The dependence DCC (deleted in colorectal cancer) defines an alternative mechanism for activation

Christelle Forcet*, Xin Ye†‡, Laure Granger*, Ve´ ronique Corset*, Hwain Shin†‡, Dale E. Bredesen†‡§, and Patrick Mehlen*§

*Apoptosis͞Differentiation Laboratory, Molecular and Cellular Genetic Center, Centre National de la Recherche Scientifique Unite´Mixte de Recherche 5534, University of Lyon, 69622 Villeurbanne, France; †Buck Institute for Age Research, Novato, CA 94945; and ‡Neuroscience Department, University of California, San Diego, CA 92093

Edited by H. Robert Horvitz, Massachusetts Institute of Technology, Cambridge, MA, and approved December 28, 2000 (received for review August 9, 2000) The expression of DCC (deleted in colorectal cancer) is often The link between the putative role of DCC as a tumor markedly reduced in colorectal and other cancers. However, the suppressor and the ability of DCC to bind netrin-1 and to rarity of point identified in DCC coding sequences and mediate axon guidance was, however, not at all clear. Recently, the lack of a tumor predisposition phenotype in DCC hemizygous we have shown that DCC is a dependence receptor (13) and mice have raised questions about its role as a tumor suppressor. therefore functionally related to other dependence receptors DCC also mediates axon guidance and functions as a dependence such as p75NTR, the common receptor, the andro- receptor; such receptors create cellular states of dependence on gen receptor, and RET (14–17). Such receptors create cellular their respective ligands by inducing when unoccupied by states of dependence on their respective ligands by inducing ligand. We now show that DCC drives cell death independently apoptosis when unoccupied by ligand but inhibiting apoptosis in of both the mitochondria-dependent pathway and the death the presence of ligand (13–16). Hence, we have shown that the receptor͞caspase-8 pathway. Moreover, we demonstrate that DCC expression of DCC induces apoptosis in the absence of netrin-1, interacts with both caspase-3 and caspase-9 and drives the acti- but in the presence of netrin-1, DCC is antiapoptotic. Further- vation of caspase-3 through caspase-9 without a requirement for more, DCC was demonstrated to be a caspase substrate, with the cytochrome c or Apaf-1. Hence, DCC defines an additional pathway major site of cleavage at Asp-1290. The caspase cleavage of DCC for the apoptosome-independent caspase activation. was shown to be required for DCC to exert its proapoptotic effect, just as it has been shown for the and ogelstein and his colleagues (1) have shown that the devel- RET (13, 16, 17). Functionally, therefore, DCC serves as a Vopment of colonic carcinoma from normal colonic epithe- caspase amplifier in the absence of ligand via exposure of a lium is associated with the of a specific set of genes. proapoptotic domain lying in the amino-terminal region of the Allelic deletions (loss of heterozygosity) on chromosome 18q in intracellular domain, proximal to the cleavage site. more than 70% of primary colorectal tumors prompted the We investigated the mechanism by which DCC induces cell search for a tumor suppressor gene at that locus. This search led death when cleaved by . We demonstrate that DCC to the cloning of a putative cell-surface receptor, DCC (deleted induces apoptosis in a caspase-9-dependent pathway, yet by a in colorectal cancer) (1). DCC expression was then shown to be mechanism that is independent of the intrinsic (mitochondria- markedly reduced in more than 50% of colorectal tumors. dependent) apoptotic pathway. We also show that DCC recruits Moreover, the loss of DCC expression is not restricted to colon caspase-3 and caspase-9, resulting in the activation of caspase-3 carcinoma but has been observed in other tumor types, including via caspase-9. Hence, DCC defines an additional pathway for the carcinoma of the stomach, pancreas, esophagus, prostate, blad- apoptosome-independent caspase activation. der, breast, male germ cell tumors, neuroblastomas, gliomas, and Methods some leukemias (2, 3). However, proof that DCC is a tumor suppressor gene remains inconclusive (4, 5). Cells, Transfection Procedures, and Constructs. The neuroblastoma cell line IMR32 constitutively expressing DCC was from ECA cell DCC encodes an approximately 200-kDa type I membrane Ϫ͞Ϫ of 1,447 amino acids, which displays homology in its lines. The Apaf1 SAK-2 cells were obtained from P. Gruss, extracellular domain with cell adhesion molecules (2), suggesting Max Planck Institute for Biophysical Chemistry, Go¨ttingen, Ger- that DCC may play a role in cell–cell or cell–matrix interactions (6). many, and transient transfection was performed by using Lipo- fectamineϩ (Life Technologies, Grand Island, NY) according to However, DCC-mediated cell aggregation has not been firmly Ϫ͞Ϫ established (7). Recently, Tessier-Lavigne and collaborators (8, 9) the manufacturer’s instructions. The caspase-9 cells were have suggested that DCC may function as a component of a obtained from R. Flavell, Yale Univ. School of Medicine, New Haven, CT, and transient transfection was performed by using receptor complex that mediates the effects of the axonal chemoat- ϩ tractant netrin-1. The role of DCC in mediating growth cone Lipofectamine (Life Technologies). The 293-EBNA-netrin-1 extension has been supported by the analysis of the DCC knockout producing netrin-1 were obtained from M. Tessier-Lavigne, Univ. mice, which display abnormal brain development (4). However, the of California San Francisco. Transient transfection of colorectal cell signaling transduction of netrin-1 through DCC that results in axon line REGb or of human embryonic kidney 293T cells was per- outgrowth is mainly unknown. In response to netrin-1 binding, formed as described (13) by using the following plasmids. pDCC- DCC has been shown to interact with other netrin-1 receptors like

UNC5H (i.e., three members UNC5H1, -2, and -3) (10) or the This paper was submitted directly (Track II) to the PNAS office. adenosine A2b receptor shown to transduce cAMP production Abbreviation: TUNEL, terminal deoxynucleotidyltransferase-mediated dUTP end labeling. upon netrin-1 binding (11). Recently, it also has been proposed that §To whom reprint requests should be addressed. E-mail: [email protected] or Frazzled, the Drosophila ortholog of DCC, is not, in certain [email protected]. circumstances, a transducing receptor but rather a carrier for the The publication costs of this article were defrayed in part by page charge payment. This cue netrin-1 that allows netrin-1 distribution in specific regions of article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. the nervous system (12). §1734 solely to indicate this fact.

3416–3421 ͉ PNAS ͉ March 13, 2001 ͉ vol. 98 ͉ no. 6 www.pnas.org͞cgi͞doi͞10.1073͞pnas.051378298 Downloaded by guest on October 1, 2021 CMV-S, pDCC-CMV.D1290N, and pGNET1myc have been de- 100 ␮l of lysate, and protein-A Sepharose was used to pull them scribed (13). pBabe-p35 was a gift of P. Friesen, Univ. of Wisconsin, out. A similar immunodepletion of caspase-9 was performed by Madison. pcDNA-Crma was a gift of David Pickup, Duke Univ. using caspase-9 antibody (PharMingen). Medical Center, Durham, NC. pcDNA-Bcl2 was obtained by cloning the 1,855-bp cDNA fragment of Bcl-2 (18) into pcDNA-3.1. Results pcDNA-casp3, pcDNA-casp6, pcDNA-casp7, pcDNA-casp8, and DCC-Induced Apoptosis via Caspase-9. As DCC cleavage by caspases pcDNA-casp9 were gifts of G. S. Salvesen, Burnham Institute, La has been shown to be required for DCC proapoptotic activity Jolla, CA. Dominant negative caspase-encoding plasmids were (13), we first assessed whether caspase inhibition is sufficient to obtained by mutating the active site cysteines to alanine via a block DCC-induced cell death. Baculovirus protein p35, a potent Quikchange strategy (Stratagene). pcDNA-DNcasp3, pcDNA- and general caspase inhibitor (22), was expressed in the presence DNcasp6, pcDNA-DNcasp7, pcDNA-DNcasp8, and pcDNA- of DCC in human embryonic kidney 293T cells, and cell death DNcasp9 encode, respectively, the dominant negative mutants for was measured by trypan blue staining. Fig. 1A shows the caspase-3, -6, -7, -8, and -9. complete suppression of DCC-induced cell death in the presence of p35, confirming the role of caspases in DCC-induced apo- Immunoblotting and Immunoprecipitation. One-dimensional immu- ptosis. A similar block of cell death was observed when the noblots by using antibodies raised against DCC (Oncogene caspase inhibitor zVAD-fmk was used (Fig. 1C). Hence, caspase Research Products, Cambridge, MA), Flag M2 epitope (Sigma), activation is required for DCC-induced cell death. and caspase-3 and caspase-9 (PharMingen) were performed as To specify what caspase was involved, we either coexpressed described (19). Coimmunoprecipitations were performed on DCC with dominant negative (catalytic) mutants of caspases or cotransfected 293T or IMR32 lysed in 50 mM Hepes, pH 7.6͞125 expressed DCC in the presence of caspase inhibitors. Coexpression mM NaCl͞5 mM EDTA͞0.1% Nonidet P-40, by using a of the dominant negative mutant of caspase-8 failed to block caspase-3, caspase-9, or Flag M2 antibody and protein-A Sepha- DCC-induced apoptosis (Fig. 1B). Similarly, the addition of the rose (Sigma). DCC interaction with caspases was monitored by caspase-8 inhibitor IETD-fmk showed no effect on cell death immunoblot by using anti-DCC antibody. induction by DCC (Fig. 1C). The caspase-8 inhibitor CrmA was also unable to block DCC-induced cell death (not shown). Hence, Cell Death Analysis and Caspase Activity Measurement. Cell death caspase-8 is not required for DCC-induced cell death. The coex- was analyzed by using trypan blue as described in Rabizadeh et al. pression of the dominant negative mutants of caspase-6 and -7 also (20). Briefly, 48 h after transfection, all of the cells contained in a failed to block DCC-induced cell death (not shown), suggesting that 35-mm dish were washed and resuspended in DMEM. The per- caspase-6 and -7 are not required either. The expression of the centage of cell death was then obtained by enumerating the trypan dominant negative mutant of caspase-3 also had no influence on blue positive cells versus the total number of cells. Caspase-8 and cell death induction by DCC (Fig. 1B), but the treatment with the caspase-3 activity were measured by using the ApoAlert caspase-8, caspase-3 inhibitor DEVD-fmk blocked DCC-induced cell death ApoAlert caspase-3 assays from CLONTECH and the caspase-3 (Fig. 1C). This observation suggested that either caspase-3 is activity assay from Boehringer Mannheim. These assays use the required but can be replaced in certain circumstances by another Ac-IETD-AFC and Ac-DEVD-AFC substrates, respectively. These caspase-3-like caspase or that the proapoptotic activity of DCC activities were determined according to the manufacturer’s instruc- required a caspase-3-like caspase displaying an affinity for the tions, and caspase activation is presented as the ratio between the DEVD motif. caspase activity of the sample and that measured in 293T cells Of interest, coexpression of the dominant negative mutant of transfected with pCMV. Terminal deoxynucleotidyltransferase- caspase-9 completely blocked cell death induced by DCC (Fig. mediated UTP end-labeling (TUNEL) assays also were performed 1B). An inhibition of DCC-induced cell death by the dominant by using the ApoAlert DNA fragmentation assay (CLONTECH) negative mutant of caspase-9 also was observed in the colorectal according to manufacturer’s instructions on either adherent SAK-2 REGb cell line (not shown). Similarly, the addition of the or cytospun 293T and IMR32 cells. TUNEL reactivity was exam- caspase-9 inhibitor LEHD-fmk suppressed the cell death induc- ined and photographed with a Zeiss Axioscope photomicroscope. tion by DCC (Fig. 1C). To determine whether DCC-induced cell The number of TUNEL-positive cells was estimated by counting at death requires caspase-9 per se or another caspase that could be least 100 TUNEL-positive cells from six separate fields. inhibited by the dominant negative mutant caspase-9 and by the LEHD-fmk inhibitor, we analyzed the DCC effect in caspase-9 Two-Hybrid Analysis. The two-hybrid system Matchmaker II, knockout cells. Remarkably, whereas, as a positive control, a developed by CLONTECH, was used according to the manu- caspase-8 transfection induced cell death, DCC transfection was facturer’s instructions. As bait, the coding sequence of DCC-IC unable to drive caspase-9 knockout cell death (Fig. 1D). (1121–1447) from pSVK3-HA-DCC-IC (13) was inserted in Finally, to investigate whether in cells endogenously expressing pAS21 through NdeI–SalI insertion. The human brain library DCC, netrin-1 withdrawal led to a caspase-9-dependent cell death, was also from CLONTECH. Y190 yeast strain was cotrans- we first ‘‘conditioned’’ the neuroblastoma cell line IMR-32 endog- formed with the DCC-IC fused to the DNA binding domain of enously expressing DCC by allowing their growth in the presence GAL4 and the brain library fused to the activating domain of of netrin-1. After such ‘‘conditioning’’, netrin-1 withdrawal led to an GAL4. Cells were then allowed to grow in the absence of leucine, increased number of apoptotic cells measured either by trypan blue tryptophan, and histidine and in the presence of 55 mM 3-amino- staining (Fig. 1E) or TUNEL reactivity (Fig. 1F). Of interest, 1,2,4-triazole. Forty-five positive clones representing nine inhibition of caspase-9 with the LEHD-fmk inhibitor blocked cell unique cDNAs were isolated from 107 screened. death induced by netrin-1 withdrawal. Hence, apoptosis induced by DCC required caspase-9, and because the only known apoptotic In Vitro Translation, S100 293T Lysate, and Caspase Cleavage Reac- pathway by using caspase-9 involves apoptosome formation with tions. Plasmids pCR-DCC-IC, pCR-DCC-IC.D1290N, pCR- caspase-9͞Apaf-1͞cytochrome c͞(d)ATP, these data suggested the DCC-IC͞1121–1290, or pCR-DCC-IC͞1290–1447 described in involvement of this mitochondria-dependent apoptotic pathway in ref. 13 were transcribed by using T7 polymerase, then translated DCC-induced cell death (23). by using the TNT system (Promega) for2hat30°C. The S100 lysate was prepared as described by Li et al. (21) from 108 293T, DCC-Induced Apoptosis Is Independent of the Mitochondria-Depen- L929, or SAK-2 cells. To immunodeplete S100 lysate of cyto- dent Apoptotic Pathway. We therefore analyzed the ability of chrome c,1␮g of anti-cytochrome c (PharMingen) was added to DCC to induce cell death in the presence of inhibitors of this MEDICAL SCIENCES

Forcet et al. PNAS ͉ March 13, 2001 ͉ vol. 98 ͉ no. 6 ͉ 3417 Downloaded by guest on October 1, 2021 Fig. 2. DCC-induced cell death does not require a mitochondria-dependent apoptotic pathway. (A) 293T cells were transiently transfected as described in Fig. 1 with either the DCC expression plasmid pDCC-CMV-S (DCC) or pCMV control plasmid (cont.) in the presence or absence of the Bcl-2 expression construct pcDNA-Bcl2. Cell death was then monitored 48 h after transfection as described in Fig. 1. Two tested quantities of DCC plasmid (0.5 and 2.5 ␮g͞60-mm dish) are shown; 0.5 ␮g͞60-mm dish is the lowest quantity of DCC-expressing plasmid tested that was able to induce a significant increase of cell death. Bcl-2 activity was controlled by measuring death of Bcl-2-transfected cell after addition of 5 ␮M staurosporine for 4 h. (B) Control or DCC-transfected cells were incubated the last 18 h with 5 ␮M cyclosporin A. Cell death was then monitored 48 h after trans- fection. As a control assay, pCMV-transfected cells were treated for 4 h with 5 ␮M staurosporine. (C) Apaf-1 Ϫ͞Ϫ SAK-2 cells were transiently transfected with either the pCMV control plasmid (cont.) or the DCC expression plasmid pDCC- CMV-S (DCC). 48 h after transfection, the mock-transfected cells were then treated (Stauro.) or not for 4 h with 2 ␮M staurosporine. Cell death was then Fig. 1. Cell death induction by DCC requires caspase-9. Human embryonic monitored at 48 h as described in Fig. 1. kidney 293T cells were transiently transfected as described (13) with the pCMV control plasmid (cont.), the DCC expression plasmid pDCC-CMV-S (DCC), or the DCC expression plasmid and the netrin-1-encoding plasmid pGNET1myc (DCC the proapoptotic activity of DCC. A similar effect was observed ϩ net.) in the presence of the p35 expression construct pBabe-p35 (A), or of the in colorectal REGb (not shown) and DLD cells (24). Bcl-2 has dominant negative caspase-3, -8, or -9 expression construct (B). Cell death was been shown to block the cell-death pathway involving cyto- analyzed by using trypan blue as described in ref. 13. (C) Mock (cont.) or DCC chrome c release and cytochrome c-dependent activation of (DCC) transfected 293T cells were incubated or not with 20 ␮M Zvad-fmk (Zvad), IETD-fmk (IETD), DEVD-fmk (DEVD), or LEHD-fmk (LEHD) 24 h after the caspase-9 (23, 25). Hence, we analyzed the ability of an inhibitor beginning of the transfection. Cell death was then analyzed by using trypan of cytochrome c release to modulate DCC proapoptotic activity. blue (as in A and B) 24 h later. (D) Caspase-9 Ϫ͞Ϫ cells were transfected with DCC-transfected cells then were incubated in the presence of a mock plasmid (cont.), DCC-expressing plasmid (DCC), or the caspase-8- cyclosporin A, an inhibitor of the mitochondrial membrane expressing plasmid pcDNA-casp.8 (casp.8). (E and F) IMR32 cells were cultured permeability transition (26). As shown in Fig. 2B, although for 48 h in the presence of culture medium issued of 293-EBNA-netrin-1 cells cyclosporin A blocked staurosporine-induced cell death, it had producing extracellular netrin-1. After this conditioned growth, cells were either further cultured with netrin-1-containing medium (ϩ net.) or with no effect on DCC-induced apoptosis. Moreover, comparing medium devoid of netrin-1 (Ϫ net.). In the latter case, cells were incubated or control and DCC-transfected cells, no DCC-induced specific not with 20 ␮M LEHD-fmk (ϩLEHD). Cell death was then measured either by release of cytochrome c was observed, at least during the 36 h trypan blue staining at the indicated time following netrin-1 withdrawal (E)or following transfection (not shown). To further analyze whether by TUNEL reactivity after 24 h of netrin-1 withdrawal (F). the apoptosome was actually required for cell death induction by DCC, we studied the effect of DCC expression in Apaf1 Ϫ͞Ϫ knockout cells. As shown in Fig. 2C, in Apaf1 Ϫ͞Ϫ cells, mitochondria-dependent apoptotic pathway. 293T cells were although a staurosporine treatment had no effect on cell death, transfected with DCC in the presence or absence of Bcl-2 (18). DCC expression induced an increased trypan blue staining (Fig. Fig. 2A shows that although Bcl-2 expression was sufficient to 2C) and TUNEL reactivity (not shown). Hence, these data block staurosporine-induced cell death, Bcl-2 had no effect on strongly suggest that, although DCC required caspase-9, it did

3418 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.051378298 Forcet et al. Downloaded by guest on October 1, 2021 not require the mitochondria-dependent apoptotic pathway that involves cytochrome c release and apoptosome formation, which is the only known pathway for caspase-9 activation.

DCC Interacts with Caspase-3 and Caspase-9. A search for DCC- interacting that may explain this apparent paradox, by using two-hybrid screening of a human brain library with the DCC intracellular domain (DCC-IC) as bait, revealed the interaction of DCC-IC with a polypeptide whose sequence corresponds to that of caspase-3 (Fig. 3A). A coimmunoprecipitation study confirmed the interaction of caspase-3 with DCC (Fig. 3B). However, the caspase- 3–DCC interaction was detected only when DCC was expressed in the presence of a ligand or when the caspase cleavage site of DCC (Asp-1290) was mutated, both of which are conditions in which DCC does not induce cell death (13). As a control experiment, when a similar coimmunoprecipitation was performed with caspase-6, no interaction was detected (Fig. 3C). In a search of other caspases interacting with DCC, we observed an interaction of DCC with caspase-9 (Fig. 3C and ref. 13). However, the interaction was visualized predominantly when DCC was in its proapoptotic state; only weak binding was detected when DCC was expressed with netrin-1 or when DCC D1290N was used. Caspase-9–DCC inter- action was monitored either from a pull-down with caspase-9 (Fig. 3C) or from a pull-down with DCC (not shown). Moreover, interaction of DCC with caspase-9 was visualized in the neuroblas- toma cell line IMR-32 endogenously expressing DCC and caspase-9 (Fig. 3D). In a search of the interacting domain of DCC, the proapo- ptotic region of DCC (amino acids 1121–1290, ref. 13) lying upstream of its main caspase cleavage site was found to be crucial for caspase-9 interaction, whereas the region distal to the caspase site was required for caspase-3 interaction (Fig. 3E). Such an interaction of DCC with caspase-3 and -9 is reminiscent of the apoptosome complex Apaf-1, caspase-9, caspase-3, cytochrome c, and (d)ATP (21).

In Vitro Activation of Caspase-3 by DCC. Because the caspase-9 apoptosome requires cytochrome c to activate caspase-9, which then activates caspase-3 (21), whereas DCC induced apoptosis independently of cytochrome c release, we established an in vitro assay to determine whether the DCC complex could form a direct activator of caspase-3 in the absence of cytochrome c. The com- bination of in vitro-translated DCC-IC, caspase-9, and caspase-3 Fig. 3. Interaction of caspase-3 and caspase-9 with the intracytoplasmic domain was not sufficient to induce caspase-3 activation (not shown). of DCC. (A) Two-hybrid screening of DCC-IC revealed interaction with caspase-3. However, the addition of an S100 cell lysate depleted of cytochrome Y190 yeast transformed with a GAL4 DNA binding domain-DCC-IC fusion (DCC) were transformed with either a mock plasmid containing the GAL4 tran- c showed a significant activation of caspase-3 by DCC-IC monitored scriptional activation domain AD (DCC ϩ vector) or a similar construct fused to either by the cleavage of purified pro-caspase-3 (Fig. 4A)orof a brain cDNA library. Cells were then allowed to grow in the absence of leucine, Ac-DEVD-AFC (Fig. 4B). Dose-dependence analysis revealed that tryptophan, and histidine and in the presence of 55 mM 3-amino-1,2,4-triazole. DCC-IC and cytochrome c display different kinetics for caspase-3 The clone corresponding to caspase-3 is shown (DCC ϩ caspase-3). (B) Coimmu- activation (Fig. 5). Indeed, the minimum required concentration of noprecipitations were performed on 293T cells cotransfected with pcDNA- cytochrome c for pro-caspase-3 activation was found to be 3.5 times DNcasp3 (casp.3) or pcDNA-3 (cont.) and pDCC-CMV-S, pDCC-CMV.D1290N, or less than the one for DCC-IC. Moreover, the addition of exogenous pDCC-CMV-S ϩ pGNET1myc, the netrin-1-encoding plasmid. To obtain a similar dATP further increased cytochrome c-induced caspase-3 activation expression of the protein DCC in the different sample, half of the plasmid at a low concentration of cytochrome c, hence suggesting that, at pDCC-CMV-S was used when cotransfected with the netrin-1-expressing vector because the presence of DCC ligand leads to increased stability of DCC (13). low concentration, cytochrome c is a better activator than DCC-IC Anti-caspase-3 antibody was used to immunoprecipitate caspase-3, and binding (Fig. 5). However, the limiting factors present in the cell lysate were was revealed by Western blot by using anti-DCC antibody. (C) Same as in B except reached more rapidly with cytochrome c, as the maximal caspase-3 that pcDNA-DNcasp9 and pcDNA-DNcasp6 were used. In this case, anti-Flag activation was observed with DCC-IC. Hence, in a physiological antibody was used to immunoprecipitate DNcaspase-9 and DNcaspase-6. (D) cellular context, DCC might conceivably be a good activator of Immunoprecipitation of DCC by caspase-9 was performed by using IMR-32 cells. caspase activation. DCC was able to mediate caspase-3 activation in The pull-down was performed with or without anti-caspase-9 antibody. (E) the absence of cytochrome c and in an S100 lysate of Apaf-1 Ϫ͞Ϫ Coimmunoprecipitations were performed, respectively, as in B and C on 293T cells (Fig. 6A), strongly suggesting that DCC-induced caspase-3 cotransfected with pcDNA-DNcasp3 or pcDNA-DNcasp9 in the presence of either activation is apoptosome-independent. Moreover, DCC-induced pDCC-CMV-⌬1121–1290 or pDCC-CMV-⌬1290–1447. tot., immunoblots per- formed on the lysate before addition of the antibodies allowing the pull out; IP, caspase-3 activation was not accompanied with caspase-9 cleavage immunoblots by using an anti-DCC antibody on samples derived from the pull- that is very often associated with caspase-9 activation (not shown). downs. In B–E, dominant negative mutants of caspases were chosen instead of However, caspase-9 was required, as immunodepletion of caspase-9 tagged wild-type because overexpression of wild-type caspase drove massive in the S100 cell lysate led to an inhibition of the DCC-induced MEDICAL SCIENCES 293T cell death. caspase-3 activation (Fig. 6B).

Forcet et al. PNAS ͉ March 13, 2001 ͉ vol. 98 ͉ no. 6 ͉ 3419 Downloaded by guest on October 1, 2021 Fig. 4. DCC-IC activation of caspase-3 in vitro. Cytochrome c-immunodepleted S100 extract from 293T or L929 cells (19) was incubated with 0.2 ␮M of purified pro-caspase-3 in the presence of an in vitro translation of DCC-IC (DCC) or of an empty vector (cont.) for 1 h (293T) or 24 h (L929) at 30°C. As a positive control, 1 ␮M cytochrome C was added. (A) Immuno- blots developed with anti-caspase-3 (ac- tive and pro-form of caspase-3) or anti- active caspase-3 antibodies (PharMingen) are shown. (B) Caspase-3 activity was also monitored as described in Methods by using the ApoAlert caspase-3 activity assay.

Thus, DCC interacts with caspases-3 and -9 and mediates the apoptosome-associated caspase-9 activation. DCC-induced cell caspase-9-dependent activation of caspase-3 in a cytochrome c- death may thus be related to the ability of the intracellular domain independent manner. Therefore, DCC participates in an alternative of DCC to (i) recruit pro-caspase-9 and pro-caspase-3 and (ii) pathway of caspase activation in which cytochrome c is not required mediate caspase-9-dependent caspase-3 activation in a cell-free but DCC cleavage is. Indeed, when a similar in vitro assay was extract. Moreover, this mechanism of activation of caspase-3, which performed by using DCC-IC D1290N, the intracellular domain of is not only independent of mitochondria, cytochrome c release, and DCC mutated in its main caspase site, no activation of caspase-3 was apoptosome formation but also independent of caspase-8 activa- observed (Fig. 6C). Hence, cleavage of DCC by caspases is a necessary step for further caspase-3 activation. Furthermore, the tion, represents a novel mechanism for the induction of apoptosis. As with the Fas-FADD(Mort-1)͞pro-caspase-8͞FLASH (27) and proapoptotic region of DCC that binds to caspase-9 (1121–1290) ͞ ͞ was sufficient to activate pro-caspase-3, whereas the C-terminal IC cytochrome c Apaf-1 pro-caspase-9 pathways (21), DCC induces region (1290–1447) had no comparable activity. Taken together, caspase activation and apoptosis through the recruitment of these results suggest that the caspase-activating complex lies in the caspases into an activating complex. Of interest, DCC sequence proapoptotic region of DCC, a region released or exposed via the does not show any death domain, CARD motif, or death effector caspase cleavage in D1290. domain, suggesting either an indirect interaction (i.e., DCC with caspase-9) via an adaptor molecule or a direct interaction (i.e., DCC Discussion with caspase-3) via an unknown motif of interaction. The obser- Here, we present evidence that DCC-induced cell death requires vation that the DCC, caspase-9, and caspase-3 combination is not caspase-9 but does not involve cytochrome c release and subsequent sufficient to activate in vitro caspase-3 but required the addition of a S100 lysate strongly suggests that other proteins present in the lysate are required to allow a functional DCC-associated caspase- activating complex. The nature of this complex, and whether it includes additional proteins similar to FADD or Apaf-1, remains to be determined. There exists an interesting difference between the DCC- associated caspase-activating complex and the apoptosome com- plex. Indeed, within the apoptosome complex, caspase-9 once recruited is cleaved and displays an increased activity (21), whereas caspase-9 recruitment into the DCC complex is appar-

Fig. 5. Comparison of caspase-3 activation by DCC-IC versus cytochrome c.A dose-dependence analysis of DCC-induced caspase-3 activation was per- formed. In a similar in vitro assay as in Fig. 4, different quantities of DCC-IC or cytochrome c in the presence or not of 1 mM dATP were added. In the case of DCC-IC, the concentration was estimated based on comparison with purified protein (L.G. and P.M., unpublished data). Purified cytochrome c was used as attempts to activate caspase-3 in vitro with in vitro-translated cytochrome c Fig. 6. DCC-IC activation of caspase-3 required caspase-9 and DCC cleavage. The were unsuccessful, probably as a result of the absence of the cytochrome same in vitro assay described in Fig. 4A was performed, but in this case, the S100 c-associated heme (C.F. and P.M., unpublished results). The p20 cleavage lysate was generated from Apaf-1 Ϫ͞Ϫ SAK-2 cells (A), the 293T S100 lysate was fragment of caspase-3 was then detected by immunoblot and quantified by immunodepleted in caspase-9 (B), or in vitro-translated DCC-IC D1290N, DCC-IC͞ densitometry. A graph of the level of caspase-3 activation is presented. 1121–1290, DCC-IC͞1291–1447 were used instead of DCC-IC (C).

3420 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.051378298 Forcet et al. Downloaded by guest on October 1, 2021 ently not accompanied by caspase-9 processing. However, caspase-activating complex. In both models, an apparent paradox caspase-9 zymogen has been described to have proteolytic would emerge because caspase cleavage releases the C-terminal activity similar to its cleaved form because the zymogenicity (i.e., domain of DCC shown here to bind pro-caspase-3. The caspase- ratio of the processed protease activity to the nonprocessed activating complex would then appear in the N-terminal region IC protease activity) of caspase-9 is approximately 10, whereas the of DCC when the pro-caspase-3 is no longer present to be activated. zymogenicity of caspase-3 is greater than 10,000 (28). Moreover, To reconcile this apparent paradox, two possible scenarios can be as described recently (29), the proteolytic processing of pro- proposed: (i) the cleavage of DCC brings together the pro- caspase-9 has only a minor effect on the enzyme’s catalytic caspase-3 interacting with the C-terminal IC domain and the activity, whereas the key requirement for caspase-9 activation is caspase-activating complex interacting with the N-terminal IC its association with a dedicated protein cofactor Apaf-1. Because domain of DCC and allows further caspase-3 activation, or (ii)the DCC still allowed caspase-3 activation in Apaf1 Ϫ͞Ϫ knockout DCC-associated caspase-activating complex based on the presence cell lysate, another caspase-9 cofactor may be present in the of caspase-9 activates other effector caspases (e.g., caspase-7 or DCC-associated caspase-activating complex. An alternative ex- caspase-3 not associated with DCC) and pro-caspase-3 interaction planation may be that caspase-9 is required for DCC-induced cell with DCC is independent of the caspase-activating complex. Thus, death and caspase activation not because it functions as a the pro-caspase-3 interaction with DCC observed in the presence protease but rather because it displays another unknown func- of netrin-1 then could represent a reservoir of unavailable tion (e.g., adaptor). Along this line, it has been recently shown caspase-3, contributing to the antiapoptotic activity of DCC bound that in certain circumstances, caspase-9 zymogen but not its to its ligand (13). LEHD-cleaving activity is needed for cell death induction (30). DCC and its associated caspase-activating complex may thus play It is tempting to speculate that the ligand withdrawal leads to the a critical role in the regulation of axon guidance during develop- formation of the DCC-associated caspase-activating complex, as ment of the nervous system (4, 8, 9) and in tumor suppression. ligand withdrawal leads to DCC pro-apoptotic activity. The nature Recently, Yee et al. (9) demonstrated that netrin-1 knock-out mice of the DCC change allowing caspase-3 activation is probably related show not only a loss in axonal guidance but also in neural pontine to a change of conformation of the receptor similar to the one ntr cell migration, the latter of which is related to an increased cell described in the case of p75 (15). This change could then affect death (31). This result provides strong support for the notion that DCC proapoptotic activity at two levels: (i) the caspase cleavage of DCC can function in vivo to induce cell death. We also observed DCC and (ii) the recruitment of caspase-9. Both of these events are that in rat embryonic dorsal spinal cord explants, the presence of necessary to allow further caspase-3 activation. In favor of the netrin-1 not only allows axonal projections but also blocks a second effect, we indeed showed that caspase-9 binds DCC pre- caspase-9-dependent death of the explant known to express DCC dominantly when expressed in the absence of netrin-1, although we (unpublished work), suggesting that axonal guidance and cell death cannot exclude the possibility that this preferential recruitment is a could represent two dependent mechanisms. Whether netrin-1- subsequent effect of the DCC caspase cleavage. The fact that DCC mediated inhibition of caspase activation has a role in directing axon cleavage is blocked by ligand presence is, however, mostly specu- projections in the nervous system, however, remains to be shown. lative, as we have been unable to detect the caspase cleavage of Finally, with respect to the potential function of DCC as a tumor full-length DCC (ref. 13, C.F. and P.M., unpublished results). Two models could thus be proposed. A first model would be that the suppressor, these results have suggested that at least one role of presence of netrin-1 blocks DCC caspase cleavage (e.g., via a closed DCC may be to induce apoptosis in expressing cells that grow conformation), which allows the recruitment of pro-caspase-3. or metastasize beyond the limitation of ligand availability, by using Ligand withdrawal would then lead to the caspase cleavage of DCC a mechanism that requires a DCC-associated caspase-activating and the subsequent exposure and͞or release of the pro-apoptotic complex. region of DCC. This would then lead to the recruitment of a caspase-activating complex. The second model would then postu- We thank G. S. Salvesen for the gift of caspase cDNAs, M. Tessier- Lavigne for the 293-EBNA netrin-1, P. Gruss for the Apaf-1 knockout late that DCC is, whether the netrin-1 is present or not, cleaved by cells, and R. Flavell for the caspase-9 knockout cells. L.G. and V.C. were active caspase, a preliminary but not sufficient event for further supported by Association pour la Recherche sur le Cancer Grant 9036. pro-apoptotic activity. Along this line, RET, another dependence This work was supported by the Ligue Contre le Cancer, the Fondation receptor, was shown to be cleaved both in the absence and in the pour la Recherche Me´dicale,Association pour la Recherche sur le presence of its ligand (17). The presence of the ligand would then Cancer Grants 9036 and 9340, the Emergence Program, and National inhibit caspase-9 recruitment and the further formation of a Institutes of Health Grants NS33376 and CA69381.

1. Fearon, E. R., Cho, K. R., Nigro, J. M., Kern, S. E., Simons, J. W., Ruppert, J. M., Hamilton, 16. Ellerby, L. M., Hackam, A. S., Propp, S. S., Ellerby, H. M., Rabizadeh, S., Cashman, N. R., Trifiro, S. R., Preisinger, A. C., Thomas, G., Kinzler, K. W., et al. (1990) Science 247, 49–56. M. A., Pinsky, L., Wellington, C. L., Salvesen, G. S., et al. (1999) J. Neurochem. 72, 185–195. 2. Cho, K. R. & Fearon, E. R. (1995) Curr. Opin. Genet. Dev. 5, 72–78. 17. Bordeaux, M. C., Forcet, C., Granger, L., Corset, V., Bidaud, C., Billaud, M., Bredesen, 3. Fearon, E. R. (1996) Biochim. Biophys. Acta 1288, M17–M23. D. E., Edery, P. & Mehlen, P. (2000) EMBO J. 19, 4056–4063. 4. Fazeli, A., Steen, R. G., Dickinson, S. L., Bautista, D., Dietrich, W. F., Bronson, R. T., Bresalier, 18. Hockenbery, D. M., Oltvai, Z. N., Yin, X. M., Milliman, C. L. & Korsmeyer, S. J. (1993) Cell R. S., Lander, E. S., Costa, J. & Weinberg, R. A. (1997) Nature (London) 386, 796–804. 75, 241–251. 5. Tanaka, K., Oshimura, M., Kikuchi, R., Seki, M., Hayashi, T. & Miyaki, M. (1991) Nature 19. Mehlen, P., Kretz-Remy, C., Preville, X. & Arrigo, A.-P. (1996) EMBO J. 15, 2695–2706. (London) 349, 340–342. 20. Rabizadeh, S. Ye, X., Sperandio, S., Wang, J. L., Ellerby, H. M., Ellerby, L. M., Giza, C., Andrusiak, 6. Hedrick, L., Cho, K. R., Fearon, E. R., Wu, T. C., Kinzler, K. W. & Vogelstein, B. (1994) R. L., Frankowski, H., Yaron, Y., Moayen, N. N., et al. (2000) J. Mol. Neurosci., 15, 215–230. Genes Dev. 8, 1174–1183. 21. Li, P., Nijhawan, D., Budihardjo, I., Srinivasula, S. M., Ahmad, M., Alnemri, E. S. & Wang, 7. Chuong, C. M., Jiang, T. X., Yin, E. & Widelitz, R. B. (1994) Dev. Biol. 164, 383–397. X. (1997) Cell 91, 479–489. 8. Keino-Masu, K., Masu, M., Hinck, L., Leonardo, E. D., Chan, S. S., Culotti, J. G. & 22. Fisher, A. J., Cruz, W. D., Zoog, S. J., Schneider, C. L. & Friesen, P. D. (1999) EMBO J. Tessier-Lavigne, M. (1996) Cell 87, 175–185. 18, 2031–2039. 9. Serafini, T., Colamarino, S. A., Leonardo, E. D., Wang, H., Beddington, R., Skarnes, W. C. 23. Green, D. R. & Reed, J. C. (1998) Science 281, 1309–1312. & Tessier-Lavigne, M. (1996) Cell 87, 1001–1014. 24. Chen, Y. Q., Hsieh, J. T., Yao, F., Fang, B., Pong, R. C., Cipriano, S. C. & Krepulat, F. (1999) 10. Hong, K., Hinck, L., Nishiyama, M., Poo, M.-M., Tessier-Lavigne, M. & Stein, E. (1999) Cell Oncogene 18, 2747–2754. 97, 927–941. 25. Yang, J., Liu, X., Bhalla, K., Kim, C. N., Ibrado, A. M., Cai, J., Peng, T. I., Jones, D. P. & 11. Corset, V. T., Nguyen-Ba-Charvet, K., Forcet, C., Moyse, E., Che´dotal, A. & Mehlen, P. Wang, X. (1997) Science 275, 1129–1132. (2000) Nature (London) 407, 747–750. 26. Susin, S. A., Lorenzo, H. K., Zamzami, N., Marzo, I., Snow, B. E., Brothers, G. M., Mangion, 12. Hiramoto, M., Hiromi, Y., Giniger, E. & Hotta, Y. (2000) Nature (London) 406, 886–889. J., Jacotot, E., Costantini, P., Loeffler, M., et al. (1999) Nature (London) 397, 441–445. 13. Mehlen, P., Rabizadeh, S., Snipas, S. J., Assa-Munt, N., Salvesen, G. S. & Bredesen, D. E. 27. Ashkenazi, A. & Dixit, V. M. (1998) Science 281, 1305–1308. (1998) Nature (London) 395, 801–804. 28. Stennicke, H. R. & Salvesen, G. S. (1999) Cell Death Differ. 6, 1054–1059. 14. Rabizadeh, S., Oh, J., Zhong, L. T., Yang, J., Bitler, C. M., Butcher, L. L. & Bredesen, D. E. 29. Hengartner, M. O. (2000) Nature (London) 407, 770–776. (1993) Science 261, 345–348. 30. Sperandio, S., de Belle, I. & Bredesen, D. E. (2000) Proc. Natl. Acad. Sci. USA 97, 14376–14381. 15. Bredesen, D. E., Ye, X., Tasinato, A., Sperandio, S., Wang, J. J., Assa-Munt, N. & 31. Yee, K. T., Simon, H. H., Tessier-Lavigne, M. & O’Leary, D. D. M. (1999) Neuron 24, MEDICAL SCIENCES Rabizadeh, S. (1998) Cell Death Differ. 5, 365–371. 607–622.

Forcet et al. PNAS ͉ March 13, 2001 ͉ vol. 98 ͉ no. 6 ͉ 3421 Downloaded by guest on October 1, 2021