Type-2A protein phosphatase activity is required to maintain death ...

Oncogene (2003) 22, 7677–7686

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Type-2A protein phosphatase activity is required to maintain death receptor responsiveness Ann-Sofi Ha¨rma¨la¨-Braskeń1,2,6, Andrey Mikhailov1,6, Thomas S So¨derstro¨m1,3, Annika Meinander1,4, Tim H Holmstro¨m1,3, Zahi Damuni5 and John E Eriksson*,1,4 Turku Centre for Biotechnology, University of Turku, A˚bo Akademi University, POB 123, Turku FIN-20521, Finland; Department of Biochemistry and Pharmacy, A˚bo Akademi University, POB 66, Turku FIN-20521, Finland; 3Department of Biology, A˚bo Akademi University, Turku FIN-20520, Finland; 4Department of Biology, Laboratory of Animal Physiology, University of Turku, Turku FIN-20014, Finland; 5Glogozymes, 6351 corte del Albeto Suite 101, Carlsbad, California 92009, USA 1 2

Type-2A protein phosphatase (PP2A) is a key regulator in many different cell signaling pathways and an important determinant in tumorigenesis. One of the signaling targets of PP2A is the mitogen-activated protein kinase (MAPK/ ERK) cascade. In this study, we wanted to determine whether PP2A could be involved in regulation of death receptor activity through its capacity to regulate MAPK/ ERK. To this end, we studied the effects of two different routes of protein phosphatase inhibition on death receptormediated apoptosis. We demonstrated that the apoptosis mediated by Fas, TNF-a, and TRAIL in U937 cells is suppressed by calyculin A, an inhibitor of type-1 and type2A protein phosphatases. The inhibition of the protein phosphatase activity was shown to subsequently increase the MAPK activity in these cells, and the level of activation corresponded to the degree of suppression of cytokine-mediated apoptosis. A more physiological inhibitor, the intracellular PP2A inhibitor protein I2PP2A, protected transfected HeLa cells in a similar way from Fas-mediated apoptosis and induced activation of MAPK in I2PP2A transfected cells. A corresponding inhibition could also be obtained by stable transfection with a constitutively active form of the MAPK kinase, MKK1 (also referred to as MEK1). The inhibitor-mediated protection was highly efficient in preventing early stages of apoptosis, as no caspase-8 cleavage occurred in these cells. The observed apoptosis suppression is likely to facilitate the tumor-promoting effect of a range of different type-2A protein phosphatase inhibitors, and could explain the reported tumor association of I2PP2A. Oncogene (2003) 22, 7677–7686. doi:10.1038/sj.onc.1207077 Keywords: PP2A; inhibition; I2PP2A; calyculin A; receptor-induced apoptosis

*Correspondence: JE Eriksson, Turku Centre for Biotechnology, University of Turku, A˚bo Akademi University, POB 123, Turku FIN20521, Finland; E-mail: john.eriksson@utu.fl 6 These two authors contributed equally to this work Received 2 January 2003; revised 25 July 2003; accepted 5 August 2003

Introduction Intracellular signal transduction pathways play an important role in the control of various cellular processes, such as cell growth, transformation, metabolism, and apoptosis. The primary mechanism to control the activity of these pathways is reversible phosphorylation of the involved regulatory proteins and kinases (for reviews, see Cohen and Cohen, 1989; Shenolikar and Nairn, 1991; Brautigan, 1997; Lewis et al., 1998). The type-1 (PP1) and type-2A (PP2A) serine/threonine protein phosphatases are major phosphatases for many kinases involved in the pathways of cellular homeostasis. In many cases, the task of, especially, PP2A is to keep kinases at steady-state activity (Brautigan, 1997). Many of the signaling pathways that are regulated by PP1 and/or PP2A are crucial determinants of cell fate, as they may control survival, growth, and differentiation of cells. Hence, they often govern cell proliferation and programmed cell death or apoptosis. There is a diverse group of natural toxins, like calyculin A, okadaic acid, and microcystin (Ishihara et al., 1989; Cohen et al., 1990; Eriksson et al., 1990; for review, see Fernandez et al., 2002), that all inhibit the type-1 and type-2A protein phosphatases. These toxins have also been reported as potent tumor promoters (Fujiki and Suganuma, 1993). These findings have led to the suggestion that PP2A, and possibly also other phosphatases, may function as tumor suppressors (Cohen and Cohen, 1989), and their inhibition would be an important step in cell transformation (Hahn et al., 2002). Another connection between protein phosphatases and cellular transformation is the finding that small DNA tumor viruses, such as Simian virus 40 (SV40) and Polyoma virus, synthesize proteins that bind to and reduce the enzymatic activity of PP2A (Pallas et al., 1990). It was shown that the alteration in the phosphatase activity and subsequent changes in phosphorylation levels were crucial steps in the transformation by these viruses (Pallas et al., 1990; Walter and Mumby, 1993). The interaction of SV40 with PP2A leads to increased activity of the mitogen-activated kinases (MAPK/ERK) and induces cell growth (Sontag et al., 1993). Related to the finding of tumor promotion

PP2A as a modulator of apoptosis A-S Ha¨rma¨la¨-Braskeń et al

7678

by small T-antigen is the observation that the PP2Aspecific inhibitor protein I2PP2A, equivalent to the acute undifferentiated leukemia-associated SET protein (von Lindern et al., 1992), has also been shown to act as an inductor in Wilm’s tumor formation (Al-Murrani et al., 1999). In the regulation of tumor formation, the ratio between cell growth and cell death is of crucial importance. The rate of cell death in a given tissue or tumor is determined by factors regulating the rate of apoptosis. Apoptosis occurs during embryonic development, tissue remodeling, immune regulation, and tumor regression. Different cell surface receptors, such as tumor necrosis factor (TNF) receptor-1, Fas/Apo-1 (CD95), and TNF-related apoptosis-inducing (TRAIL) receptor-1 (DR4) and -2 (DR5), act as physiological mediators of apoptosis. The signaling pathways by which these receptors induce apoptosis are rather similar. Ligand binding induces receptor oligomerization, followed by recruitment of adaptor proteins to form the death-inducing signaling complex (DISC). The adaptor proteins then recruit the initiator caspase, caspase-8, thereby connecting receptor signaling to the apoptotic effector machinery (reviewed by SchulzeOsthoff, 1998; Holmstro¨m and Eriksson, 2000). An important determinant in tumor elimination is the susceptibility of a given tumor cell to the proapoptotic cytokines of the immune cells and neighboring cells, for example, FasL, TRAIL, and TNF-a. Defects in the apoptotic and antiapoptotic signaling pathways have been implicated in different pathological conditions, including cancer (Fisher et al., 1994). Activation of the MAPK pathway usually leads to an increased celldivision rate and many tumor cell lines have increased levels of MAPK/ERK activity (Lengyel et al., 1997; Lewis et al., 1998). It is known that PP2A is involved in the regulation of MAPK/ERK signaling (Amaral et al., 1993), and we have shown that activation of MAPK is a potent inhibitor of Fas-mediated apoptosis (Holmstro¨m et al., 1998, 1999, 2000; Tran et al., 2001; So¨derstro¨m et al., 2002). Inhibitors of serine/threonine phosphatases were early implicated as being protective from various forms of receptor-induced apoptosis (Ohoka et al., 1993; Song and Lavin, 1994). It is also known that a number of known tumor promotors are able to block apoptosis in a variety of cell types (Wright et al., 1994). This could indicate that the tumor induction caused by protein phosphatase inhibitors is linked to suppression of apoptosis. In the present study, we have examined the effects of protein phosphatase inhibitors on cytokine-mediated apoptosis. As PP2A is an important regulator of ERK and MAPK kinase (MKK1), inhibition of PP2A is likely to result in activation of MAPK in any given cell type. While we have shown that increased MAPK activity protects cells against cytokine-mediated apoptosis, we wanted to study whether the observed tumor promotion caused by protein phosphatase inhibitors could partly be explained by MAPK-mediated protection against cytokine-mediated apoptosis. Our study shows efficient suppression of TNF-a-, Fas-, and TRAIL-mediated Oncogene

apoptosis by the protein phosphatase inhibitor calyculin A and physiological inhibitor protein I2PP2A, specific to PP2A. The inhibitor-mediated suppression occurs at the level of or upstream of the initiating caspase, caspase-8. Taken together, these results link protein phosphatase inhibitor-mediated MAPK activation to protection against cytokine-mediated apoptosis.

Materials and methods Cell culture The human monocytic cell line U937 was maintained in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS). HeLa and 3T3 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% inactivated FCS. Cells were grown in a humidified incubator with 5% CO2 in air at 371C. U937 cells were incubated at a density of 106 with an agonistic anti-human Fas IgM (100 ng/ml; MBL, Nagoya, Japan), TNF-a (50 mg/ml; Upstate Biotechnology, Lake Placid, NY, USA) or TRAIL (100 ng/ml; Alexis Corporation, Switzerland) for the indicated time periods in the presence or absence of calyculin A (cl-A; 5–50 nM; Alexis Corporation, Switzerland). HeLa cells were incubated with 3 mM cycloheximide (CHX; Sigma, St Louis, MO, USA) and 100 ng/ml antihuman Fas IgM for 8–10 h. Analysis of phosphatidylserine exposure and DNA fragmentation To detect phosphatidylserine exposure by flow cytometry, 200 ml U937 cell suspension was incubated for 10 min on ice with 200 ml of binding buffer (2.5 mM HEPES pH 7.4, 35 mM NaCl, 0.625 mM CaCl2) with 0.5 ml Annexin V-FITC (Alexis, Laufelfingen, Switzerland) and analysed on a FACScan flow cytometer (Becton Dickinson, NJ, USA). Detection of DNA fragmentation into oligonucleosomal DNA fragments by agarose gel electrophoresis was performed as described (Sorenson et al., 1990). To display the long-term functional viability of cells that survived from the different treatments, the surviving cells were sorted out and were allowed to recover. For this purpose, the treated cells were stained with Annexin V-FITC as above and additionally stained with propidium iodide (500 ng/ml). To sort out the surviving cells, exactly 104 cells double negative for Annexin V-FITC and propidium iodide staining were sorted in separate tubes with 1 ml of Iscov’s DMEM media supplemented with 20% FCS. Cells were transferred into wells of a 12-well plate, were grown for 120 h in a CO2 incubator, and then counted. Serine/threonine phosphatase assay Cells (1 106/sample) were lysed with 100 ml of lysis buffer (20 mM Hepes, pH 7.4, 30 mM beta-mercaptoethanol, 10% glycerol, 1 mM EGTA, 0.1% Nonidet P-40, 1 mM PMSF, and 1 mg/ml each of aprotinin, leupeptin, and pepstatin). The total protein phosphatase activities in the cell extracts were measured with a Protein Phosphatase Assay System (GibcoBRL Life Technologies, UK), according to the manufacturer’s instructions. Briefly, the assay uses 32P-labeled phosphorylase a as a substrate; the reaction was started by adding the substrate to the cell extracts. The cell extracts were diluted 1 : 100 in the protein phosphatase assay. The reaction was allowed to proceed for 10 min at 301C and was stopped by adding 20% trichloroacetic acid. The samples were placed on


7679 ice for 20 min and then centrifuged at 10 kG for 3 min. An aliquot of the clear supernatant was taken to determine the amount of released 32Pi in the reaction. The radioactivity from the released 32Pi was measured with a Microbeta Liquid Scintillation Counter (Wallac, Finland). MAPK activity assay Cells (2 106/sample) were lysed with 200 ml of lysis buffer (PBS, pH 7.4, 1% Nonidet P-40, 0.5% sodium deoxycholate, 1 mM Na3VO4, 0.1% SDS, 1 mM EDTA, 1 mM EGTA, 20 mM NaF, 1 mM PMSF, and 1 mg/ml each of aprotinin, leupeptin, and pepstatin). For MAPK immunoprecipitation, cell lysates were centrifuged (3000 g for 15 min), and the supernatant was incubated with mouse antibodies against human p42 MAPK and ERK2 (Transduction Laboratories, Lexington, KY, USA) coupled to protein G-sepharose (Pharmacia). Immunoprecipitates were then washed three times in lysis buffer and three times in kinase assay buffer (10 mM Tris, pH 7.4, 150 mM NaCl, 10 mM MgCl2, 0.5 mM dithiothreitol). The kinase reaction was carried out by adding 20 ml of kinase assay buffer to the immunoprecipitate. The kinase assay buffer included 25 mM ATP, 2.5 mCi of [g-32P]ATP (Amersham, Buckinghamshire, UK), and myelin basic protein (MBP; 1 mg/ml, Sigma) as substrate. The reaction was carried out for 15 min at 371C and stopped by addition of 3 Laemmli sample buffer. The samples were resolved on a 12.5% SDS–polyacrylamide gel (PAGE). Immunoblotting techniques Immunoblotting was performed after lysing cells in Laemmli sample buffer and then resolving the proteins on SDS–PAGE (12.5%). The separated proteins were transferred to nitrocellulose filter (Schleicher&Schuell, Dossel, Germany). The filters were blocked with either 5% bovine serum albumin (BSA) in PBS/Tween (137 mM NaCl, 2.7 mM KCl, 10.1 mM Na2HPO4, 1.8 mM KH2PO4/0.1% Tween-20) or 5% nonfat dried milk in TBS/Tween (10 mM Tris pH 7.5, 150 mM NaCl/0.1% Tween20). Phospho-MAPK (p44/p42) antibodies (New England BioLabs, Boston, MA, USA) were used at a dilution of 1 : 1000, a-Active MAPK (Promega) at 1 : 5000, caspase-8 (a kind gift from Dr Peter Krammer, Tumor Immunology G0300, German Cancer Research Center, D-69120 Heidelberg, Germany) at 1 : 8, followed by reaction with the appropriate HRP-conjugated secondary antibodies. Detection was carried out with the ECL system (Amersham Pharmacia Biotech, Buckinghamshire, UK). Transfection studies HeLa and 3T3 cells were transiently transfected using Lipofectin (GibcoBRL, UK) in Optimem media. Cells were seeded at 105 cells/ml in 24-well plates on coverslips and grown up to 20% confluence, and then transfected with the I2PP2A construct using lipofectin : DNA (1 : 2). Control and MOCK transfections were performed with lipofectin and noncoding vector alone. After 24 h of growth, HeLa cells were placed on 10% FCS. The treatment with Fas and cycloheximide was started at 48 h and cells were harvested after 8–10 h. In the case of 3T3 cells, 10% FCS was added after 8 h growth, media was changed at 10 h, and cells were harvested at 26 h. 3T3 cells were also transiently transfected by electroporation (1000 mF, 240 V) using a Bio-Rad Gene Pulser. Cells (5 105) were washed, resuspended in 0.4 ml of Optimem (Gibco-BRL), and 30 mg of plasmid DNA (I2PP2A or pMCL-HA-MKK1-S218E/S222D) was used. Cells were cultured on coverslips for 2 h with 10%

FCS, in order to attach. After 2 h serum was withdrawn and after 24 h serum starvation cells were harvested. A stable U937 cell line that expressed constitutively active MKK1 was produced using the DNA construct pMCL-HAMKK1-S218E/S222D (a kind gift from Natalie Ahn, University of Colorado, Boulder, CO, USA). The cell line was established as follows: 106 U937 cells were transfected by electroporation (320 V, 700 mF) in the presence of 30 mg of DNA. Cells were seeded into a 10 ml culture flask containing RPMI media supplemented with 15% FCS. After 48 h, 100 mg/ ml of hygromycin (Roche, Switzerland) was added, the concentration of hygromycin was increased to 250 mg/ml after 1 week and cells were seeded into 8 96-well plates 100l/wel (100 ml/well). After 2 weeks selection, six live clones remained and they were transferred to 24-well plates. The clones were transferred to 10 ml culture flasks after 1 week, and positive clones were assayed by Western blot with an antibody against HA. Two positive clones were found (U937.11-MKK1 and U937.12-MKK1), which were then used for further studies. For detection of transfected cells, the cells were fixed for 15 min with 3% formaldehyde in PBS. The cells were then washed once with PBS and permeabilized with 0.1% Nonidet P-40 (Sigma) for 5 min at room temperature. After washing with PBS, cells were incubated for 1 h at room temperature with a blocking solution, 20% boiled normal goat serum (BNGS). After blocking, cells were incubated for 1 h with a monoclonal Flag-specific antibody (1 : 500, Sigma)7pERK antibody (1 : 500, Transduction Laboratories) in PBS with 5% BNGS. Cells were then washed three times with PBS and incubated further for 1 h with FITC-conjugated goat antimouse antibody (1 : 1000, Alexat, Molecular Probes)7 TRITC-conjugated goat anti-rabbit antibody (1 : 1000, Alexat, Molecular Probes). The DNA was stained with DAPI (0.3 mg/ml, 40 6-diamidino-2-phenylindol, Sigma) for 5 min, coverslips were washed three times before being mounted with Mowiol 4-88 (Aldrich) containing 2.5% DAPCO (1,4-diazobicyclo-[2.2.2]-octan, Aldrich) or Vectashield (Vector Laboratories). The cells were analysed using a Leica DMR fluorescence microscope equipped with a digital Hamamatsu ORCA CCD camera. Images were further processed using Adobe Photoshop and PowerPoint software. For bar graph data representation, 150–500 cells were counted and the results are average7range of values from two different experiments.

Results Fas-, TNF-a-, and TRAIL-mediated apoptosis is suppressed by protein phosphatase inhibitors It has been shown that PP2A is involved in the regulation of MAPK by suppressing the MAPK signaling (Braconi Quintaje et al., 1996; Chung and Brautigan, 1999). Previously, we have shown that MAPK is involved in the regulation of Fas-mediated apoptosis (Holmstro¨m et al., 1998; 1999; 2000; So¨derstro¨m et al., 2002; Tran et al., 2001). In order to study how inhibition of protein phosphatases affect cytokine-induced apoptosis, we used two different approaches and inhibitors: (1) calyculin A (a natural toxin, which acts as a specific inhibitor of protein phosphatase type 1 and type 2A) and (2) I2PP2A (an intracellular PP2A inhibitor; Al-Murrani et al., 1999), focusing on how they affect Fas-, TNF-a-, and TRAILmediated apoptosis in U937 cells that respond by Oncogene


7680

a control

Fas

Number of events

6.5 %

19.8 %

TRAIL

TNF

26.8 %

33.4 %

FITC-fluorescence intensity

b 6 Relative amount of apoptotic cells

apoptosis when stimulated with any of these three death receptor types. U937 cells were incubated with the apoptosis-inducing death receptor agonists, Fas antibody (100 ng/ml), TNF-a (50 mg/ml), and TRAIL (100 ng/ml). The degree of apoptosis was assessed by flow-cytometric analysis of cells exposing phosphatidylserine, a feature typical for apoptotic cells. After 4 h stimulation of the Fas, TNF-a, and TRAIL receptors, respectively, approximately 20– 30% of the cells became apoptotic. The degree of apoptosis was shown to be dependent on the ligand used (Figure 1a). However, when cells were pretreated for 10 min with 30 nM calyculin A before addition of the respective apoptosis-inducing ligands, apoptosis was inhibited to a varying degree. Since calyculin A alone was shown to affect the degree of apoptotic cells by increasing it by 4–5%, the relative amount of apoptotic cells is shown in Figure 1b. The results show that calyculin A caused a significant inhibition of cytokinemediated apoptosis. To determine whether apoptosisinduced DNA fragmentation could also be inhibited by calyculin A, the induction of apoptotic chromatin cleavage into oligonucleosomal-sized DNA fragments was measured by agarose gel electrophoresis. Our results show that pretreatment with calyculin A suppresses the formation of DNA fragments in a concentrationdependent manner (Figure 1c). To ensure that the cells protected with calyculin A were really alive and functional, we sorted out 104 Annexin/PI-negative cells (at aseptic conditions, sorting performed on FACSVantage SE cell sorter upgraded with the new DiVa digital acquisition platform) and allowed them to recover. Although the sorting process itself caused noticeable lag in cell division, there were no significant differences in cell recovery between control, calyculin-A-, and calyculin A/TRAIL-treated cells (Table 1), which implied that the phosphatase inhibition really provided a long-term protection for the cells.

5

Control 30 nM Calyculin A

33.4% 26.8%

4 19.8%

3

1

18.4%

15.0%

14.4%

2 6.5% 10.2%

0 CTRL

c

Fas

Fas

TNFα

TNF-α

cl A - 50 - 5 10 50 - 5 10 50

TRAIL

TRAIL - 5 10 50 nM

Transient transfection of HeLa cells with the PP2A inhibitor protein I2PP2A As calyculin A affects a broad range of protein phosphatase activities, we wanted to test whether similar effects could be obtained with a more physiological and well-targeted inhibitor of PP2A. The PP2A inhibitor protein I2PP2A (fusion construct with Flag tag) was transiently transfected into HeLa cells in order to study whether its specific PP2A inhibition could interfere with Fas-mediated apoptosis. Control and transfected HeLa cells were incubated with CHX (3 mM) and Fas (100 ng/ ml) for 8–10 h in order to induce apoptosis. The transfected cells were visualized by a-Flag immunoreactivity and the nuclear morphology by DNA staining. Expression of the PP2A inhibitor protein a-Flag-I2PP2A itself did not affect the number of apoptotic cells among the transfected cell population, but it did clearly protect the a-Flag-I2PP2A transfected HeLa cells from Fasinduced apoptosis (Figure 2A, panel g) when compared to nontransfected cells treated with CHX and Fas (Figure 2A, panel c). The effect of expression of I2PP2A on Oncogene

Figure 1 Inhibition of Fas-, TNF-a-, and TRAIL-mediated apoptosis in calyculin-A-treated U937 cells. U937 cells were incubated with anti-FasR antibody (100 ng/ml), TNF-a (50 ng/ml) or TRAIL (100 ng/ml þ 2 mg/ml of enhancer ab) for 4 h. (a) Annexin –V-FITC labeling was determined by using a FACScan flow cytometer. Bars indicate the percentage of apoptotic cells. (b) Cells were pretreated with 30 nM cl A for 10 min before addition of Fas, TNF-a or TRAIL. The amount of apoptotic cells was determined by a FACScan flow cytometer and the relative ratios of apoptotic cells are shown (control ¼ 1). Numbers on top of the bars show the actual percentage of apoptotic cells from one representative experiment. (c) The formation of oligonucleosome-sized DNA fragments in U937 cells were studied by agarose gel electrophoresis

Fas-induced apoptosis was quantified by counting apoptotic and nonapoptotic cells (Figure 2B). These results further corroborate the hypothesis that PP2A is an important member in the signaling system regulating cytokine-mediated apoptosis.


7681 Table 1

Functional recovery after calyculin A treatment Number of cells after 120 h

Unsorted, untreated cells

Untreated, sorted cells

Calyculin A-treated, sorted cells

Calyculin A and TRAIL-treated, sorted cells

1.0670.45 106

0.1170.05 106

0.0970.06 106

0.1470.09 106

Cells were incubated in the presence or absence of calyculin A (30 nM) and TRAIL (100 ng/ml), respectively. Of the treated and untreated cells, 104 surviving cells were sorted out by flow cytometry (see Materials and methods), and the cells were allowed to recover for 120 h, after which they were counted. In parallel, 104 nonsorted cells were taken manually as control to grow at the same conditions (three samples/treatment)

MAPK activation in U937 cells correlates with calyculinA-induced inhibition of protein phosphatase activity We wanted to test whether the observed inhibition of apoptosis could be related to the levels of protein phosphatase and MAPK activities. The protein phosphatase activity was measured with phosphorylase a as substrate in U937 cells, incubated for 30 min in the presence of increasing concentrations of calyculin A. A 50–60% inhibition of the total protein phosphatase activity was obtained with calyculin A concentrations in the range of 10–50 nM (Figure 3a). Calyculin A concentrations in the range of 1–5 nM added to cell extracts in vitro will give the same amount of inhibition (Ha¨rma¨la¨-Braskeń, A-S, Mikhailov, A and Eriksson, JE, unpublished observations). The remaining phosphatase activity observed at high concentrations of calyculin A, as well as the differences in phosphatase inhibition in vivo vs in vitro, is likely to be a result of dissociation of the inhibitor from the protein phosphatases during the experiment. To determine if the calyculin-A-dependent inhibition of protein phosphatase activity could affect MAPK activity, we used MAPK assays and an antibody specific for active MAPK. The results show that calyculin A induced a clear concentration-dependent increase in MAPK activity, as determined by MBP phosphorylation in the kinase assay (Figure 3b), with peak values at 20–50 nM calyculin A, approximately 5– 6-fold higher than the control values. Western blotting with the antiactive MAPK antibody shows a similar activation of MAPK (Figure 3c). To confirm that the observed effect on MAPK activity was not due to changes in the amount of MAPK, the same samples were immunoblotted with a p44/p42-specific antibody (Figure 3d). From these results, we can clearly conclude that the calyculin-A-induced inhibition of protein phosphatase activity corresponds well with the increased MAPK activity in these cells. The PP2A inhibitor protein I2PP2A induces MAPK activity in 3T3 cells In order to further study the relationship between PP2A inhibition and MAPK/ERK activation, 3T3 cells were transiently transfected with the PP2A inhibitor protein a-Flag-I2PP2A. Immunostaining with a-Flag and ERK antibodies revealed the elevation of ERK activity in cells positive for a-Flag-I2PP2A (Figure 4a). As a control, we also used constitutively active HA-MKK1, which we transiently transfected into

3T3 cells and studied the effects on ERK. Immunostaining with HA and p-ERK antibodies demonstrates that the HA-MKK1-expressing cells also show parallel elevation of p-ERK activity (Figure 4A). The immunostaining results were further confirmed by Western blotting with p-ERK of transfected and nontransfected cells (Figure 4B). Stable transfection with constitutively active MKK1 protects cells from Fas-, TNF-a-, and TRAIL-mediated apoptosis In order to study in parallel the effects of an elevated MAPK activity on cytokine-induced apoptosis in U937 cells, we established stable transfectants of a constitutively active form of MKK1. Wild-type and stable transfected cells were treated with Fas, TNF-a, and TRAIL for 4 h. The amount of apoptotic cells was determined by flow cytometry as Annexin V positivity. U937 cells expressing the constitutively active form of MKK1 were efficiently protected against Fas-, TNF-a-, and TRAIL-induced apoptosis. In the stably transfected cell line Fas-, TNF-a-, and TRAIL could not further elevate the degree of apoptosis, as demonstrated when the relative amount of apoptotic cells was compared in wt U937 cells and U937 stably transfected with constitutively active MKK1. A clear protection against cytokine-mediated apoptosis could be observed in cells expressing constitutively active MKK1 (Figure 5). Treatment of MKK1-transfected cells with calyculin A did not increase the level of protection against cytokineinduced apoptosis (data not shown). These results provide further evidence that the increased MAPK activity is involved in protecting the U937 cells from apoptosis. The effect of calyculin A on caspase activity in U937 cells treated with Fas-, TNF-a-, and TRAIL In order to study where in the apoptosis signaling cascade the inhibitory effect of calyculin A is targeted, we performed Western blotting assay of caspase-8 activation with an antibody recognizing the inactive, intermediate, and active forms of caspase-8. The Western blot shows that cells pretreated with calyculin A before addition of Fas, TNF-a, and TRAIL show inhibition of the cleavage of caspase-8 to the active p-18 subunit (Figure 6a). In U937 cells, with stable expression of constitutively active MKK1, almost no active p18 could be detected in contrast to wild-type U937 cells Oncogene


7682

A

α-Flag / FITC

DAPI

b

c

d

e

f

g

h

I2PP2A-Flag CHX + Fas

I2PP2A-Flag Control

CHX + Fas

Control

a

B

35 % Apoptotic cells

30 25 20 15 10 5 0 CTRL Fas + CXH CTRL Fas + CHX Flag-I2PP2A Figure 2 Transfection of HeLa cells with the PP2A inhibitor protein I2PP2A protects cells from Fas-induced apoptosis in the presence of cycloheximide (CHX). (A) Control cells and cells transiently transfected with a-Flag-I2PP2A were incubated for 8–10 h in the absence (a–b, e–f) or in the presence (c, d, g, h) of anti-Fas (100 ng/ml) and CHX (3 mM). (a) A representative immunofluorescence graph shows nuclear alterations visualized by DAPI staining (left panels) and transfected cells (arrowheads on (e) and (g)), as detected by FITC immunofluorescence of a-Flag (right panels). In (c) and (g), the arrows indicate apoptotic cells. (B) The percentage of apoptosis in control and transfected cells was assessed by judging chromatin condensation viewed under an epifluorescence microscope

Oncogene


7683

a Protein phosphatase activity (%)

100 80 60 40 20 0 C

0.1

1

5

10

20

50

nM cl A

b — MBP C

0.1

1

5

10

20

50 nM cl A

c 46 —

— α-active MAPK C

d 46 —

0.1

1

5

10

20

50 nM cl A — p42/p44

Figure 3 Correlation between calyculin-A-mediated inhibition of protein phosphatase activity and increased MAPK activity in U937 cells. Cells were incubated for 30 min with indicated concentrations of calyculin A. (a) Protein phosphatase activity was measured by an assay with phosphorylase a as substrate, (control ¼ 100%). (b) The MAPK activity was measured by an immune complex assay with MBP as a substrate. (c) The activation of MAPK was confirmed by immunoblotting with a specific antibody to a-active MAPK. (d) To confirm equal loading and to check whether the observed effect could be due to changes in the amounts of MAPK, samples from (c) were immunoblotted with an p42/p44-specific antibody

that had been stimulated on their Fas, TNF-a, and TRAIL receptors (Figure 6b).

Discussion An important step in growth-regulatory signal transduction pathways is the reversible phosphorylation of proteins. The net balance of phosphatase/kinase activities maintains most of the regulatory pathways of any given cell in a functional state. An indication that serine/ threonine phosphatases are directly involved in growth control came through the discovery that okadaic acid is a specific inhibitor of PP1 and PP2A (Bialojan and Takai, 1988), and also a very potent tumor promoter (Rosenberger and Bowden, 1996). These findings have also led to the suggestion that PP2A may function as a tumor suppressor (Cohen and Cohen, 1989; Hahn et al., 2002). Different inhibitors of serine/threonine protein phosphatases have been shown to profoundly affect the

susceptibility to programmed cell death in a variety of cell types (Gjertsen and Doskeland, 1995). Here, our basic assumption is that the tumor promotion caused by protein phosphatase inhibitors could be a combined effect of growth stimulation and apoptosis inhibition. Our results show that the protein phosphatase inhibitor calyculin A was efficiently suppressing apoptosis, as triggered by death-inducing cytokines such as Fas, TNF-a, or TRAIL in U937 cells. It has previously been shown that different tumor promoters inhibit apoptosis in a variety of cell lines (Wright et al., 1994). However, as shown in different studies, inhibitors of PP2A are, depending on the concentration, able to both induce (Fladmark et al., 1999) and prevent (Chiang et al., 2001) apoptosis. The obvious controversy is likely to be a result of the differences in the employed cellular models and the great range of inhibitor doses used in different studies. The PP2A-specific inhibitor protein I2PP2A is highly expressed in Wilms’ tumor formation (Carlson et al., 1998) and during early development (Von Lindern et al., 1992). This protein is an endogenous component and suggested to be a selective inhibitor of PP2A. When cells in the present study were transiently transfected with this inhibitor, the death receptor-mediated apoptosis was clearly inhibited. The degrees of apoptosis suppression caused by both artificial exogenous as well as physiological endogenous protein phosphatase inhibitors are very similar. In this study, we showed that inhibition of serine/ threonine phosphatase activity by calyculin A and by I2PP2A leads to an elevated MAPK activity, which was not due to changes in the protein levels of MAPK. Previously, MAPK activation was found to suppress Fas-mediated apoptosis (Holmstro¨m et al., 1998). Here we demonstrate that there is a good agreement between inhibition of protein phosphatase activity, elevated MAPK activity, and inhibition of apoptosis. This indicates that the calyculin-A-induced protection against cytokine-mediated apoptosis is likely to be via MAPK. Stable transfectants with constitutively active MKK1 (upstream initiator of MAPK/ERK) showed an intense inhibition of Fas-, TNF-a-, and TRAILmediated apoptosis, very similar in degree and intensity to what we observed with PP2A inhibitors. Apart from the primary implicated involvement of the MAPK/ERK pathway in the antiapoptotic effects of PP2A inhibitors, there remains the possibility of cooperative actions with other pathways, for example the PI3/PKB pathway, which is also known to be affected by PP2A and having antiapoptotic properties (Sato et al., 2000). Execution of apoptosis is passing several consecutive stages. Our results show that the calyculin-A-induced inhibitory effect on Fas-, TNF-a-, and TRAIL-mediated apoptosis is seen at the very early stages of apoptosis, including the disappearance of activated caspase-8. It led us to the conclusion that one of the major targets of antiapoptotic action of calyculin A is located quite upstream in apoptosis machinery, probably at the level of DISC assembly or activation of initiator caspases. It was shown before that regulatory A subunit of PP2A Oncogene


7684

A

I2PP2A-Flag

Control

MKK1-HA

b

c

d

e

f

g

h

i

p-ERK

α-Flag/HA TRITC

DAPI

a

B 46 KD – p-ERK –

46 KD – p44 – p42 – p-ERK

0.440

0.196

1.002

ERK

0.551

0.432

0.747

ratio

0.798

0.454

1.341

C

C

I2PP2A

+ vector Figure 4 Activation of p-ERK in 3T3 cells transiently transfected with the PP2A inhibitor protein I2PP2A. (A) Representative immunofluorescent micrographs of 3T3 cells transfected with the inhibitor protein I2PP2A or with constitutive active MKK1. Nuclei were stained with DAPI (panels a–c). Monoclonal a-Flag and HA were developed by Alexa586-conjugated secondary antibody in order to detect the presence of Flag-tagged I2PP2A and HA-tagged MKK1 in transfected cells (panels d–f). An antibody against p-ERK developed with FITC-conjugated secondary antibody was used to detect MAPK activation (panels g–i). Cells expressing I2PP2A and constitutively active MKK1 show elevated p-ERK activity. The arrows indicate transfected cells. (B) Western blotting with p-ERK antibody of lysates from 3T3 cells transiently transfected with I2PP2A: cells transfected with I2PP2A and control transfection with noncoding vector. The lower part is the loading control for the samples developed with antibodies against p42/p44 ERK. The table shows numbers from the pixel count of p-ERK and ERK bands in the Western blot (done with Imaging Research MCID M4 software)

can interact with caspase-3, leading to ablation of this subunit with consequential increase in PP2A activity and decrease of activity of MAPK/ERK during execution of apoptosis (Santoro et al., 1998). We are currently investigating the possibility of a similar interaction pattern between PP2A and upstream caspases. Taken together, here we propose that MAPK/ERK activation is the main antiapoptotic effector of PP2A inhibition in Oncogene

cells, and its antiapoptotic action is targeted at the level of initiator caspases. This study was focused on the short-term effects on cell populations. Nevertheless, the presented results agree well with the suggested role of PP2A as a tumor suppressor (Cohen and Cohen, 1989) and that its inhibition would be an important step in cell transformation (Hahn et al., 2002). One crucial step in tumor


7685 WT U937 MKK1.11-U937 _ 26.8%

Fas

CTRL

Fas

TNF-α

_

+

_

_ p18 cl A

+

TRAIL

TNF-α

TRAIL

Figure 5 Constitutively active MKK1 inhibits Fas-, TNF-a-, and TRAIL-induced apoptosis in U937 cells. U937 cells stable transfected with constitutively active MKK1 were treated for 4 h with FasR antibody (100 ng/ml), TNF-a (50 ng/ml) or TRAIL (100 ng/ml). The number of apoptotic cells was determined by Annexin V-FITC staining and FACScan flow cytometer; the results are presented as the relative amounts of apoptotic U937 cells (WT and transfected with constitutively active MKK1), treated with Fas, TNF-a, and TRAIL (Control ¼ 1). The data represent mean values7s.d. from three different experiments

promotion and cell transformation is the obstruction of death receptor signaling, as T-cell-mediated surveillance and destruction of transformed cells take place through activation of death receptor signaling (Holmstro¨m and Eriksson, 2000). Our results imply that already partial inhibition of the PP2A activity will lead to abrogation of death receptor signaling, in addition to the growthpromoting effect that has been implicated by PP2A inhibition, for example, through the action of small Tantigen (Pallas et al., 1990, Walter and Mumby, 1993). Moreover, the outcome of PP2A inhibition-mediated modulation of death receptor signaling could be a shift in the operational mode of the death receptors, from apoptosis-promoting to growth-promoting, as has been seen in several tumor cell lines (Holmstro¨m et al., 1999). However, more detailed studies will be required to more formally demonstrate the relationship between PP2A activity, death receptor signaling, and tumor promotion in in vivo settings. To this end, stable transfection of

b

TRAIL

22.9%

TNF-α

21.4%

TRAIL

20.6%

TNF-α

6.5% 19.4%

Fas

0

+

19.8%

2 1

_

Fas

3

+

Control

4

_ _ p43/41

Control

Relative amount apoptotic cells

5

__ caspase-8a/b

a

33.4%

__ caspase-8a/b _ _ p43/41 _ p18 U937.11-MKK1

U937.12-MKK1

Figure 6 Caspase-8 activation in Fas-, TNF-a-, and TRAILstimulated cells is affected by calyculin A and elevated MAPK activity. Cells were preincubated (10 min) with cl A (30 nM) before addition of FasR antibody (100 ng/ml), TNF-a (50 ng/ml) or TRAIL (100 ng/ml þ 2 mg/ml of crosslinking antibodies), after which the incubation continued for 4 h. (a) Immunoblotting with an antibody to caspase-8 showed the appearance of active caspase fragments (p-18) was inhibited in calyculin A pretreated cells. (b) Two stable transfected cell lines expressing constitutively active MKK1 were treated for 4 h with FasR antibody (100 ng/ml), TNFa (50 ng/ml) or TRAIL (100 ng/ml þ 2 mg/ml of crosslinking antibodies) and immunoblotted against the caspase-8 antibody. No active fragments of caspase-8 could be detected

suitable cell lines with the I2PP2A inhibitor or the small Tantigen, followed by transfer into an appropriate mouse model will be designed, including parallel analysis of apoptosis, cell growth, and tumor formation. Acknowledgements We thank Helena Saarento for excellent technical assistance. Financial support from the Academy of Finland, the Finnish Cancer Foundations, the Sigrid Juse´lius Foundation, the Victoria Foundation, and A˚bo Akademi University are gratefully acknowledged.

References Al-Murrani SW, Woodgett JR and Damuni Z. (1999). Biochem. J., 341, 293–298. Amaral MC, Casillas AM and Nel AE. (1993). Immunology, 79, 24–31. Bialojan C and Takai A. (1988). Biochem. J., 256, 283–290. Braconi Quintaje SB, Church DJ, Rebsamen M, Valloton MB, Hemmings BA and Lang U. (1996). Biochem. Biophys. Res. Commun., 221, 539–547. Brautigan DL. (1997). Adv. Second Messenger Phosphoprotein Res., 31, 113–124. Brautigan DL and Pinault FM. (1993). Mol. Cell. Biochem., 127–128, 121–129. Carlson SG, Eng E, Kim EG, Perlman EJ, Copeland TD and Ballermann BJ. (1998). J. Am. Soc. Nephrol., 9, 1873–1880. Chiang CW, Harris G, Ellig C, Masters SC, Subramanian R, Shenolikar S, Wadzinski BE and Yang E. (2001). Blood, 97, 1289–1297.

Chung H and Brautigan DL. (1999). Cell Signal., 11, 575–580. Cohen P and Cohen PTW. (1989). J. Biol. Chem., 264, 21435–21438. Cohen P, Holmes CFB and Tsukitani Y. (1990). TIBS, 15, 98–102. Eriksson JE, Gro¨nberg L, Nyga˚rd S, Slotte JP and Meriluoto JAO. (1990). Biochem. Biophys. Acta, 1025, 60–66. Fernandez JJ, Candenas ML, Souto ML, Trujillo MM and Norte M. (2002). Curr. Med. Chem., 9, 229–262. Fisher DE. (1994). Cell, 78, 539–542. Fladmark KE, Brustugun OT, Hovland R, Boe R, Gjertsen BT, Zhivotovsky B and Doskeland SO. (1999). Cell Death Differ., 6, 1099–1108. Fujiki H and Suganuma M. (1993). Adv. Cancer Res., 61, 143–194. Gjertsen BT and Doskeland SO. (1995). Biochim. Biophys. Acta, 1269, 187–199. Oncogene


7686 Hahn WC, Dessain SK, Brooks MW, King JE, Elenbas B, Sabatini DM, DeCaprio JA and Weinberg RA. (2002). Mol. Cell. Biol., 22, 2111–2123. Hahn WC and Weinberg RA. (2002). Nat. Rev. Cancer, 2, 331–341. Holmstro¨m T, Chow SC, Elo I, Coffey ET, Orrenius S, Sistonen L and Eriksson JE. (1998). J. Immunol., 160, 2626–2636. Holmstro¨m T and Eriksson JE. (2000). Crit. Rev. Immunol., 20, 121–152. Holmstro¨m T, Tran SEF, Johnson VL, Ahn NG, Chow SC and Eriksson JE. (1999). Mol. Cell. Biol., 19, 5991–6002. Holmstro¨m TH, Schmitz I, So¨derstro¨m TS, Poukkula M, Johnson VL, Chow SC, Krammer PH and Eriksson JE. (2000). EMBO J., 19, 5418–5428. Ishihara H, Martin BL, Brautigan DL, Karaki H, Ozaki H, Kato Y, Fusetani N, Watabe S, Hashimoto K, Uemura D and Hartshorne DJ. (1989). Biochem. Biophys. Res. Commun., 159, 871–877. Lengyel E, WangH, Gum R, Simon C, Wang Y and Boyd D. (1997). Oncogene, 29, 2563–2573. Lewis TS, Shapiro PS and Ahn NG. (1998). Adv. Cancer Res., 74, 49–149. Ohoka Y, Nakai Y, Mukai M and Iwata M. (1993). Biochem. Biophys. Res. Commun., 197, 916–921. Pallas DC, Shahrik LK, Martin BL, Jaspers S, Miller TB, Brautigan DL and Roberts TM. (1990). Cell, 60, 167–176.

Oncogene

Rosenberger SF and Bowden GT. (1996). Oncogene, 12, 2301–2308. Santoro MF, Annand RR, Robertson MM, Peng YW, Brady MJ, Mankovich JA, Hackett MC, Ghayur T, Walter G, Wong WW and Giegel DA. (1998). J. Biol. Chem., 273, 13119–13128. Sato S, Fujita N and Tsuruo T. (2000). Proc. Natl. Acad. Sci. USA, 97, 10832–10837. Schulze-Osthoff K, Ferrari D, Los M, Wesselborg S and Peter ME. (1998). Eur. J. Biochem., 254, 439–459. Shenolikar S and Nairn AC. (1991). Adv. Second Messenger Phosphoprotein Res., 23, 1–121. So¨derstro¨m TS, Poukkula M, Holmstro¨m TH, Heiskanen KM and Eriksson JE. (2002). J. Immunol., 169, 2851–2860. Song Q and Lavin MF. (1994). Biochem. Biophys. Res. Comm., 190, 47–52. Sontag E, Fedorov S, Kamibayashi C, Robbins D, Cobb MH and Mumby M. (1993). Cell, 75, 887–897. Sorenson CM, Barry MA and Eastman A. (1990). J. Natl. Cancer Inst., 8, 749–755. Tran SE, Holmstrom TH, Ahonen M, Kahari VM and Eriksson JE. (2001). J. Biol. Chem., 276, 16484–16490. Von Lindern M, vaan Baal S, Wiegant J, Raap A and Grosveld G. (1992). Mol. Cell. Biol., 12, 3346–3355. Walter G and Mumby M. (1993). Biochem. Biophys. Acta, 1155, 207–226. Wright SC, Zhong J and Larrick JW. (1994). FASEB J., 8, 654–660.