Caspase-Dependent Mitochondrial Pathway Induce Apoptosis in T

Siva-1 and an Alternative Splice Form Lacking the Death Domain, Siva-2, Similarly Induce Apoptosis in T Lymphocytes via a Caspase-Dependent Mitochondrial Pathway This information is current as of September 29, 2021. Bénédicte Py, Christian Slomianny, Patrick Auberger, Patrice X. Petit and Serge Benichou J Immunol 2004; 172:4008-4017; ; doi: 10.4049/jimmunol.172.7.4008 http://www.jimmunol.org/content/172/7/4008 Downloaded from

References This article cites 49 articles, 25 of which you can access for free at:

http://www.jimmunol.org/content/172/7/4008.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 29, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2004 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Siva-1 and an Alternative Splice Form Lacking the Death Domain, Siva-2, Similarly Induce Apoptosis in T Lymphocytes via a Caspase-Dependent Mitochondrial Pathway1

Be´ne´dicte Py,* Christian Slomianny,‡ Patrick Auberger,§ Patrice X. Petit,2† and Serge Benichou2,3*

Siva-1 is a death domain-containing proapoptotic protein identiﬁed as an intracellular ligand of CD27 and of the glucocorticoid- induced TNFR family-related gene, which are two members of the TNFR family expressed on lymphoid cells. Although Siva-1 expression is up-regulated in multiple pathological processes, little is known about the signaling pathway underlying the Siva- induced apoptosis. In this study, we investigated the mechanism of the proapoptotic activity of Siva-1 and an alternative splice form lacking the death domain of Siva-1, Siva-2, in T lymphocytes in which Siva proteins, CD27, and glucocorticoid-induced Downloaded from TNFR family-related gene are primarily expressed. Overexpression of Siva proteins triggers a typical apoptotic process manifested by cell shrinkage and surface exposure of phosphatidylserine, and conﬁrmed by ultrastructural features. Siva-induced apoptosis is related to the CD27-mediated apoptotic pathway and results in activation of both initiator and effector caspases. This pathway involves a mitochondrial step evidenced by activation of Bid and cytochrome c release, and is modulated by overexpression of Bcl-2 or Bcl-xL. The determinants for Siva-induced apoptosis are not contained within the death domain found in the central part of

Siva-1, but rather in both the N-terminal and C-terminal regions shared by both Siva proteins. The N-terminal region also http://www.jimmunol.org/ participates in the translocation of both Siva proteins into the nuclear compartment. These results indicate that Siva-1 and Siva-2 mediate apoptosis in T lymphocytes via a caspase-dependent mitochondrial pathway that likely involves both cytoplasmic and nuclear events. The Journal of Immunology, 2004, 172: 4008–4017.

poptosis, or programmed cell death, plays a major role cellular triggering events or from various extracellular stimuli in- for proper functions and homeostasis of all multicellular volving engagement of death receptors (3). A organisms (1–3). In humans, this process is required in Death receptors, such as TNFR or Fas, are members of a su- development, tissue homeostasis, and elimination of extraneous or perfamily of receptors that play pivotal roles in transducing ex- damaged cells. Dysregulation of apoptosis is therefore involved in trinsic cell death or survival signals (6, 7). The TNFR superfamily by guest on September 29, 2021 a wide variety of severe pathological processes, such as neurode- consists of more than 20 structurally related type I transmembrane generative disorders, autoimmune and immunodeficiency diseases, proteins that can be divided into two subclasses depending on cancers, or viral infections (4, 5). Cells undergoing apoptosis are whether they contain a death domain (DD)4 within their cytoplas- characterized by stereotypical changes including condensation of mic region. Ligation of death receptors by their ligands provokes cytoplasmic and nuclear contents, plasma membrane blebbing, the recruitment of a death-inducing signaling complex through ho- fragmentation of nuclei, and ultimately the breakdown into mem- motypic interactions between the death cytoplasmic domain of the brane-bound apoptotic bodies that are rapidly phagocytosed. Cel- receptor and the adaptor molecules also containing a DD. Forma- lular signals leading to these apoptotic features can occur through tion of the death-inducing signaling complex then triggers a sig- multiple independent pathways that are initiated either from intra- naling cascade that leads to the final execution of cell death (8– 10). Unlike death receptors, members of the TNFR superfamily that do not contain a DD within their cytoplasmic tail primarily trans- De´partements de *Maladies Infectieuses and †Geńe´tique De´veloppement et Patholo- gies Molećulaires, Institut Cochin, Institut National de la Santeét de la Recherche duce signals for cell growth and survival. However, some receptor Me´dicale U567, Centre National de la Recherche Scientifique UMR8104, Universite´ members of the family can be involved, under different environmental ‡ Paris 5, Paris, France; Laboratoire de Physiologie Cellulaire, Institut National de la circumstances, in either cell proliferation or apoptosis (7). Santeét de la Recherche Me´dicale EMI0228, Universite´Lille 1, Villeneuve d’Ascq, France; and §Institut National de la Santeét de la Recherche Me´dicale U526, Faculte´ CD27 is such a receptor of the TNFR superfamily expressed on de Me´decine Pasteur, Nice, France B and T cells, which can equally provide stimulatory signals for Received for publication March 14, 2003. Accepted for publication January 20, 2004. cell growth as well as apoptosis (11–13). The binding to CD27 of The costs of publication of this article were defrayed in part by the payment of page its ligand CD70, a member of the TNF family also expressed on charges. This article must therefore be hereby marked advertisement in accordance discrete subpopulations of lymphocytes, results in T and B cell with 18 U.S.C. Section 1734 solely to indicate this fact. proliferation and B cell Ig production (14–17). The costimulatory 1 This work was supported in part by Institut National de la Santeét de la Recherche Me´dicale, Centre National de la Recherche Scientifique, Universite´Paris 5, Associ- function of CD27 in the generation and maintenance of T cell ation pour la Recherche contre le Cancer Grant 4493 (to P.X.P.) and by the Agence immunity was confirmed in vivo in CD27-deficient mice (18). Nationale de Recherche sur le SIDA (to S.B.). S.B. is also supported in part by However, it was documented that ligation of CD27 could also SIDAction. 2 P.X.P. and S.B. contributed equally to this work as co-senior authors. 3 Address correspondence and reprint requests to Dr. Serge Benichou, De´partement 4 Abbreviations used in this paper: DD, death domain; DDHR, DD homology region; de Maladies Infectieuses, Institut Cochin, Institut National de la SanteétdelaRe- GITR, glucocorticoid-induced TNFR family-related gene; GFP, green fluorescent cherche Me´dicale U567, Batiment Gustave Roussy, 27 Rue du Faubourg Saint- protein; DAPI, 4Ј,6Ј-diamidino-2-phenylindole; PS, phosphatidyl-serine; TRAF, Jacques, 75014 Paris, France. E-mail address: [email protected] TNFR-associated factor.

induce an apoptotic process, despite its cytoplasmic region lacking troporation. A total of 10 ϫ 106 cells were resuspended in 200 ␮l of RPMI a DD (13). This CD27-mediated apoptosis has been related to the 1640 medium with Glutamax-1 and supplemented with 10% FCS and 10 ␮ identification of a new intracellular ligand of CD27, called Siva, mM HEPES buffer. Electroporations were performed with 10 g of plasmid DNA using 4-mm gap cuvettes (EquiBio, Needham Heights, MA) in which is highly expressed in lymphoid cells and exhibits proapop- a Bio-Rad Gene Pulser II (260 V, 950 microfarads). Cells were then totic activity (13, 19). The full-length predominant form of Siva, washed once in culture medium supplemented with 10 mM HEPES buffer and referred to as Siva-1, is a 175 aa protein containing a DD homol- cultured subsequently for the indicated period of time. Where indicated, HPB- ogy region (DDHR) in its central part (13), but a minor form, ALL transfected cells were cultured for 48 h in the presence of the caspase inhibitors, Boc-Asp-fluoromethylketone (boc-D-fmk; 40 ␮M) or N-benzy- Siva-2, which is generated by alternative splicing and lacks most loxycarbonyl-Val-Ala-Asp-fluoromethylketone (z-VAD-fmk; 20 ␮M) pur- of the DDHR coding sequence, is also expressed (20). Interest- chased from Calbiochem (La Jolla, CA), or were stimulated 24 h after trans- ingly, Siva can also bind to the cytoplasmic domain of glucocor- fection by incubation with an anti-CD27 mAb (1 ␮g/ml) purchased from ticoid-induced TNFR family-related gene (GITR) (21), another Immunotech (clone 1A4-CD27; Marseilles, France). The hepatocarcinoma member of the TNFR family that is mainly expressed on T lym- cell line Hep3B (kindly provided by C. Desbois-Mouthon, Hopital Saint- Antoine, Paris, France) were maintained in MEM (Invitrogen) supple- phocytes and involved in the control of T cell activation and pro- mented with 10% FCS and 100 U/ml penicillin/streptomycin at 37°C under

grammed cell death (22). The cytoplasmic tails of CD27 and GITR 5% CO2. Where indicated, Hep3B cells were treated with cisplatin (16 display a striking homology and associate with the cysteine-rich ␮g/ml) for 16 h. region found at the C-terminal end of both Siva forms (13, 20). Flow cytometry assays Several studies show that Siva-1 is overexpressed in various pathological circumstances, such as acute ischemic injury (23) and The ﬂow cytometry analyses were performed on GFP-positive cells with a

Coxsackievirus infection (24). Moreover, up-regulation of Siva FACSCalibur cytofluorimeter (BD Biosciences, San Jose, CA) while gat- Downloaded from ing the forward and the side scatters on viable cells. A minimum of 1 ϫ 104 was also detected in hepatocarcinoma cells treated with the cis- events were acquired in list mode and analyzed with the CellQuest soft- platin anticancer drug (25), whereas the metastasis suppressor ware (BD Biosciences). Cell shrinkage, cell viability, and phosphatidyl- TIP30, which inhibits metastasis of the small cell lung carcinoma serine (PS) exposure at the cell surface were analyzed as previously de- by predisposing cells to apoptosis, also induces Siva overexpres- scribed (32). Briefly, cell shrinkage was evaluated by analysis of the light sion (26). Although Siva-1 participates in multiple apoptotic pro- scattering properties, whereas viability and PS exposure were assessed by analysis of the cell permeability to propidium iodide (1 ␮g/ml) and by cell cesses in lymphoid cells in which both CD27 and GITR are pri- surface staining with Alexa 633-conjugated annexin V (Molecular Probes, http://www.jimmunol.org/ marily expressed (13, 19, 22, 24, 27), little is known about the Eugene, OR), respectively. Caspase-3 activity was determined using the signaling pathway underlying the Siva-induced apoptosis in T caspase-3 Intracellular Activity Assay kit II purchased from Calbiochem, lymphocytes. We have thus explored the mechanism of the pro- whereas caspase-8 and caspase-9 activities were determined using the Caspase-8 Detection kit (Red-IETD-fmk) and Caspase-9 Detection kit apoptotic activity of both Siva-1 and Siva-2 in T cells. Siva-over- (Red-LEHD-fmk), purchased from Oncogene Research Products (San Di- expressing cells display stereotypical features of apoptosis, includ- ego, CA). The determination of the caspase activities was done following ing caspase activation, mitochondrial events, and morphological the manufacturer’s instructions. changes, which are modulated by overexpression of Bcl-2 or Electron microscopy Bcl-xL antiapoptotic proteins. In contrast with previous reports, we show that the DDHR of Siva-1 is dispensable for induction of After 48 h transfection, 7 ϫ 105 HPB-ALL GFP-positive cells were sorted by guest on September 29, 2021 apoptosis, whereas the main determinants required for this activity are by flow cytometry using an Epics-ELITE EST cytometer and then pro- located in both the N- and C-terminal regions shared by both Siva-1 cessed for electron microscopy analysis. The cell pellets were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.4, postfixed in 1% osmium and Siva-2 proteins. These results indicate that both Siva proteins are tetroxide in the same buffer and subsequently processed for embedding in able to mediate apoptosis in lymphoid cells through activation of a Epon resin. The contrasted sections were observed on a Hitachi H600 caspase-dependent mitochondrial signaling pathway. transmission electron microscope. Materials and Methods Western blot analysis Plasmid construction A total of 7 ϫ 105 HPB-ALL GFP-positive cells were sorted 24 or 48 h after transfection as previously described and then lysed in a buffer con- Vectors for expression of the human wild-type Siva-1, Siva-2, and deletion taining 10 mM Tris-HCl (pH 7.6), 150 mM NaCl, 2 mM EDTA, 0.5% mutants of Siva-1 (see Fig. 7A) fused to the green fluorescent protein (GFP) Triton X-100, and an antiprotease mixture (Sigma-Aldrich, St. Louis, MO). were constructed by PCR with specific primers, and the amplified frag- The protein concentration of the cell lysates was measured (Bio-Rad, Her- ments were inserted between the EcoRI and BamHI sites of pEGFPC1 cules, CA), and proteins (50 ␮g) were then separated by SDS-PAGE and (Clontech Laboratories, Palo Alto, CA). Vector for expression of Siva-1 transferred to a polyvinylidene difluoride (PVDF) membrane (Amersham fused to the flag M2 epitope at its N terminus was also constructed by PCR Biosciences, Piscataway, NJ). The membrane was saturated with 5% non- with specific primers in the pSG-flag vector (28). Vector for expression of fat dry milk in TBS containing 0.5% Tween 20 and then incubated with the CD27 receptor (pcDNA3-CD27) was kindly provided by K. V. Prasad primary Ab in the saturation solution for1hatroom temperature. The (University of Illinois, Chicago, IL) (13). following Abs were used to detect actin, pro-caspase-3, pro-caspase-8, and Cell culture and transfections Bid, respectively: a goat polyclonal anti-actin (I-19; Santa Cruz Biotech- nology), a rabbit polyclonal anti-caspase-3 (CCP32; BD PharMingen, San CD4/CD8 double-positive human peripheral blood-acute lymphocytic leu- Diego, CA), a goat polyclonal anti-caspase-8 (T-16; Santa Cruz Biotech- kemia (HPB-ALL) T cells were kindly provided by G. Bismuth (Institut nology, Santa Cruz, CA) and a goat polyclonal anti-Bid (AF860; R&D Cochin, Paris, France), and were maintained in RPMI 1640 medium with Systems, Minneapolis, MN). The membrane was then incubated with an Glutamax-1 (Invitrogen, San Diego, CA) supplemented with 10% FCS and appropriate HRP-conjugated secondary Ab (DAKO, Carpinteria, CA) in

100 U/ml penicillin/streptomycin at 37°C under 5% CO2. CD4 single- TBST for 30 min at room temperature. The proteins were visualized with positive Jurkat T cells stably expressing Bcl-2 (29) (kindly provided by N. the ECL kit (Amersham Biosciences), and the signals were quantiﬁed using

Israel, Institut Pasteur, Paris, France) or Bcl-xL (30) (kindly provided by K. NIH Image software. The expression of the endogenous Siva protein, either Schulze-Osthoff, University of Muenster, Muenster, Germany) as well as in HPB-ALL cells or in Hep3B cells treated with cisplatin, was determined the control Jurkat cells stably transfected with the neomycin-resistant vec- by Western blot analysis on the cell lysate (25 ␮g of total protein) using tor were cultured as HPB-ALL cells, but in the presence of 500 ␮g/ml anti-Siva (C-20; Santa Cruz Biotechnology). As a control, ␣-tubulin was G418 (Invitrogen). SKW6.4 B cells stably overexpressing Bcl-2 and the detected using anti-␣-tubulin (TU-02; Santa Cruz Biotechnology). The ex- control SKW6.4 cells stably transfected with the neomycin-resistant vector pression of wild-type Siva-1, Siva-2, and the Siva-1 deletion mutants was (kindly provided by B. Fadeel, Institute of Environmental Medicine, Stock- monitored by Western blotting of total lysates of 1 ϫ 106 HPB-ALL trans- holm, Sweden) were cultured as already described (31). Transient transfected cells using a mixture of anti-GFP mAbs (clones 7.1 and 13.1; Boer- fections of HPB-ALL, Jurkat, and SKW6.4 cells were performed by elec- hinger Mannheim, Mannheim, Germany). 4010 SIVA-INDUCED APOPTOSIS IN T LYMPHOCYTES

Immunoﬂuorescence staining HPB-ALL cells (2 ϫ 106) were concentrated in 300 ␮l of culture medium without FCS and put on cover glass pretreated with poly-L-lysine (Sigma- Aldrich). After adhesion, cells were washed in PBS, ﬁxed with 4% para- formaldehyde, and treated with 0.1 M glycine. Cells were then permeabilized for 30 min at room temperature with 0.05% saponin (Sigma-Aldrich) in PBS containing 0.2% BSA for anti-cytochrome c staining. For anti-GFP staining, cells were permeabilized either with 0.1% Triton X-100 for 10 min or with 0.05% digitonin (Sigma-Aldrich) in a buffer containing Tris- HCl (pH 7.5) 10 mM, NaCl 150 mM, EDTA 1 mM for 5 min. Cytochrome c was detected with the 6H2.B4 mAb (BD PharMingen) and GFP fusions with the mixture of anti-GFP mAbs, followed by incubation with a Cy3- conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA). Cells were then mounted in Vectashield mounting medium (Vec- tor Laboratories, Burlingame, CA) containing 4Ј,6Ј-diamidino-2-phenylindole (DAPI) for staining of the nuclear DNA. Images were acquired with a Zeiss Axiophot microscope 80ϫ (Analytical Imaging Facility, Albert Einstein College of Medicine, Bronx, NY) equipped with a CCD camera controlled by Metaview software (Universal Imaging, Downingtown, PA).

Results Overexpression of Siva in T lymphocytes results in cell death Downloaded from induction The capacity of the full-length Siva-1 protein (see Fig. 7A)to trigger an apoptotic process in lymphoid cells was investigated by transient overexpression in T cell lines. A vector for expression of Siva-1 fused at its N terminus to the GFP was generated and used http://www.jimmunol.org/ to transfect HPB-ALL (Fig. 1) or Jurkat T cells (data not shown). We first explored whether Siva overexpression induces a cell death process associated with some features of apoptosis. HPB-ALL cells were transfected with the GFP-Siva-1 expression vector, and FIGURE 1. Cell death induction in T lymphocyte overexpressing 48 h later, cell shrinkage, cell viability, and PS exposure at the Siva-1. HPB-ALL T cells were transfected with a vector for expression of outer face of the plasma membrane were assessed by flow cytom- either GFP-Siva-1 or GFP. GFP-positive cells were then assayed by flow cytometry for cell shrinkage, cell viability, and PS exposure at the cell etry on GFP-positive cells. As shown in Fig. 1A, 67.5% of cells surface. A, Cell shrinkage was evaluated by a biparametric analysis of overexpressing GFP-Siva-1, vs only 16.3% of control cells ex- forward scatter and side scatter. B, Cell viability and PS exposure were pressing GFP, had a reduced forward scatter and a slightly in- analyzed by propidium iodide (PI) permeability and cell surface staining by guest on September 29, 2021 creased side scatter, corresponding to a smaller size and a higher using annexin V-Alexa 633, respectively. A and B, Data are representative cell granularity reminiscent of cell shrinkage. Additionally, a of four independent experiments and were measured 48 h after transfection. higher percentage of cells expressing GFP-Siva-1 (51% vs Ͻ10% C, Kinetic analysis of GFP-Siva-1 (Ⅺ) or GFP (f) cells displaying surface of GFP expressing cells) showed PS exposure as detected by the PS exposure evaluated at the indicated time after transfection by cell sur- cell surface annexin V staining (Fig. 1B). This significant PS ex- face annexin V staining. Values are the mean of four independent exper- ternalization was observed among GFP-Siva-1 expressing cells as iments. Error bars represent 1 SD from the mean. early as 24 h after transfection, with progressive exposure exhib- ited in following days (Fig. 1C). The PS exposure was accompa- nied by a decrease of the mitochondrial membrane potential in a condensed at the nuclear periphery and detached from the nuclear significant fraction of GFP-Siva-1-expressing cells (data not membrane. Compared with control cells, mitochondria from GFP- shown), suggesting that mitochondria contribute to the Siva-in- Siva-1 expressing cells were electron translucent and swollen, and duced cell death. A similar cell death induction was observed by often appeared either with a bulging swollen matrix or with a overexpression of the well-characterized DD-containing proapop- partly disrupted outer membrane (Fig. 2, lower panels), suggesting totic Fas-associated DD protein in HPB-ALL cells (data not that intermembrane space proteins were released into the cytosol shown). Altogether, these results indicate that overexpression of during cell death. All these morphological changes demonstrate Siva-1 in T lymphocytes leads to an apoptotic cell death process. that Siva-1 overexpression results in the induction of a stereotypical apoptotic cell death, including mitochondrial damage events. Siva-induced cell death displays morphological features of apoptosis Siva participates in the CD27-mediated apoptotic pathway in T The ultrastructural features of HPB-ALL cells overexpressing lymphocytes Siva-1 were analyzed by transmission electron microscopy to fur- To analyze the role of Siva in apoptotic signaling pathways, we ther define the morphological changes induced by Siva-1 in T lym- first checked that the level of Siva overexpression induced in our phocytes. HPB-ALL cells expressing either GFP-Siva-1 or GFP transient experimental system was comparable to the up-regulation were sorted 48 h after transient transfection by flow cytometry, of Siva expression observed in some physiopathological processes, fixed, and then processed for electron microscopy (Fig. 2). Al- such as cell treatment with the cisplatin anticancer drug. In agree- though control cells overexpressing the GFP protein appeared with ment with the results previously reported (25), Hep3B hepatocar- a normal phenotype (Fig. 2, right panels), dying cells expressing cinoma cells treated with cisplatin led to a 6.2-fold overexpression GFP-Siva-1 displayed many characteristic changes of apoptosis of the Siva protein (Fig. 3A). This up-regulation was similar to (Fig. 2, left panels). The vacuolar system as well as the endoplas- what we estimated from the Western blot analysis of HPB-ALL mic reticulum were dramatically expanded, and the chromatin was transfected cells using anti-Siva Abs to detect both GFP-Siva and The Journal of Immunology 4011

expression vector, HPB-ALL cells were cultured for 2 days in the presence of the broad-spectrum caspase inhibitors, boc-D-fmk and z-VAD-fmk. GFP-positive cells were then assessed for PS exposure and cell viability by flow cytometry analysis. As shown in Fig. 4A, both caspase inhibitors largely decreased the level of GFP-Siva-1-expressing cells displaying PS exposure and also pre- served the cell viability (data not shown), indicating that Siva-1 induces apoptosis in a caspase-dependent manner. Direct activation of caspases by Siva-1 overexpression was further explored by analysis of activation of caspase-8 and caspase-3, which are the major initiator and effector caspases, respectively (8). HPB-ALL cells expressing GFP-Siva-1 were sorted 24 and 48 h after transfection, and caspase-3- and caspase-8-activation was first analyzed by Western blot detection of the proteolytic cleavage of the pro- caspase precursors (Fig. 4B). Compared with untransfected cells or cells expressing the GFP control, a significant decrease of the pro- caspase-8 was already observed in cell lysate from GFP-Siva-1- expressing cells as soon as 24 h after transfection, whereas the

decrease of the pro-caspase-3 could be detected only 48 h after Downloaded from transfection, even though a similar amount of protein was loaded, as evidenced by probing the same blots with an anti-actin Ab. This FIGURE 2. Morphological features of T lymphocytes expressing suggests that activation of the initiator caspase-8 precedes activa- Siva-1. HPB-ALL cells were transfected with either the GFP-Siva-1 (left tion of the effector caspase-3 in the Siva apoptotic pathway. In panels) or GFP (right panels) expression vector. GFP-positive cells were addition, the Siva-induced apoptosis was significantly reduced by sorted by flow cytometry 48 h later, fixed, and then processed for trans- expression of the viral CrmA inhibitor (data not shown), which http://www.jimmunol.org/ mission electron microscopy analysis as described in Materials and Meth- inhibits caspase-8 without affecting caspase-3 activity (33). ods. General cellular (top panels) morphology (scale bar, 2 ␮m) and mor- To verify that the disappearance of the pro-caspase forms pro- phological details (bottom panels) (scale bar, 500 nm) of the mitochondria voked by Siva-1 expression was related to a specific cleavage lead- are shown. Arrows indicate mitochondrial alterations. cCh, condensed ing to active caspases, direct induction of the caspase-3, caspase-8, chromatin; dNE, detachment from the nuclear envelope; ER, endoplasmic and also caspase-9 activity was finally documented in a flow cy- reticulum; V, vacuoles. tometry assay. HPB-ALL cells were transfected with either GFP or GFP-Siva-1 expression vector, and caspase activities were assayed on GFP-positive cells using specific substrates leading to a fluo- the endogenous Siva protein (Fig. 3B). Because the transfection rescent product after cleavage (Fig. 4C). Caspase-8 and caspase-9 by guest on September 29, 2021 efficiency of the GFP-Siva expression vector in HPB-ALL cells activities were detectable in a significant proportion of GFP-Siva- was ϳ20% as evaluated by FACS analysis (Fig. 3B), we calculated 1-expressing cells as soon as 24 h after transfection, whereas that a 5-fold overexpression of Siva is induced in our transient caspase-3 activity was not yet detected. We have checked that the expression experimental system. caspase activities assessed in this flow cytometry assay were in- Because it was reported that Siva interacts with the cytoplasmic hibited by treatment of the Siva-expressing cells with the z-VAD- tail of the CD27 receptor (13), we investigated whether the apo- fmk caspase inhibitor (data not shown). These results demonstrate ptosis process triggered by Siva overexpression was related to the that Siva-1 expression induces apoptosis in T lymphocytes through CD27-mediated apoptotic pathway. HPB-ALL cells were cotrans- activation of a caspase-dependent pathway. This signaling path- fected with a vector expressing CD27 in combination with either way involves activation of both initiator and effector caspases, GFP or GFP-Siva-1 expression vector, and stimulated 24 h later such as caspase-8, -9, and -3. with an anti-CD27 Ab directed against the extracellular region of CD27. The percentage of apoptotic cells was then evaluated by cell The Siva-1 apoptotic pathway triggers mitochondrial events surface PS exposure on GFP-positive cells (Fig. 3C). In nonstimu- The mitochondria play a pivotal role in multiple caspase-depen- lated cells, coexpression of GFP-Siva-1 in combination with CD27 dent apoptotic pathways (9, 10). Because the main molecular link did not affect the proapoptotic activity of Siva-1 alone. In contrast, connecting caspase and mitochondrial pathways is the caspase-8- there was a significant increase of cell surface PS exposure after mediated activation of Bid, a proapoptotic member of the Bcl-2 stimulation with anti-CD27 in cells coexpressing GFP-Siva-1 and protein family (34), we initially explored whether overexpression CD27. This stimulatory effect is specific because it was not ob- of Siva-1 leads to the activating proteolytic cleavage of Bid. HPB- served in cells expressing GFP-Siva-1 alone or in cells expressing ALL cells were transfected with either GFP-Siva-1 or GFP ex- CD27 in combination with the GFP control. These results indicate pression vector and then sorted 24 and 48 h later by flow cytom- that Siva-1 is involved in the apoptotic pathway triggered by CD27 etry. Processing of Bid was evaluated on GFP-positive cells by stimulation. Western blot using an anti-Bid Ab (Fig. 5A). Compared with untransfected and GFP-expressing cells, lysate from GFP-Siva-1-ex- Siva-induced apoptosis is related to a caspase-dependent pressing cells showed a marked disappearance of Bid as soon as pathway 24 h after transfection, whereas probing the same blot with anti- Caspase proteases are key factors involved in both the initiation actin evidenced that a similar amount of cellular proteins was and execution phases of most apoptotic processes (8). To charac- loaded. This indicates that Siva-1 overexpression induces the ac- terize the signaling pathway triggered by Siva-1 overexpression, tivation of Bid, which usually results in the permeabilization of we first investigated whether caspase inhibitors could block the mitochondrial membranes and the release of mitochondrial apop- Siva-induced apoptosis. After transfection with the GFP-Siva-1 togenic components, such as cytochrome c (10). We thus checked 4012 SIVA-INDUCED APOPTOSIS IN T LYMPHOCYTES Downloaded from

FIGURE 3. Siva overexpression increases the apoptotic process triggered by CD27 ligation. A, Hep3B cells were cultured in the presence or not of cisplatin for 16 h. An identical amount (25 ␮g) of protein from cell lysates was then separated by SDS-PAGE, and expression of the endogenous Siva protein was analyzed by Western blotting with anti-Siva Abs. The intensity of the bands was quantified using NIH Images Software and the Siva signal (top panel) was normalized to the signal corresponding to ␣-tubulin (bottom panel). The values indicated represent the percentage of the signal intensity relative to nontreated cells (100%), and are representative of three independent experiments. B, HPB-ALL cells were transfected with the GFP-Siva-1 http://www.jimmunol.org/ expression vector. The transfection efficiency was determined by flow cytometry analysis (upper panel), and Siva protein expression was analyzed by Western blotting with anti-Siva (lower left panel) and anti-GFP (lower right panel) Abs. The intensity of the bands was quantified using NIH Images Software, and the values indicated between parentheses represent the percentage of the signal intensity relative to the endogenous Siva protein (100%). C, HPB-ALL cells were cotransfected with either GFP or GFP-Siva-1 expression vector in combination or not with a vector expressing CD27 (pcDNA3- CD27). Twenty-four hours after transfection, cells were stimulated with an anti-CD27 Ab, and GFP-positive cells were then assayed 24 h later for PS exposure by annexin V staining and flow cytometry. Values are the mean of three independent experiments. Error bars represent 1 SD from the mean. by guest on September 29, 2021 that the Siva-mediated apoptosis led to cytochrome c release from abrogated by overexpression of either Bcl-2 or Bcl-xL antiapop- mitochondria by indirect immunofluorescence staining. As shown totic proteins. Because studies on the Fas-mediated apoptosis sig- in Fig. 5B, Siva-1 expression induced a change in the cytochrome naling pathways have highlighted two types of cells (35) depend- c staining pattern. A typical punctuate mitochondrial profile was ing on the requirement (type II cells such as Jurkat cells) or not observed in a large majority of both untransfected and GFP-ex- (type I cells) of a mitochondrial amplification loop, we explored pressing cells, whereas a significant proportion of GFP-Siva-1- the Siva-induced apoptosis in a type I cell line such as SKW6.4 B expressing cells displayed a diffuse cytosolic distribution of cyto- lymphoid cells overexpressing Bcl-2 (35). In contrast to Jurkat chrome c as soon as 24 h posttransfection. Quantitative analysis cells, overexpression of Bcl-2 in SKW6.4 cells only reduced by from immunofluorescence images showed that around 30 to 40% 50% the proportion of apoptotic GFP-Siva-1-expressing cells (Fig. of GFP-Siva-1-expressing cells display a mitochondrial release of 5D). This result indicates that the Siva-induced apoptosis is also, at cytochrome c 2 days after transfection, whereas such a phenotype least in part, modulated by antiapoptotic members of the Bcl-2 was observed in fewer than 5% of cells expressing the GFP con- protein family in type I cells. Altogether, the data reported in Fig. trol. As previously mentioned, overexpression of Siva-1 also pro- 5 indicate that the mitochondria form a pivotal step in the signaling voked a marked drop in the membrane mitochondrial potential apoptotic pathway triggered by Siva-1 both in type I and type II (data not shown), thus confirming that the Siva-1 apoptotic path- lymphocytes. way includes a mitochondrial step in T lymphocytes. In contrast to Bid, the antiapoptotic members of the Bcl-2 pro- Siva-1 accumulates in nuclear dot-like structures tein family, such as Bcl-2 and Bcl-xL, are known to preserve the integrity of the mitochondrial membrane (34). We thus explored To analyze the subcellular distribution of Siva-1 in T lymphocytes, whether the Siva-induced apoptosis in T cells could be modulated HPB-ALL cells were transfected with the GFP-Siva-1 expression by overexpression of Bcl-2 or Bcl-xL. Jurkat T cells overexpress- vector and the fusion was then visualized by direct fluorescence at ing either Bcl-2 or Bcl-xL were transiently transfected with the various time points after transfection. GFP-Siva-1 was detected by GFP-Siva-1 expression vector, and cell surface PS exposure was Western blot analysis as soon as 3 h after transfection (data not analyzed 48 h later on GFP-positive cells (Fig. 5C). Expression of shown), and accumulated in the nucleus in a diffuse staining pat- GFP-Siva-1 in control Jurkat cells stably transfected with the neo- tern (Fig. 6A). This nucleoplasmic localization was still observable mycin vector resulted in an apoptotic level absolutely similar to 6 and 9 h after transfection, but a fraction of the fusion appeared that observed in HPB-ALL T cells. In contrast, the percentage of to form dot-like structures. At 24 h after transfection, GFP-Siva-1 GFP-Siva-1-expressing cells displaying a cell surface exposure of was almost exclusively localized into dot-like structures, whereas PS was reduced to a level corresponding to the GFP control in cells the GFP control showed a diffuse nucleo-cytoplasmic staining dis- coexpressing Bcl-2 or Bcl-xL. These results show that the pro- tribution. The GFP-Siva-1-positive dots were often found at the apoptotic activity of GFP-Siva-1 in Jurkat T cells can be totally periphery of the nucleus, and their number varied between 5 and The Journal of Immunology 4013

X-100 or digitonin and using anti-GFP Abs (Fig. 6B). Digitonin permeabilizes the plasma membrane without affecting the integrity of the nuclear envelope, whereas Triton X-100 permeabilizes all the cellular membranes. In cells permeabilized with Triton X-100, GFP-Siva-1 detected with anti-GFP stained dot-like structures that colocalized perfectly with the intrinsic fluorescence of the fusion. By contrast, no staining was detected with anti-GFP on digitonin- permeabilized cells, even though GFP-Siva-1 expression was evidenced by the dots visualized by direct fluorescence. Because digitonin permeabilization also preserves the integrity of the mitochondrial membrane, we have checked that GFP-Siva-1 did not colocalize with mitochondrial markers (data not shown). These localization studies indicate that Siva-1 is predominantly a nuclear protein that accumulates into dot-like structures. Despite our ef- forts to characterize the Siva-positive nuclear structures, we failed to colocalize Siva-1 with specific markers of known nuclear sub- structures (36) such as promyelocytic leukemia oncoprotein bodies, nucleoli, nuclear pore complexes, or nuclear lamina (data not shown). Downloaded from Deleted forms of Siva-1 corresponding to the three distinct regions of the 175 aa long protein (see Fig. 7A) were then generated to determine which regions were responsible for the nuclear-dotted localization of the protein. Although Siva-1 is the predominant form expressed in T cell populations, the minor Siva-2 form, generated by alternative splicing and lacking most of the DDHR, is http://www.jimmunol.org/ also produced (19, 20). HPB-ALL cells expressing the wild-type or deleted GFP-Siva forms were examined 24 h after transfection by direct fluorescence microscopy (Fig. 7). Whereas both GFP- SivaDDHR and GFP-SivaC were equally distributed between the nucleus and the cytoplasm, the GFP-SivaN mutant still retained the ability of GFP-Siva-1 to stain nuclear dot-like structures. A similar dotted staining was observed in cells expressing the GFP-Siva-2

form. These findings indicate that the determinants required to by guest on September 29, 2021 FIGURE 4. Caspase activation in the Siva-induced apoptotic process. target Siva-1 and Siva-2 into the nuclear dot-like structures are HPB-ALL cells were transfected with either the GFP-Siva-1 or GFP expression vector. A, Inhibition of Siva-induced PS exposure by caspase in- contained within the N-terminal part of the proteins. hibitors. Transfected cells were cultured for 48 h in the presence or not of the boc-D-fmk and z-VAD-fmk caspase inhibitors, and GFP-positive cells were then assayed for PS exposure by annexin V staining and flow cytom- The N- and C-terminal regions of Siva-1 and Siva-2 proteins etry. Values are the mean of four independent experiments. Error bars contain the proapoptotic determinants represent 1 SD from the mean. B, Activation of pro-caspase-8 and pro- Because the Siva-2 form lacks most of the DDHR (see Fig. 7A), caspase-3 in Siva-1-expressing cells. GFP- or GFP-Siva-1-expressing cells we explored whether Siva-2 was also able to induce apoptosis in T were sorted 24 h (left panels)or48h(right panels) after transfection by flow cytometry. An identical amount (50 ␮g) of protein lysate from trans- lymphocytes to characterize the role of the DDHR in the activity fected and untransfected cells (mock) was then separated by SDS-PAGE, of the protein. As evidenced by caspase-3 activation, GFP-Siva-2 and proteolytic activation of pro-caspase-8 and pro-caspase-3 was analyzed overexpression led to an apoptosis induction as efficient as the by Western blotting with specific anti-pro-caspase Abs. The intensity of the full-length Siva-1 form in HPB-ALL cells (Fig. 8A). Similar re- bands was quantified using NIH Images Software and the pro-caspase-8 sults were obtained by analyzing PS exposure on GFP-Siva-2-ex- (upper panel) and pro-caspase-3 (middle panel) signals were normalized to pressing cells (data not shown). These results show that both Siva the signal corresponding to actin (lower panel). The values indicated rep- forms display proapoptotic activity, and also indicate that the resent the percentage of the signal intensity relative to untransfected cells DDHR of Siva-1 is dispensable for this activity. (100%, mock), and are representative of three independent experiments. C, The three deletion Siva mutants (SivaN, SivaDDHR, and SivaC) Caspase-3, caspase-8, and caspase-9 activities in cells expressing Siva-1. At the indicated time after transfection, GFP-positive cells were analyzed were then expressed in HPB-ALL cells and assayed for induction by flow cytometry for caspase activities using specific substrates. Values of PS exposure at the cell surface to identify the regions of the are the mean of three independent experiments. Error bars represent 1 SD protein containing the determinants responsible for the proapop- from the mean. totic effects. As indicated in Fig. 8B, the kinetic of cells displaying PS exposure was similar in cells expressing either the GFP-SivaN, GFP-SivaC, or GFP-Siva-1 fusions. In contrast, PS exposure in 20. In addition, identical dotted structures were detected by indi- cells expressing the GFP-SivaDDHR mutant was not significantly rect immunofluorescence on HPB-ALL cells expressing a flag- different from the GFP-expressing cells, even though this mutant tagged Siva-1 form (Fig. 6A), indicating that this typical localiza- was expressed at the same level as the SivaN and SivaC mutants tion is independent of the epitope tag added to the Siva-1 sequence. (see Fig. 7B). These results confirm that the DDHR of Siva-1 is not We finally checked that the GFP-Siva-1-positive structures were required for the induction of apoptosis in T lymphocytes, whereas localized inside the nucleus by immunostaining performed on both the N-terminal and the C-terminal cysteine-rich regions par- HPB-ALL transfected cells permeabilized with either Triton ticipate in this activity. 4014 SIVA-INDUCED APOPTOSIS IN T LYMPHOCYTES Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021 FIGURE 5. Mitochondrial activation in the Siva-induced apoptotic process. A, Activation of Bid in Siva-1-expressing cells. HPB-ALL cells were transfected with either the GFP-Siva-1 or GFP expression vector. GFP-positive cells were sorted 24 h (left panels)or48h(right panels) after transfection by flow cytometry. An identical amount (50 ␮g) of protein lysate from transfected and untransfected cells (mock) was then separated by SDS-PAGE, and proteolytic activation of Bid was analyzed by Western blotting performed on cell lysates with anti-Bid Abs. The intensity of the bands was quantified using NIH Images Software and the Bid signal (upper panels) was normalized to the signal corresponding to actin (lower panels). The values indicated represent the percentage of the signal intensity relative to untransfected cells (100%, mock), and are representative of three independent experiments. B, Immuno- fluorescence analysis of mitochondrial cytochrome c release. Forty-eight hours after transfection, HPB-ALL cells expressing GFP-Siva-1 (upper panels) or GFP (lower panels) were fixed, permeabilized, and cytochrome c (right panels) was detected with a mouse anti-cytochrome c mAb followed by staining with Cy3-conjugated anti-mouse IgG. Nuclear DNA was stained with DAPI (left panels). Scale bar represents 20 ␮m. The percentage of cells expressing GFP or GFP-Siva-1 displaying cytochrome c release at different times after transfection is indicated on the right. Values are the mean of three independent experiments and were determined from immunofluorescence analysis on at least 100 transfected cells. Error bars represent 1 SD from the mean. C, Siva-1 activity in Jurkat T cells stably overexpressing Bcl-2 or Bcl-xL. Cells were transiently transfected with the GFP-Siva-1 or GFP expression vector. Neomycin (Neo) corresponds to control Jurkat cells stably transfected with the neomycin vector, and transiently transfected with the GFP-Siva-1 or GFP expression vector. GFP-positive cells were assayed 48 h later for PS exposure by annexin V staining. Values are the mean of three independent experiments. Error bars represent 1 SD from the mean. D, Relative proapoptotic activity of Siva-1 in Jurkat and SKW6 cells overexpressing Bcl-2. Cells stably overexpressing Bcl-2 or control cells (Neo) stably transfected with the neomycin vector were transiently transfected with the GFP-Siva-1 or GFP expression vector. Siva-1 activity was assessed as described (C), but the results are expressed as the percentages of the specific activity determined in Bcl-2 overexpressing cells relative to that determined in control neomycin cells. Values are the mean of three independent experiments. Error bars represent 1 SD from the mean.

Discussion induced cell death. Indeed, both initiator caspase-8 and caspase-9, Overexpression of Siva-1 was observed in several physiopatho- and effector caspase-3 are activated in Siva-1 expressing cells. Fi- logical apoptosis processes (23–26), but the signaling pathway in- nally, the Siva apoptotic pathway requires, at least in type II Jurkat volved in these processes has not yet been characterized. We have cells, a mitochondrial step initiated by alteration of the membrane thus explored the proapoptotic activity of Siva proteins in lym- potential and resulting in the cytosolic release of mitochondrial phoid cells in which both Siva proteins and their cellular interac- components, such as cytochrome c. Both pro- and antiapoptotic tion partners, i.e., CD27 and GITR, are highly expressed (13, 15, members of the Bcl-2 protein family (34) largely regulate these 16, 19, 21, 22). The results indicate that overexpression of Siva-1 mitochondrial events. in T lymphocytes triggers the activation of a caspase-dependent Both the N- and C-terminal domains of the Siva proteins equally death pathway including a pivotal mitochondrial step. Because this display the ability to induce a cell death process in T lymphocytes. apoptotic activity is greatly affected by caspase inhibitors, this in- These two regions could act in synergy in the same apoptotic path- dicates that caspases are the primary executioners of the Siva- way or participate in various pathways because the two domains The Journal of Immunology 4015

FIGURE 6. Cellular localization of GFP-Siva in T lymphocytes. HPB-ALL cells were transfected with the GFP-Siva-1, flag-Siva-1, or GFP expression vector. A, Cells were fixed at the indicated time after transfection. Although cells expressing GFP-Siva-1 and GFP were directly examined by fluorescence microscopy, cells expressing flag-Siva-1 were permeabilized with Triton X-100 before staining by immunofluorescence with an anti-flag mAb, and then examined by fluorescence microscopy. B, Cells expressing GFP-Siva-1 were fixed 24 h after transfection and permeabilized with either Tri- ton X-100 (upper panels) or digitonin (lower panels). Cells were visualized by direct fluorescence for exami- nation of the GFP-Siva-1 distribution or processed for indirect immunofluorescence with anti-GFP before exam- Downloaded from ination by fluorescence microscopy. Nuclear DNA was stained with DAPI. Scale bar represents 10 ␮m. http://www.jimmunol.org/

show a distinct pattern of intracellular distribution when they are Zn-finger motif (13). This cysteine-rich region of Siva proteins is separately expressed. The C-terminal domain equally distributes involved in the association with CD27 and GITR, and thus plays between the cytoplasm and the nucleus, whereas the N-terminal a pivotal role for the Siva-transducing apoptotic signal provided by region mediates the accumulation of both Siva forms into nuclear ligation of these receptors (13, 21). Interestingly, some TNFR- dot-like structures. Therefore, Siva-1 and Siva-2 are nuclear pro- associated factor (TRAF) proteins, which like Siva contain both teins that initially show a diffuse nucleoplasmic localization and Ring- and Zn-finger motifs, also directly interact with the CD27 by guest on September 29, 2021 subsequently translocate in dot-like structures at steady state. In- and GITR cytoplasmic tails (43, 44). Because it is proposed that terestingly, many recent studies reported that some DD-containing TRAF proteins function as important mediators of antiapoptotic proteins, including DAXX, DEDAF, DEDD, or TRADD are in- signals (45), we can hypothesize that Siva and TRAF molecules volved in the regulation of various apoptotic pathways by playing are respectively involved in either the cell death or survival signals specific roles in either the cytoplasmic or nuclear compartments mediated by CD27 or GITR in T lymphocytes. (34, 37–42). Similarly, Siva-1 and Siva-2 could be nucleo-cyto- In contrast to the N- and C-terminal regions, the central part of plasmic shuttling proteins that lead to the activation of distinct the long Siva-1 form containing the DDHR is not necessary for the cytoplasmic and nuclear apoptotic pathways through the two N- proapoptotic activity of the protein, at least in T lymphocytes. and C-terminal modular domains, respectively. Although in some DD-containing proteins (46), such as TRADD, Although the C terminus of the Siva proteins is required for the DD is sufficient by itself to trigger apoptosis (42, 47, 48), the direct association with the cytoplasmic domain of the CD27 and Siva-1 DDHR is, like the DD of the Fas-associated DD protein, GITR receptors (13, 21), the N-terminal region likely forms the devoid of intrinsic apoptotic activity (49). Overexpression of a effector domain of the proteins in the nucleus. In agreement with deletion mutant covering this region is not able to induce apopto- this modular organization, it was reported that Siva-1 directly in- sis, whereas the Siva-2 form, lacking most of this region, still teracts with the antiapoptotic Bcl-xL protein, but not with Bcl-2, retains an apoptotic activity. These data are in contradiction with through a putative amphipathic helical structure encompassing the the previous reports (20, 27) indicating that both human and mu- N-terminal region of the protein (19). This binding may specifi- rine Siva-2 proteins lacked the proapoptotic activity of Siva-1 in cally inhibit Bcl-xL activity, leading to an increased susceptibility nonlymphoid cells. Again, these discrepancies may reflect differ- of Siva-1-expressing cells to UV radiation-induced apoptosis. ences in the mechanism used by Siva proteins to induce apoptosis However, these observations do not fit with our data showing that in T lymphocytes. both Bcl-2 and Bcl-xL are equally able to prevent the apoptotic The signaling pathway triggered by Siva-1 in T cell lines results process provoked by Siva-1 overexpression. Because the previous in both activation of caspases and induction of mitochondrial analyses were done in nonlymphoid cells (13, 19, 27), our results events. Although our analysis suggests that Siva-1 is involved in suggest that Siva-1 uses distinct mechanisms to induce apoptosis various apoptotic pathways, it likely participates in a prototypical in T lymphocytes. Siva-1 is indeed a ubiquitous protein (13) that death receptor-initiated pathway in lymphoid cells. Siva-1 interacts could participate in various apoptotic pathways in a cell type-de- through its C-terminal region with the cytoplasmic tails of CD27 pendent manner. and GITR (13, 20, 22), two members of the TNFR family that, in The N-terminal domain of Siva-1 and Siva-2 does not show any contrast to Fas or TNFR, lack a DD in their cytoplasmic domain homology with known proteins, but the C-terminal region contains (7). Association of Siva-1 with CD27 or GITR may thus mediate 13 cysteine residues that potentially form a Ring-finger and also a subsequent activation of the initiator caspase-8. However, we 4016 SIVA-INDUCED APOPTOSIS IN T LYMPHOCYTES Downloaded from http://www.jimmunol.org/

FIGURE 8. Apoptotic activity of Siva deletion mutants. HPB-ALL cells were transfected with the vector for expression of the wild type or deleted forms of Siva-1 and Siva-2. A, Caspase-3 activity is shown in cells ex- FIGURE 7. The N-terminal region of Siva-1 and Siva-2 is responsible pressing Siva-1, Siva-2, and SivaDDHR. Forty-eight hours after transfec- for the nuclear translocation of the proteins. A, Schematic representation of tion, GFP-positive cells were analyzed by flow cytometry for caspase-3 by guest on September 29, 2021 the Siva-1, Siva-2, and Siva deletion mutants is shown. The 175 aa long activity as indicated in Fig. 4. B, PS exposure on cells expressing the wild Siva-1 protein corresponds to the predominant form expressed, whereas type or deleted forms of Siva-1 is shown. The percentage of GFP-positive Siva-2 is the shorter form generated by alternative splicing. The DDHR (o) cells displaying surface PS exposure was evaluated at the indicated time and the cysteine-rich region (f) are shown. Numbering of the N- and after transfection by cell surface annexin V staining. Values are the mean C-terminal ends of the deletion mutants refers to the Siva-1 form. B, Expres- of four independent experiments. Error bars represent 1 SD from the mean. sion of GFP-Siva-1, GFP-Siva-2, and GFP-Siva deletion mutants in HPB-ALL cells is shown. Lysates from cells transfected with the vector for expression of the wild type or deleted forms of Siva-1 and Siva-2 were analyzed by Western preceding caspase-3 activation in Siva-induced cell death. How- blotting with anti-GFP Abs. C, Cellular distribution of GFP-Siva deletion mu- ever, the results obtained in SKW6.4 type I cells suggest that the tants. Twenty-four hours after transfection, HPB-ALL cells expressing GFP- Siva-induced activation of caspase-8 can also bypass the mito- Siva-1, GFP-Siva-2, or GFP-Siva deletion mutants were fixed and directly examined by fluorescence microscopy. Nuclear DNA was stained with DAPI chondria to directly activate the caspase-3. (middle row). Scale bar represents 10 ␮m. In summary, our results show that both Siva-1 and Siva-2 display proapoptotic activity in T lymphocytes. Because Siva proteins participate in CD27- and GITR-mediated apoptosis pathways (this failed to detect a direct binding of Siva-1 to caspase-8 in two- study and Ref. 22), they likely contribute to lymphocyte function hybrid assay (data not shown), and Siva probably requires a down- and homeostasis regarding the essential roles of CD27 and GITR stream adaptor protein not yet identified. Because DDs usually in the T cell functions in vivo (18, 44). Because several recent form an interface for homotypic interactions (46), the DDHR of studies indicate that the expression of Siva-1 is modulated in mul- Siva-1 could similarly form a module required for homodimeriza- tiple pathological circumstances, further investigations will aim to tion of the protein (B. Py and S. Benichou, unpublished observa- pinpoint the physiological roles of Siva proteins to understand tions) and for interactions with some other DD-containing pro- their contribution in these pathological processes. teins. Caspase-8 then leads to activation of effector caspases either directly or through a Bid-mediated mitochondrial pathway (35). Acknowledgments Siva-induced apoptosis resulted in proteolytic activation of Bid We thank G. Bismuth, N. Israel, K. V. Prasad, K. Schulze-Osthoff, and release of cytochrome c from mitochondria, and was effi- B. Fadeel, and C. Desbois-Mouthon for the kind gift of reagents, A. Benmerah and I. Kim for active support and critical reading of the ciently blocked by overexpression of Bcl-2 or Bcl-xL proteins, at least in type II Jurkat cells. Cytosolic cytochrome c then associates manuscript, and E. Dewailly for skillful technical assistance. with Apaf-1 in a complex with caspase-9 (the apoptosome) allow- ing activation of effector caspases, such as caspase-3, and subse- References 1. Steller, H. 1995. Mechanisms and genes of cellular suicide. Science 267:1445. quently the final execution of the apoptotic process (9, 10). In 2. Jacobson, M. D., M. Weil, and M. C. Raff. 1997. Programmed cell death in agreement with such a pathway, we observed caspase-9 activity animal development. Cell 88:347. The Journal of Immunology 4017

3. Vaux, D. L., and S. J. Korsmeyer. 1999. Cell death in development. Cell 96:245. 28. Depienne, C., P. Roques, C. Creminon, L. Fritsch, R. Casseron, D. Dormont, 4. Roulston, A., R. C. Marcellus, and P. E. Branton. 1999. Viruses and apoptosis. C. Dargemont, and S. Benichou. 2000. Cellular distribution and karyophilic prop- Annu. Rev. Microbiol. 53:577. erties of matrix, integrase, and Vpr proteins from the human and simian immu- 5. Johnstone, R. W., A. A. Ruefli, and S. W. Lowe. 2002. Apoptosis: a link between nodeficiency viruses. Exp. Cell Res. 260:387. cancer genetics and chemotherapy. Cell 108:153. 29. Aillet, F., H. Masutani, C. Elbim, H. Raoul, L. Chene, M. T. Nugeyre, C. Paya, 6. Ashkenazi, A., and V. M. Dixit. 1998. Death receptors: signaling and modulation. F. Barre-Sinoussi, M. A. Gougerot-Pocidalo, and N. Israel. 1998. Human immu- Science 281:1305. nodeficiency virus induces a dual regulation of Bcl-2, resulting in persistent in- 7. Locksley, R. M., N. Killeen, and M. J. Lenardo. 2001. The TNF and TNF re- fection of CD4ϩ T- or monocytic cell lines. J. Virol. 72:9698. ceptor superfamilies: integrating mammalian biology. Cell 104:487. 30. Belka, C., J. Rudner, S. Wesselborg, A. Stepczynska, P. Marini, 8. Leist, M., and M. Jaattela. 2001. Four deaths and a funeral: from caspases to A. Lepple-Wienhues, H. Faltin, M. Bamberg, W. Budach, and alternative mechanisms. Nat. Rev. Mol. Cell Biol. 2:589. K. Schulze-Osthoff. 2000. Differential role of caspase-8 and BID activation dur- 9. Wang, X. 2001. The expanding role of mitochondria in apoptosis. Genes Dev. ing radiation- and CD95-induced apoptosis. Oncogene 19:1181. 15:2922. 31. Strasser, A., A. W. Harris, D. C. Huang, P. H. Krammer, and S. Cory. 1995. Bcl-2 10. Van Loo, G., X. Saelens, M. Van Gurp, M. MacFarlane, S. J. Martin, and and Fas/APO-1 regulate distinct pathways to lymphocyte apoptosis. EMBO J. P. Vandenabeele. 2002. The role of mitochondrial factors in apoptosis: a Russian 14:6136. roulette with more than one bullet. Cell Death Differ. 9:1031. 32. Arnoult, D., I. Tatischeff, J. Estaquier, M. Girard, F. Sureau, J. P. Tissier, 11. Sugita, K., M. J. Robertson, Y. Torimoto, J. Ritz, S. F. Schlossman, and A. Grodet, M. Dellinger, F. Traincard, A. Kahn, et al. 2001. On the evolutionary C. Morimoto. 1992. Participation of the CD27 antigen in the regulation of IL- conservation of the cell death pathway: mitochondrial release of an apoptosis- 2-activated human natural killer cells. J. Immunol. 149:1199. inducing factor during Dictyostelium discoideum cell death. Mol. Biol. Cell 12. Sugita, K., T. Hirose, D. M. Rothstein, C. Donahue, S. F. Schlossman, and 12:3016. C. Morimoto. 1992. CD27, a member of the nerve growth factor receptor family, ϩ 33. Zhou, Q., S. Snipas, K. Orth, M. Muzio, V. M. Dixit, and G. S. Salvesen. 1997. is preferentially expressed on CD45RA CD4 T cell clones and involved in Target protease specificity of the viral serpin CrmA: analysis of five caspases. distinct immunoregulatory functions. J. Immunol. 149:3208. J. Biol. Chem. 272:7797. 13. Prasad, K. V., Z. Ao, Y. Yoon, M. X. Wu, M. Rizk, S. Jacquot, and S. F. Schlossman. 1997. CD27, a member of the tumor necrosis factor receptor 34. Adams, J. M., and S. Cory. 2001. Life-or-death decisions by the Bcl-2 protein family, induces apoptosis and binds to Siva, a proapoptotic protein. Proc. Natl. family. Trends Biochem. Sci. 26:61. Downloaded from Acad. Sci. USA 94:6346. 35. Scaffidi, C., S. Fulda, A. Srinivasan, C. Friesen, F. Li, K. J. Tomaselli, 14. Hintzen, R. Q., S. M. Lens, K. Lammers, H. Kuiper, M. P. Beckmann, and K. M. Debatin, P. H. Krammer, and M. E. Peter. 1998. Two CD95 (APO-1/Fas) R. A. van Lier. 1995. Engagement of CD27 with its ligand CD70 provides a signaling pathways. EMBO J. 17:1675. second signal for T cell activation. J. Immunol. 154:2612. 36. Dundr, M., and T. Misteli. 2001. Functional architecture in the cell nucleus. 15. Kobata, T., K. Agematsu, J. Kameoka, S. F. Schlossman, and C. Morimoto. 1994. Biochem. J. 356:297. CD27 is a signal-transducing molecule involved in CD45RAϩ naive T cell co- 37. Ko, Y. G., Y. S. Kang, H. Park, W. Seol, J. Kim, T. Kim, H. S. Park, E. J. Choi, stimulation. J. Immunol. 153:5422. and S. Kim. 2001. Apoptosis signal-regulating kinase 1 controls the proapoptotic

16. Goodwin, R. G., M. R. Alderson, C. A. Smith, R. J. Armitage, T. VandenBos, function of death-associated protein (Daxx) in the cytoplasm. J. Biol. Chem. http://www.jimmunol.org/ R. Jerzy, T. W. Tough, M. A. Schoenborn, T. Davis-Smith, K. Hennen, et al. 276:39103. 1993. Molecular and biological characterization of a ligand for CD27 deﬁnes a 38. Zheng, L., O. Schickling, M. E. Peter, and M. J. Lenardo. 2001. The death new family of cytokines with homology to tumor necrosis factor. Cell 73:447. effector domain-associated factor plays distinct regulatory roles in the nucleus 17. Kobata, T., S. Jacquot, S. Kozlowski, K. Agematsu, S. F. Schlossman, and and cytoplasm. J. Biol. Chem. 276:31945. C. Morimoto. 1995. CD27-CD70 interactions regulate B-cell activation by T 39. Stegh, A. H., O. Schickling, A. Ehret, C. Scafﬁdi, C. Peterhansel, T. G. Hofmann, cells. Proc. Natl. Acad. Sci. USA 92:11249. I. Grummt, P. H. Krammer, and M. E. Peter. 1998. DEDD, a novel death effector 18. Hendriks, J., L. A. Gravestein, K. Tesselaar, R. A. van Lier, T. N. Schumacher, domain-containing protein, targeted to the nucleolus. EMBO J. 17:5974. and J. Borst. 2000. CD27 is required for generation and long-term maintenance 40. Schickling, O., A. H. Stegh, J. Byrd, and M. E. Peter. 2001. Nuclear localization of T cell immunity. Nat. Immun. 1:433. of DEDD leads to caspase-6 activation through its death effector domain and 19. Xue, L., F. Chu, Y. Cheng, X. Sun, A. Borthakur, M. Ramarao, P. Pandey, inhibition of RNA polymerase I dependent transcription. Cell Death Differ. M. Wu, S. F. Schlossman, and K. V. Prasad. 2002. Siva-1 binds to and inhibits 8:1157.

BCL-xL-mediated protection against UV radiation-induced apoptosis. Proc. Natl. 41. Roth, W., F. Stenner-Liewen, K. Pawlowski, A. Godzik, and J. C. Reed. 2002. by guest on September 29, 2021 Acad. Sci. USA 99:6925. Identification and characterization of DEDD2, a death effector domain-containing 20. Yoon, Y., Z. Ao, Y. Cheng, S. F. Schlossman, and K. V. Prasad. 1999. Murine protein. J. Biol. Chem. 277:7501. Siva-1 and Siva-2, alternate splice forms of the mouse Siva gene, both bind to 42. Morgan, M., J. Thorburn, P. P. Pandolfi, and A. Thorburn. 2002. Nuclear and CD27 but differentially transduce apoptosis. Oncogene 18:7174. cytoplasmic shuttling of TRADD induces apoptosis via different mechanisms. 21. Spinicelli, S., G. Nocentini, S. Ronchetti, L. T. Krausz, R. Bianchini, and J. Cell Biol. 157:975. C. Riccardi. 2002. GITR interacts with the pro-apoptotic protein Siva and induces 43. Akiba, H., H. Nakano, S. Nishinaka, M. Shindo, T. Kobata, M. Atsuta, apoptosis. Cell Death Differ. 9:1382. C. Morimoto, C. F. Ware, N. L. Malinin, D. Wallach, et al. 1998. CD27, a 22. Nocentini, G., L. Giunchi, S. Ronchetti, L. T. Krausz, A. Bartoli, R. Moraca, member of the tumor necrosis factor receptor superfamily, activates NF-␬B and G. Migliorati, and C. Riccardi. 1997. A new member of the tumor necrosis factor/ stress-activated protein kinase/c-Jun N-terminal kinase via TRAF2, TRAF5, and nerve growth factor receptor family inhibits T cell receptor-induced apoptosis. NF-␬B-inducing kinase. J. Biol. Chem. 273:13353. Proc. Natl. Acad. Sci. USA 94:6216. 44. Riccardi, C., M. G. Cifone, and G. Migliorati. 1999. Glucocorticoid hormone- 23. Padanilam, B. J., A. J. Lewington, and M. R. Hammerman. 1998. Expression of induced modulation of gene expression and regulation of T-cell death: role of CD27 and ischemia/reperfusion-induced expression of its ligand Siva in rat kid- GITR and GILZ, two dexamethasone-induced genes. Cell Death Differ. 6:1182. neys. Kidney Int. 54:1967. 24. Henke, A., H. Launhardt, K. Klement, A. Stelzner, R. Zell, and T. Munder. 2000. 45. Bradley, J. R., and J. S. Pober. 2001. Tumor necrosis factor receptor-associated Apoptosis in Coxsackievirus B3-caused diseases: interaction between the capsid factors (TRAFs). Oncogene 20:6482. protein VP2 and the proapoptotic protein Siva. J. Virol. 74:4284. 46. Weber, C. H., and C. Vincenz. 2001. The death domain superfamily: a tale of two 25. Qin, L. F., T. K. Lee, and I. O. Ng. 2002. Gene expression profiling by cDNA interfaces? Trends Biochem. Sci. 26:475. array in human hepatoma cell line in response to cisplatin treatment. Life Sci. 47. Hsu, H., J. Xiong, and D. V. Goeddel. 1995. The TNF receptor 1-associated 70:1677. protein TRADD signals cell death and NF-␬B activation. Cell 81:495. 26. Xiao, H., V. Palhan, Y. Yang, and R. G. Roeder. 2000. TIP30 has an intrinsic 48. Park, A., and V. R. Baichwal. 1996. Systematic mutational analysis of the death kinase activity required for up-regulation of a subset of apoptotic genes. EMBO domain of the tumor necrosis factor receptor 1-associated protein TRADD. J. 19:956. J. Biol. Chem. 271:9858. 27. Cao, C., X. Ren, S. Kharbanda, A. Koleske, K. V. Prasad, and D. Kufe. 2001. The 49. Chinnaiyan, A. M., K. O’Rourke, M. Tewari, and V. M. Dixit. 1995. FADD, a ARG tyrosine kinase interacts with Siva-1 in the apoptotic response to oxidative novel death domain-containing protein, interacts with the death domain of Fas stress. J. Biol. Chem. 276:11465. and initiates apoptosis. Cell 81:505.