL1/FAP-1 expression was not responsible for Fas resistance in human breast cancer cells. DISCUSSION Regulation of apoptosis is a fundamental feature in the control of normal development but also in the progression of can-cer. The role of PTP in this process in breast cancer is poorly documented, and the relation between PTPL1 and apoptosis is still a matter of debate in the literature and seems to be highly dependent on cell type or species. In this study, we have addressed the role of PTPL1 in the model of antiestrogen-induced apoptosis in hormone-responsive human breast cancer cells positive or defective for this particular tyrosine phosphatase. We have shown that PTPL1 is necessary for the early inhibition of IRS-1/PI3-K/Akt pathway, which leads to increased apoptosis and inhibition of IGF-I survival effect. Therefore these results first demonstrate that PTPL1 has a pro-apoptotic effect in human breast cancer cells.
PTPL1 Is a Pro-apoptotic Enzyme Acting on the PI3-kinase/ Akt Pathway-Several studies have shown that antiestrogen inhibits growth factor effects on proliferation (6, 7) as well as on activation of AP-1-controlled genes (8) and IRS-1 phosphorylation (37). We have presently demonstrated that OH-Tam not only emphasizes the apoptotic effect of steroid withdrawal but also inhibits the survival effect of IGF-I. In many cell lines, the importance of the IRS-1/PI3-K/Akt pathway in survival has been demonstrated (25), but alternative pathways implying mitogen-activated protein kinase have been evidenced (38). In the 32D cell line devoid of IRS-1 two alternative pathways were demonstrated; one involves mitogen-activated protein kinase activation via Shc phosphorylation, and a second one, independent of mitogen-activated protein kinase and PI3-K, implies a mitochondrial translocation of Raf (26). In MCF7 cells, the use of specific inhibitors showed the predominant role of PI3-K pathway for the protective action of IGF-I. Whereas PD98059 was without effect, the results obtained with OH-Tam and LY294002 were analogous, suggesting that the antiestrogen can act as an inhibitor of the PI3-K pathway. Studies from our laboratory show that some of the IGF-I effects such as immediately early gene regulation or stimulation of glyceraldehyde-3-phosphate dehydrogenase expression were not inhibited by short term (8) or long term tamoxifen treatment, 2 indicating that some but not all the transducing pathways of IGF-I receptor were discontinued after antiestrogen treatment. We have now shown that pre-treatment with OH-Tam induced a reduction of IRS-1 phosphorylation and PI3-K activation as well as an inhibition of IGF-I-induced Akt activation. Our results clearly demonstrate that OH-Tam inhibits IGF-I survival effect through a blockade of the IRS-1/PI3-K/Akt pathway.
We have previously shown that the up-regulation of PTPL1 is crucial for the overall effect of tamoxifen on growth factor activity (16). Here, using stable antisense PTPL1 clones, we have evaluated the implication of this enzyme in the apoptotic process. We now demonstrate that OH-Tam had a reduced effect (7-day treatment) or no effect at all (5-day treatment) on IGF-I survival effect in cells defective in PTPL1. This is clear cut evidence for the role of PTPL1 in the anti-survival and pro-apoptotic effects of OH-Tam. To further understand PTPL1 implication and mechanism of action in apoptosis, we have tested how the absence of PTPL1 expression affected the ability of OH-Tam to inhibit IGF-I-induced PI3-K and Akt activities. We showed that there was no blockade of the IRS-1/PI3-K/Akt pathway in PTPL1-defective clones, indicating the predominant role of PTPL1 in this inhibition. It has previously been shown by other groups (39,40) that OH-Tam does not directly affect the expression of IRS-1/2 and IGF-IR, although it inhibits estrogen-induced expression of IRS-1 and IGF-IR. These results, associated with the present data on IRS-1 phosphorylation, point at an action of the tyrosine phosphatase at the level of some sites of the major tyrosine-phosphorylated proteins implicated in this signal transduction pathway. To identify the target(s) of PTPL1/FAP-1 in this transduction pathway we are now developing a panel of PTPL1 mutants able to specifically trap its endogenous substrates (41).
PTPL1/FAP-1 Apoptotic Effect Is Independent of the Fas Pathway-This study is one of the first evidences of a proapoptotic effect of a tyrosine phosphatase counteracting the adverse protective effect of a tyrosine kinase in breast cancer. Effectively, the role of these enzymes in cancer is poorly understood. Keane et al. (18) demonstrate by inducible transfection that DEP-1 had an antiproliferative effect in MCF7 cells, and Ardini et al. (42) show a positive correlation between high expression of PTP ␣ and delayed tumor growth and metastasis. Only SHP-1 was clearly implicated in the apoptotic process in breast cancer; the membrane recruitment of this enzyme by the somatostatin receptor or Fas has been shown to be an early event in their signaling that induced intracellular acidificationdependent apoptosis (17). Wild-type MCF7 (MCF7-wt), mock-transfected clone (P4), and PTPL1 antisensetransfected clones (B3 and B10) cultured in 1% DCC/FCS were treated for 4 days with 50 nM OH-Tam or vehicle alone (ethanol) before 10 min of IGF-I (20 nM) stimulation. A, cell lysates were analyzed by immunoblotting with an antibody specific for phospho-Akt (Akt-P; bottom panel). Equivalent amounts of Akt were confirmed by reprobing the blots with anti-Akt antibody (Akt; top panel). B, the signals from the experiments were quantified by scanning densitometry, and the level of phosphorylation in each sample was normalized to the level of Akt.
The relation between PTPL1/FAP-1 and apoptosis has been addressed by several groups and remains controversial. The first evidence of such a relation was brought up by Sato et al. (15) with the demonstration of the interaction of FAP-1 with the pro-apoptotic receptor Fas. In this initial study and a recent one from the same group (43) it was shown that in transfectant clones of Jurkat or TMK cells overexpressing FAP-1 the apoptotic effect of a Fas agonist antibody was inhibited by 50%. The three C-terminal amino acids (SLV) of human Fas were necessary and sufficient for its interaction with the second PDZ domain of PTPL1 (44). The direct cytoplasmic microinjection of this tripeptide resulted in the induction of Fas-mediated apoptosis in a colon cancer cell line expressing Fas and FAP-1. This observation indicated that an interaction between Fas C-terminal domain and the PDZ domain of a protein modulates Fas signal transduction inhibition, and the authors propose that FAP-1 could be this inhibitory protein (45). On the other hand, Cuppen et al. (46) show that there was no interaction between murine Fas, which lacked the SLV terminal sequence, and PTP-BL (mouse homologue of PTPL1/FAP-1). These data strongly suggest that this interaction, which is not conserved along the evolution, does not play a crucial role in FAP-1 function. In the same study, Cuppen et al. (46) could not generate an inhibition of Fas-induced apoptosis by overexpressing PTP-BL in mouse T cell lymphoma transfected with human Fas, which interacted with PTP-BL, indicating that this interaction is not sufficient for inhibiting Fas transduction in this cell type.
Several laboratories have looked for a correlation between PTPL1/FAP-1 expression and resistance to Fas-induced apoptosis in different cell lines or tissues. Some studies have provided evidence for a high level of FAP-1 mRNA expression in Kaposi's sarcoma (31) and, in pancreatic adenocarcinomas (33), a higher expression in T helper cells type 1 (which are resistant to apoptosis) than in T helper cells type 2 (which are sensitive to Fas ligand) (32). Other groups failed to show any correlation between Fas resistance and FAP-1 expression in colon cancer (34), in subclones of S49.1 murine T lymphocytes (35), in adult T cell leukemia cell lines (36), and in testicular germ cells where exposure to mono-(2-ethylhexyl) phthalate induces massive cell death and increases FAP-1 expression (47). We have presently evaluated the sensitivity to Fas-mediated apoptosis of three breast cancer cell lines expressing various levels of PTPL1/FAP-1 mRNA; they are T47D cells, which expressed high levels of enzyme, wild-type MCF7 cells, expressing moderate levels of PTPL1, and the B3 transfectant MCF7 subclone, in which enzyme expression was switched off. T47D cells were sensitive, whereas MCF7 cells were resistant to Fas-mediated apoptosis as previously described (48), indicating the absence of correlation between PTPL1/FAP-1 expression and Fas sensitivity. Moreover, B3 and MCF7 cells were equally resistant, indicating that in MCF7 cells, PTPL1 expression is not responsible for the resistance to Fas-mediated apoptosis. Altogether, these studies suggested a dual effect of PTPL1 on apoptosis. On one side, in human cells sensitive to Fas-mediated apoptosis, overexpression of PTPL1/FAP-1 could inhibit Fas signaling in a cell-and species-specific manner. On the other hand, in cells not expressing Fas or in which Fas is not efficient as is the case in most breast cancer cell lines, PTPL1/FAP-1 could have a pro-apoptotic effect by inhibition of the IRS-1/PI3-K/Akt signaling pathway.
Conclusions-We have presently demonstrated that the protein-tyrosine phosphatase PTPL1/FAP-1, whose expression is stimulated by antiestrogens in hormone-sensitive breast cancer cells, plays a major role in the inhibition of IGF-I survival pathway in these tumor cells by interfering on the IRS-1/PI3-K/Akt pathway. The pro-apoptotic effect of PTPL1 in breast cancer could suggest new therapeutic routes and points at the necessity of further studies on the mechanisms by which such enzymes are regulated by yet unknown endogenous factors in breast tumors.