Regulation of the Extrinsic Apoptotic Pathway by the Extracellular Matrix Glycoprotein EMILIN2

Cells and other materials.HT1080, SK-UT-1, HeLa, and RD-ES cell lines were obtained from the ATCC and cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (FBS) (Gibco BRL). 293-EBNA cells were a gift from Rupert Timpl (Max Planck, Munich, Germany) and were cultivated in the same medium with 250 μg/ml of G418. Normal human foreskin fibroblasts (BJ) and transformed BJ/ERM cells were cultivated in minimal essential medium supplemented with nonessential amino acids and 10% FBS. Wild-type A3 and caspase-8 mutant I 9.2 Jurkat cell lines were obtained from the ATCC and were cultured in RPMI supplemented with 10% FBS. Cells were maintained at 37°C in a humidified 5% CO₂ atmosphere. Anti-EMILIN2 monoclonal antibody 828B3B3 was obtained upon immunization of BALB/c mice with 100 μg of the recombinant gC1q domain of EMILIN2. The anti-cleaved PARP polyclonal antibody AB3620, which detects only the cleaved large 89-kDa subunit and not full-length PARP, the FLICE/caspase-8 fluorimetric protease assay kit, and ApopTag apoptosis detection kit were obtained from CHEMICON International (Temecula, CA). Anti-caspase-10 polyclonal antibody was obtained from MBL (Woburn, MA). Anti-caspase-3 and anti-TNFR1 were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). The anti-DR4 antibodies were obtained from Sigma-Aldrich (Milan, Italy) and Alexis (Florence, Italy). Monoclonal anti-DR5 antibody was obtained from Alexis. Anti-Myc antibody was obtained from Cell Signaling Technology Inc. (Danvers, MA). Secondary horseradish peroxidase-conjugated antibodies were obtained from Amersham (GE Healthcare), or Trueblot was obtained from eBiosciences (San Diego, CA). A secondary antibody conjugated with Alexa Fluor 680, Alexa Fluor dye-labeled cholera toxin subunit B, human epithelial growth factor (EGF), and insulin-like growth factor I (IGF-I) were obtained from Invitrogen (Milan, Italy). The general caspase inhibitor Z-VAD-FMK, anti-caspase-8 monoclonal antibody, and anti-CD95 monoclonal antibody were obtained from R&D Systems Inc. (McKinley Place, MN). Cell Death Detection ELISA^PLUS and In situ Cell Death Detection kit, fluorescein, were purchased from Roche Diagnostics (Milan, Italy). The human annexin V-fluorescein isothiocyanate kit was obtained from Bender MedSystems GmbH (Vienna, Austria). Staurosporine was obtained from Sigma-Aldrich (Milan, Italy). Reagents for RNA Interference siGenome SMARTpool targeting EMILIN2 (catalog number M-014777-00) and TRAIL (catalog number M-011524-01) and a nontargeting small interfering RNA (siRNA) pool (catalog number D-001206-13) as well as DHARMAfect 3 were obtained from Dharmacon RNA Technologies (Lafayette, CO).

DNA constructs.Human EMILIN2 cDNA was retrotranscribed from total human kidney RNA and cloned into the pCEP-Pu vector containing the sequence of the BM40 signal peptide. The following oligonucleotides were used: 5′-CTAGCTAGCAGGCCCGCAGCCCGGG-3′ (forward) and 5′-CGGGATCCTTAATGGTGATGGTGATGATGGAGGTGGGAAAGGAA-3′ (reverse). EMILIN2 was subcloned into pcDNA3.1/Myc-His and the retroviral vector pLPC by HindIII and BamHI restriction. The EMILIN2 C-terminal fragment, containing the sequence coding for the proline-rich region, the collagenous stalk, and the gC1q-like domain, was generated by NarI restriction. Wild-type DR4 (wtDR4) and a truncated DR4 mutant (tDR4) were reverse transcribed from total RNA extracted from the HeLa cell line and cloned into pcDNA 3.1/Myc-His. The following oligonucleotides were used: 5′-CCCAAGCTTATGGCGCCACCACCAGCT-3′ (forward) and 5′-TGGATATCTCTCCAAGGACACGGCAGAGC-3′ (reverse); the reverse oligonucleotide for tDR4 was 5′-GCTCTAGAGAGACCCAAGCGCCAGAA-3′. The forward oligonucleotide for TRAIL was 5′-CTAGCTAGCAGACTACAAGGACGACGATGACAAGACCTCTGAGGAAACCATTTC-3′, and the reverse oligonucleotide was 5′-CCCAAGCTTCAGGTCAGTTAGCCAACT-3′. The fragment was subsequently cloned into vector pCEP-Pu and expressed in 293-EBNA cells. The extracellular regions of DR4 and two splice variants of DR5 (DR5_S and DR5_L) were cloned into vector pGEX-KG with the following oligonucleotides: 5′-CGGGATCCGCGAGTGGGACAGAGGCA-3′ (forward) and 5′-GGAATTCATTATGTCCATTGCCTGA-3′ (reverse). The forward oligonucleotide for the DR5 constructs was 5′-CGGGATCCGAGTCTGCTCTGATCAC-3′; the reverse oligonucleotide for DR5_S was 5′-GGAATTCTGATTCTTTGTGGACACA-3′, and that for DR5_L was 5′-GGAATTCTGAGAGAGAACAGGGAGA-3′.

Cell transfection, expression, and purification of recombinant proteins.293-EBNA cells were transfected by electroporation with the pCEP-Pu constructs and selected in the presence of 0.5 μg/ml of puromycin and 250 μg/ml of G418. After 24 to 48 h of incubation in serum-free medium, conditioned media were collected. The various recombinant protein concentrations used in the tests were estimated to be ∼20 to 30 nM by Western blotting using purified proteins of known concentrations as a reference. EMILIN2 was purified with Ni-nitrilotriacetic acid (NTA) resin (QIAGEN, Milan, Italy) and eluted with 250 mM imidazole. TRAIL containing a FLAG tag at the N terminus was purified by using anti-FLAG M2 affinity gel (Sigma-Aldrich, Milan, Italy). In addition, EMILIN2 was also synthesized by means of the RTS 500 E. coli HY rapid translation system kit (Roche Diagnostics, Milan, Italy) according to the manufacturer's instructions; as a template, 15 μg of the EMILIN2 pcDNA3.1 Myc-His construct was used, and synthesis was carried out for 24 h. The protein was ultimately purified by means of Ni-NTA resin dialyzed against phosphate-buffered saline and concentrated. SK-UT-1 and HT1080 cell lines were transfected using FuGene6 reagent (Roche Diagnostics) either transiently or stably upon selection with 600 μg/ml and 500 μg/ml G418, respectively. Retroviral gene transfer was performed as described previously (20), and HT1080 cells were selected with 1 μg/ml of puromycin. Protein expression was analyzed by Western blotting.

Immunoblot analysis, immunoprecipitation, and glutathione S-transferase (GST) pull-down assay. (i) Immunoblot.Forty micrograms of protein lysates or immunoprecipitates was resolved in 4 to 20% Criterion precast gels (Bio-Rad Laboratories) and transferred onto Hybond-ECL nitrocellulose membranes (Amersham, GE Healthcare). Membranes were blocked with 5% dry milk in Tris-buffered saline-Tween (TBS-T) buffer for 1 h and then incubated with the various antibodies for 1 h at room temperature. Following three washes with TBS-T buffer, the membranes were incubated with the appropriate secondary antibodies followed by three washes with TBS-T buffer.

Blots were developed using ECL (Amersham Biosciences). Alternatively, the Odyssey infrared imaging system was used (Li-COR Biosciences). To evaluate EMILIN2 expression by BJ/ERM cells, the cells were incubated in serum-free medium or in the presence of IGF-I or EGF (50 ng/ml) for 48 h, and specific bands were detected by Western blotting following immunoprecipitation.

(ii) Immunoprecipitation.For immunoprecipitations, following 1 h of incubation with a 12.5 nM concentration of EMILIN2 at 37°C, cells were lysed in HNTG buffer (20 mM HEPES [pH 7.5], 150 mM NaCl, 0.1% Triton X-100, 10% glycerol) in the presence of complete protease inhibitor cocktail (Roche Diagnostics) and immunoprecipitated with anti-DR4, anti-DR5, anti-Fas, anti-TNFR1, or anti-Myc antibody as indicated. The EMILIN2 band was detected by means of the 828B3B3 antibody. Ewing's sarcoma biopsies were lysed with the same buffer. Three hundred micrograms of proteins of each sample was immunoprecipitated overnight at 4°C with 1 μg of the indicated antibodies, and the immunocomplexes were captured with protein A/G PLUS-agarose (Santa Cruz Biotechnology) and separated by 4 to 20% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

(iii) GST pull-down assay.Ten micrograms of recombinant DR4 and two DR5 splice variant (DR5_S and DR5_L) extracellular regions fused to the GST sequence was immobilized on glutathione-Sepharose 4B beads (Amersham Biosciences) at 4°C for 1 h in the presence of TEN buffer (20 mM Tris, 0.1 mM EDTA, 100 mM NaCl). GST alone was used as a control. One microgram of purified recombinant EMILIN2 or in vitro-transcribed and -translated molecules with the TNT-coupled reticulocyte lysate system (Promega Italia) in the presence of [³⁵S]Met (Amersham Biosciences) diluted in TEN buffer containing 2% polyoxyethylene(9)nonylphenyl ether (Igepal; Sigma-Aldrich) were added to the beads, and the incubation was carried out at 4°C for 1 h. Proteins were separated by SDS-PAGE and revealed by Western blotting and/or by exposure to X-ray film. Alternatively, the GST pull-down experiment using GST-DR4 and equal amounts of [³⁵S]Met-labeled EMILIN2 was performed following preincubation of the GST chimera with increasing concentrations of recombinant TRAIL (from 0 to 1 μg of protein).

(iv) DISC formation.DISC formation analysis was conducted as previously described (31). Briefly, 2 × 10⁷ HeLa cells were collected and resuspended in serum-free medium, medium containing an equimolar amount of EMILIN2 or TRAIL (37.5 nM), or plain serum- free medium as a control and incubated for 30 min at 37°C. Cells were then washed and lysed with lysis buffer (30 mM Tris-HCl [pH 7.5], 150 mM NaCl, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, protease inhibitor cocktail [Roche Diagnostics, Milan, Italy], 1% Triton X-100, and 10% glycerol), incubated on ice for 30 min, and clarified by centrifugation. Lysates were then immunoprecipitated with 2 μg of anti-DR4 antibody for 4 h at 4°C with constant rotation, and the complexes were captured with protein A/G PLUS-agarose. Beads were washed three times with lysis buffer, and the proteins were detached from the beads by resuspension in SDS-PAGE loading buffer and boiling. Proteins were resolved on 4 to 20% Criterion precast gels and transferred for immunoblot; the Trueblot secondary antibody (eBiosciences, San Diego, CA) was used to avoid the detection of primary antibodies used for the immunoprecipitation.

Binding studies.EMILIN2 (13 μg) was labeled with high specific activity (0.3 × 10¹⁸ cpm mol⁻¹) using Iodogen-coated tubes (Pierce Biotechnology Inc., Rockford, IL) and purified with D-Salt Polyacrylamide 6000 desalting columns (Pierce) according to the manufacturer's instructions. For saturation binding, increasing concentrations of ¹²⁵I-labeled EMILIN2 were used to incubate 3 × 10⁵ HT1080 cells transfected with wtDR4 or with the empty vector. Cells were incubated for 1.5 h at 4°C. Nonspecific binding was determined by displacement with an excess of cold EMILIN2. Calculations and graphs were obtained by using Sigma Plot 9.0 software.

MTT and TUNEL assays. (i) MTT.Cells (10 × 10³) were plated in quadruplicates in 96-well plates; after EMILIN2 treatment, 5 mg/ml of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium (MTT) was added, and the cells were incubated for three extra hours. The formazan crystals were solubilized with dimethyl sulfoxide (DMSO), and the absorbance was detected at 560 nm. For the tests performed using wild-type A3 and caspase-8 mutant I 9.2 Jurkat cell lines, 10 × 10⁵ cells were seeded into 96-well plates; after the treatments, the medium was removed following a brief centrifugation, and the pellet was solubilized with DMSO. For each graph, a 100% viability index was assigned to the higher absorbance value.

(ii) TUNEL.HT1080 cells were treated with EMILIN2 with or without Z-VAD-FMK (100 μM). Apoptosis was analyzed after overnight incubation with the Cell Death Detection ELISA^PLUS terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) assay. To precisely determine the percentage of apoptotic cells, following the various treatments, the ApopTag Fluorescein In Situ Oligo Ligation apoptosis detection kit (CHEMICON International, Temecula, CA) was used according to the manufacturer's instructions, and the positive nuclei were counted.

Indirect immunofluorescence.Cells (10⁴) were plated in a Lab-Tek chamber slide (Nalge Nunc Corp.) and grown overnight at 37°C. EMILIN2 or fibronectin (FN) (15 nM) was added. After 30 min to 24 h of incubation, the cells were fixed with 4% (wt/vol) paraformaldehyde, blocked with 2% bovine serum albumin, and incubated with the appropriate antibodies. When needed, 1 μg/ml of Hoechst 33258 (Sigma-Aldrich) was added to label the cell nuclei. Images were acquired with a confocal system (Leica Microsystems).

Soft agar colony assay and growth in a three-dimensional (3D) Matrigel matrix. (i) Soft agar colony assay.Low-melting-point agarose (0.5%) was solidified. Retrovirally transduced HT1080 cells (5 × 10⁵) in complete medium containing 0.3% low-melting-point agarose were placed on top of the bottom agar layer. Pictures were taken after 10 days of incubation, and the clones formed by more than 10 cells were counted.

(ii) Growth in Matrigel.SK-UT-1 cells (5 × 10⁵) transfected with the pcDNA3.1 constructs were embedded in 500 μl of Matrigel diluted 1:3 with serum-free medium (BD Biosciences). Pictures were taken after 10 days of incubation. Some of the Matrigel drops were labeled with the human annexin V-fluorescein isothiocyanate kit (Bender MedSystems GmbH, Vienna, Austria) according to the manufacturer's instructions. The solutions were used in order to cover the Matrigel drop and allow sufficient diffusion. Alternatively, 5 × 10⁵ RD-ES cells were embedded in 500 μl of Matrigel diluted 1:3 with serum-free medium. In this case, the Matrigel was enriched with the addition of 37.5 nM FN or EMILIN2. Pictures were taken after 10 days of incubation.

EMILIN2 and TRAIL gene silencing.EMILIN2 expression knockdown was achieved with the DHARMACON siRNA Specific SMARTpool according to the manufacturer's instructions. Briefly, BJ/ERM cells plated onto 24-well plates were transfected with the specific and nontargeting SMARTpools by means of the DHARMAfect 3 reagent. Cells were then kept in serum-free medium for 72 h to induce apoptosis, and a TUNEL assay was performed using the In situ Cell Death Detection kit, fluorescein (Roche Diagnostics, Milan, Italy). The conditioned medium of untransfected cells and cells transfected with EMILIN2-targeting or nontargeting siRNA was added to HT1080 target cells. After 24 h of incubation, HT1080 cell viability was assessed by MTT assay. EMILIN2 knockdown was assessed by reverse transcription-PCR (RT-PCR). For the TRAIL knockdown, HT1080 cells plated onto 24-well plates were transfected with the specific or nontargeting SMARTpools, and the next day, they were incubated overnight with EMILIN2 or the control conditioned medium. Cell death was evaluated with the Apoptosis Detection kit (CHEMICON International, Temecula, CA) according to the manufacturer's instructions, and TRAIL knockdown was assessed by RT-PCR.

Statistical analysis.The data were analyzed by a Student's t test. Data were taken to be significant when the P value was <0.05. Graphs and calculations were made with SigmaPlot software, version 9 (Systat Software Inc., Point Richmond, CA). Images were assembled using Adobe Illustrator CS and Adobe Photoshop CS.

EMILIN2 inhibits the growth of tumor cell lines.When 293-EBNA cells were transfected with full-length EMILIN2 cDNA, quite unexpectedly from what was observed with EMILIN1 (25), we noticed that EMILIN2-transfected cells displayed reduced growth and that cells failed to reach confluence. To shed light on this phenomenon, we treated different cell lines (HT1080 and SK-UT-1 sarcoma cell lines and the HeLa carcinoma cell line) with EMILIN2. The presence of EMILIN2 induced a striking decrease in cell viability, as evaluated by MTT assay. This growth-suppressive effect was further confirmed by expressing EMILIN2 in SK-UT-1 and HT1080 cells by either transfection or retroviral infection (Fig. 1A). Next, His-tagged EMILIN2 was purified to homogeneity (see Fig. S1A in the supplemental material), and increasing concentrations of the recombinant molecule were used to treat HT1080 cells. The presence of EMILIN2 induced a lethal effect that was dose dependent, with a 50% reduction of cell viability at ∼10 nM (Fig. 1B). EMILIN2 action was also time dependent; in fact, after 12 h of incubation, a 50% reduction of cell viability was detected (Fig. 1C).

• Open in new tab
• Download powerpoint

FIG. 1.

EMILIN2 decreases cell survival by triggering apoptosis. (A) MTT assays performed on SK-UT-1 cells transfected with an EMILIN2 pcDNA construct (pc-E2) or a mock-transfected control (pc-m.), HT1080 cells transduced with an EMILIN2 retroviral construct (pL-E2) or a mock-transduced control (pL-m.), and HeLa cells challenged with conditioned medium from EMILIN2 (E2) or mock-transfected (C.M.) 293-EBNA cells. Incubations were carried out overnight, and representative pictures of the experiments are shown on the right side of each graph, whose values represent the means ± standard errors (SE) for five separate experiments expressed as percent viability. Bar, 40 μm. (B) MTT assay of the dose-response experiment performed on HT1080 cells incubated overnight in serum-free medium (S.F.) or increasing concentrations of purified EMILIN2 (E2). Values represent the means ± SE for three independent experiments expressed as percent viability. (C) MTT assay of the time course experiment performed on HT1080 cells treated with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells. Cell viability was evaluated at the indicated times, and values represent the means ± SE of three independent experiments and are expressed as percent viability. (D) Enzyme-linked immunosorbent assay-based TUNEL assay performed on HT1080 cells incubated overnight with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells with or without the general caspase inhibitor Z-VAD-FMK (Z) used at a concentration of 100 μM. Values represent the means ± SE for three independent experiments. (E) Representative higher-magnification picture of HT1080 cells showing cell blebbing following treatment with equimolar amounts (37.5 nM) of FN, EMILIN2 (E2), or TRAIL after overnight incubation. Bar, 10 μm. (F) TUNEL assay performed on HT1080 cells following overnight incubation under the same experimental conditions as those described above (E). Values represent the means ± SE for three independent experiments (*, P ≤ 0.05).

EMILIN2 triggers apoptosis in tumor cell lines via the extrinsic apoptotic pathway.Given the fast kinetics of EMILIN2 growth-suppressive effects, we hypothesized a proapoptotic function. To assess whether EMILIN2 killed cells by programmed cell death, we performed a TUNEL assay and found significant DNA fragmentation in the presence of EMILIN2 (Fig. 1D and F). Moreover, the addition of a pancaspase inhibitor rescued EMILIN2-induced DNA fragmentation and cell death, indicating that the phenomenon was caspase dependent (Fig. 1D). The extracellular nature of EMILIN2 suggested a possible engagement of cell surface death receptors and the activation of the extrinsic apoptotic pathway. Thus, to investigate the proapoptotic function of EMILIN2, we compared it to TRAIL, a death receptor-specific ligand. Both EMILIN2 and TRAIL were produced in 293-EBNA cells and TRAIL purified by means of a FLAG tag (see Fig. S1A and S1B in the supplemental material). As opposed to cells treated with FN, another ECM molecule used as a control, HT1080 cells exposed to equimolar amounts of EMILIN2 and TRAIL showed membrane blebbing, a hallmark of cells undergoing apoptosis (Fig. 1E). Additionally, the EMILIN2 proapoptotic activity was comparable to that of TRAIL as detected by both TUNEL and MTT assay (Fig. 1F and see Fig. S1C in the supplemental material, respectively). To rule out the possibility that the effects exerted by EMILIN2 could depend on the synthesis of other proapoptotic molecules induced by its overexpression, the recombinant protein was depleted from the medium of transfected 293-EBNA cells. The removal of the protein almost completely abolished the decrease in cell viability induced by the conditioned medium derived from EMILIN2-transfected cells, indicating that the presence of EMILIN2 is necessary and sufficient for this phenomenon to occur (Fig. 2A). Accordingly, the use of the protein synthesis inhibitor cycloheximide did not affect the activity of exogenously added EMILIN2, nor did it affect that of TRAIL, indicating that these two molecules did not require the de novo synthesis of additional proteins to exert their effects (Fig. 2B). Moreover, EMILIN2 activity was independent of the presence of TRAIL. First, TRAIL expression was not affected by EMILIN2 (Fig. 2C). Second, TRAIL silencing had no influence on EMILIN2 proapoptotic effects as detected by a TUNEL assay performed on HT1080 cells following TRAIL knockdown (Fig. 2D and E). Finally, the possibility that EMILIN2 could bind additional molecules produced by 293-EBNA cells, thus exerting an indirect effect, was excluded by its in vitro synthesis in a cell-free context. The purified molecule, despite requiring higher concentrations (possibly due to incomplete folding), was nonetheless effective at inducing apoptosis (see Fig. S1D and S1E in the supplemental material).

• Open in new tab
• Download powerpoint

FIG. 2.

EMILIN2 is necessary and sufficient for apoptotic induction. (A) MTT assay performed on HT1080 cells treated overnight with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells with or without depletion of His-tagged recombinant protein through Ni-NTA resin (Ni). Values represent the means ± SE for three independent experiments expressed as percent viability. EMILIN2 depletion was confirmed by Western blot analysis (top). (B) MTT assay performed on HT1080 cells incubated for 8 h with equimolar amounts (37.5 nM) of recombinant EMILIN2 (E2), TRAIL, or FN in the presence of cycloheximide (50 μg/ml) or DMSO. Values represent the means ± SE for three independent experiments and are expressed as percent viability. (C) RT-PCR (top) and Western blot (bottom) evaluations of TRAIL content in HT1080 cells upon treatment with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells for 6 h. Actin was used as a normalizer. (D) Levels of TRAIL mRNA upon TRAIL silencing by siRNA (si TRAIL) as detected by RT-PCR. Nontargeting siRNA (si C.) was used as a control, and actin was used as a normalizer. (E) TUNEL assay performed on HT1080 cells transfected with control (si C.) and TRAIL (si TRAIL) siRNA and subsequently treated with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells. Values represent the means ± SE for three independent experiments (*, P ≤ 0.05).

The activation of the molecules of the apoptotic machinery was monitored in different cell lines either after treatment with EMILIN2 or upon transfection with the EMILIN2 construct. To verify the hypothesis that EMILIN2 activated the extrinsic apoptotic pathway, cleavage of caspase-8 and caspase-10 was evaluated. Challenging HT1080 cells with EMILIN2 strongly induced the processing of caspase-10 and caspase-8 (Fig. 3A and B), whereas no activation was observed in cells exposed to FN. In addition, similar to what was observed with TRAIL and in support of our hypothesis, cell death occurred only in wild-type Jurkat but not in caspase-8-deficient I 9.2 cells (Fig. 3C). The overexpression of EMILIN2 reduced the levels of procaspase-3 and induced the processing of PARP (Fig. 3D), indicating that EMILIN2 thoroughly stimulated the chain of events that led to cell death. Taken together, our results supported a possible involvement of EMILIN2 in the initiation of the extrinsic apoptotic pathway. Further evidence was provided by the finding that the treatment of HeLa cells with EMILIN2 led to DISC assembly. In fact, upon preincubation of the cells with recombinant EMILIN2, immunoprecipitation of DR4 revealed the recruitment of FADD and caspase-8, similar to what was observed previously with the use of TRAIL (3). No FADD or caspase-8 recruitment was detected when cells were incubated with serum-free medium (Fig. 3E).

• Open in new tab
• Download powerpoint

FIG. 3.

EMILIN2 elicits the extrinsic apoptotic pathway. (A, top) Immunoblot analysis of procaspase-10 and the active cleaved subunit conducted on HT1080 cell lysates treated with 37.5 nM FN or EMILIN2 (E2). (Bottom) Immunoblot analysis of procaspase-8 and the active cleaved subunits conducted on HT1080 cell lysates after treatment with purified EMILIN2 (pE2) or FN as a control. Αctin was used as a normalizer of protein loading. (B) Evaluation of caspase-8 activity. HT1080 cells were treated with 400 nM staurosporine (S.) or 37.5 nM collagen type I (C.I.) or an equimolar amount of EMILIN2 (E2). After 6 h of incubation, active caspase-8 was traced by means of a specific fluorescent marker. Bar, 10 μm. (C) MTT assay performed on either wild-type A3 Jurkat cells (A3 J) or caspase-8-deficient I 9.2 Jurkat cells (I 9.2 J) following treatment with equimolar amounts of recombinant EMILIN2 (pE2) or TRAIL (pTRAIL). Values represent the means ± SE for three independent experiments and are expressed as percent viability (*, P ≤ 0.05). (D) Western blot analysis of PARP processing and of caspase-3 activation following two independent transient transfections of SK-UT-1 cells with the pcDNA EMILIN2 construct (pc-E2 a and b) compared to mock-transfected cells (pc-m.). An antibody specific for the 89-kDa subunit was used to detect PARP. Overexpression of EMILIN2 was confirmed by means of the 828B3B3 antibody. Αctin was used as a normalizer of protein loading. (E) DISC immunoprecipitation and Western blot analysis showing enhanced recruitment of procaspase-8 and FADD into the DISC in HeLa cells incubated for 30 min with serum-free medium (S.F.) or medium containing equimolar amounts (37.5 nM) of purified TRAIL (pTRAIL) or EMILIN2 (pE2).

EMILIN2 is a novel ligand for TRAIL-R1 (DR4).To investigate whether the activation of the extrinsic apoptotic pathway by EMILIN2 could occur following direct engagement of death receptors, HT1080 cells were incubated with EMILIN2, and the lysates were immunoprecipitated with antibodies directed against Fas and TNFR1 or with DR4 and DR5, death receptors known to play an important role in the regulation of solid tumor survival. No detectable binding was found with Fas or TNFR1 (see Fig. S2A in the supplemental material). On the contrary, an interaction with endogenous DR4 and, to a lesser extent, with DR5 was evident (Fig. 4A). To further verify this finding, we generated GST fusion proteins corresponding to the extracellular region of DR4 (tDR4) and to two splice variants of DR5 (tDR5_L and tDR5_S) in order to avoid possible nonspecific interactions with the transmembrane and intracellular regions. The chimeras were incubated with His-tagged recombinant EMILIN2. GST pull-down experiments provided further evidence for the EMILIN2-DR interaction. In fact, as shown in Fig. 4B, high levels of EMILIN2 coprecipitated with DR4, and weaker interactions were also detected with the two forms of DR5. No specific binding was detected with the GST negative control. To analyze this interaction in more detail, the GST-DR4 chimera was also used in a GST pull-down experiment in the presence of the in vitro-transcribed and -translated TRAIL, EMILIN2, or EMILIN2 C-terminal fragment (E2 del) (amino acids 788 to 1053). This experiment confirmed the interaction of DR4 with EMILIN2 and with the TRAIL positive control but not with the EMILIN2 C-terminal fragment, which was used as a negative control due to its lack of in vitro activity (data not shown). The latter result implies that the EMILIN2-DR4 interaction is direct and that it does not involve the presence of any additional molecule. For the same reason, the interaction must occur via the protein core, since the recombinant His₆-EMILIN2 protein is not glycosylated, and it likely depends on the N portion of the molecule (Fig. 4C).

• Open in new tab
• Download powerpoint

FIG. 4.

EMILIN2 binds to DR4, leading to its activation. (A) Western blot analysis of immunoprecipitation (IP) with anti-DR4 or anti-DR5 antibodies of HT1080 cell lysates following incubation with serum-free medium containing a 12.5 nM concentration of EMILIN2. The anti-Myc antibody was used as a control; the EMILIN2 band was revealed with the 828B3B3 specific monoclonal antibody, and the receptors were revealed with specific anti-DR4 or anti-DR5 antibodies (α-DR). (B) Western blot analysis of the GST pull-down experiment performed using the extracellular regions of DR4 and of two DR5 splice variants (DR5_S and DR5_L) fused to the GST sequence or GST alone as a control and incubated with 1 μg of EMILIN2. EMILIN2 was detected with the 828B3B3 monoclonal antibody (α-E2), and the GST fusion chimeras were detected with an anti-GST antibody (α-GST). (C) GST pull-down experiment using the in vitro-transcribed and -translated TRAIL, EMILIN2 (E2), and C-terminal fragment (E2-del) (left). The pull-down experiment (right) was performed using the GST-DR4 chimera, and the bands were revealed following overnight exposure to X-ray films; the GST-DR4 chimera was revealed by Coomassie staining. (D) Western blot analysis of immunoprecipitation showing an in vivo interaction in Ewing's sarcoma tumors. Tumor (left) or normal (right) tissues (0.3 mg) were lysed and precipitated with anti-DR4 and anti-TNFR1 antibodies. EMILIN2 was detected with the 828B3B3 monoclonal antibody (α-E2), and DR4 was detected with an anti-DR4 antibody (α-DR4). (E) Saturation binding study of ¹²⁵I-labeled EMILIN2 using HT1080 cells transfected with the DR4 pcDNA construct (wtDR4) (see below) or with the empty vector (mock). The incubation was carried out for 1 h at 4°C. Values represent the means ± SE for three independent experiments run in triplicate. (F) Clusterization and capping of DR4 induced by EMILIN2 (white arrows). HT1080 cells were treated for 6 h with EMILIN2 (E2) or FN as a control. Right panels show higher magnifications. (G) Capped DR4 localizes in lipid rafts (white arrows). Shown is double immunofluorescence of DR4 stained with a specific antibody and lipid rafts stained with fluorescent cholera toxin B (CTB). Bar, 10 μm.

To establish if the interaction between EMILIN2 and death receptors could occur in vivo, Ewing's sarcoma tumors expressing high levels of DR4 and DR5 were used (24). Extracts from Ewing's sarcoma tumor cells or normal peripheral lymphocytes were immunoprecipitated with the anti-DR4 antibody or anti-TNFR1 as a control (Fig. 4D). The immunoprecipitates, analyzed by Western blotting using anti-EMILIN2 and anti-DR4 specific antibodies, confirmed that the interaction between EMILIN2 and DR4 also occurs in vivo. Additionally, ligand binding studies using ¹²⁵I-labeled EMILIN2 performed on HT1080 cells transfected with wtDR4 or with empty vector (see below) provided further evidence of this interaction, with the calculated affinity being 3.7 nM (Fig. 4E).

Since an excess of recombinant TRAIL did not compete with EMILIN2 for binding to DR4 (see Fig. S2B in the supplemental material), it appears that the two molecules do not share the same binding site(s) within the DR4 extracellular region.

The DR4-EMILIN2 interaction triggers apoptosis.It is known that upon engagement by their ligands, death receptors trimerize and segregate. The co-redistribution of these receptors with lipid rafts correlates with the activation of programmed cell death (17, 26). We thus investigated whether EMILIN2 binding to DR4 induced similar effects. HT1080 cells were incubated with EMILIN2 at a concentration of 37.5 nM for 6 h. As revealed by specific anti-DR4 antibodies, only EMILIN2 challenge induced DR4 segregation and capping (Fig. 4F), and, similar to what was observed for the known death receptor ligands, EMILIN2 induced the localization of the DR4-segregated receptors in lipid rafts (Fig. 4G).

In order to substantiate the role of the EMILIN2-DR4 interaction in modulating cell death, the pcDNA 3.1 Myc-His constructs of the wild type (wtDR4) and of a defective form of DR4 (tDR4) lacking the intracellular region were used to stably transfect HT1080 cells. Positive clones were identified by Western blotting using anti-Myc antibody (Fig. 5A). Both the wtDR4 and tDR4 constructs were expressed at the cell surface, as shown by confocal microscopy and fluorescence-activated cell sorter analysis (Fig. 5B). Transfected clones were then incubated with EMILIN2 and evaluated for cell survival. The overexpression of wtDR4 strongly sensitized HT1080 cells to EMILIN2-induced cell death; on the contrary, the tDR4 dominant negative form prevented EMILIN2-induced cell death (Fig. 5C). Accordingly, the stable overexpression of FLIP_S in HT1080 clones (Fig. 6A) partially blocked the EMILIN2- and the TRAIL-induced proapoptotic effects, a further indication that EMILIN2 triggered cell death by activating the extrinsic apoptotic pathway (Fig. 6B). Overall, our results support the concept that the ECM component EMILIN2 induces apoptosis by a death receptor-mediated pathway.

• Open in new tab
• Download powerpoint

FIG. 5.

The cytoplasmic portion of DR4 is necessary for EMILIN2 proapoptotic activity. (A) Western blot analysis of the lysates of HT1080 clones transfected with the wtDR4 pcDNA Myc-His construct (top) or with the construct of a truncated mutant lacking the cytoplasmic region (tDR4) (bottom). Bands were revealed with an anti-Myc antibody. Clones numbers are indicated, and actin was used as a normalizer of protein loading. (B, left) Immunofluorescence analysis of fixed and permeabilized HT1080 wtDR4 clone 20 (wtDR4), HT1080 tDR4 clone 12 (tDR4), or empty vector (mock) showing localization of the recombinant proteins at the cell surface. Bars, 20 μm. (Right) Fluorescence-activated cell sorter analysis of the same cells. (C) MTT assay performed on HT1080 cells transfected with the wtDR4 or with the tDR4 construct or with empty vector (mock) following 6 h of incubation with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells Values represent the means ± SE for three independent experiments and are expressed as percent viability (*, P ≤ 0.05).

• Open in new tab
• Download powerpoint

FIG. 6.

EMILIN2 activity is hindered by the extrinsic apoptotic pathway inhibitor FLIP_S. (A) Western blot analysis of the lysates of HT1080 stable clones transfected with a FLIP_S construct (pCR-F) or empty vector (pCR-m). Actin was used as a normalizer. (B) MTT assay performed on two mock-transfected clones (pCR-m) (lanes 1 and 8) and two clones highly expressing FLIP_S (pCR-F) (lanes 6 and 17) following overnight incubation with equimolar amounts (25 nM) of collagen type I (C.I), EMILIN2 (E2), or TRAIL. Values represent the means ± SE for three independent experiments and are expressed as percent viability (*, P ≤ 0.05).

Transformation of normal fibroblasts up-regulates DR4 and DR5 and sensitizes cells to EMILIN2 effects.The induction of cell death through the ligation of EMILIN2 to DR4 and DR5 represents a novel function for an ECM constituent. To investigate whether EMILIN2 secretion affected cell behavior differently in normal versus transformed cells, EMILIN2 expression was evaluated in human BJ fibroblasts transformed by the ectopic expression of E1A, HA-Ras-V12, and MDM2 (BJ/ERM) and in normal parental BJ cells (32). RT-PCR analysis of the two cell types revealed no significant change in EMILIN2 mRNA expression in transformed versus normal fibroblasts (Fig. 7A). On the contrary, higher expression of DR4 and DR5 was found in BJ/ERM cells than in the normal cells, and this correlated with an increased sensitivity to the EMILIN2-induced proapoptotic effect (Fig. 7B).

• Open in new tab
• Download powerpoint

FIG. 7.

Fibroblast transformation leads to the up-regulation of DR4 and DR5 and subsequent sensitization to EMILIN2 activity. (A) PCR analysis of EMILIN2 (E2) (top) and of DR4 and DR5 mRNA content (bottom) in normal human fibroblasts (BJ) compared to transformed BJ cells (BJ/ERM). Αctin was used as a normalizer. (B) MTT assay performed on BJ and BJ/ERM cells upon overnight treatment with conditioned medium from EMILIN2-transfected (E2) or mock-transfected (C.M.) 293-EBNA cells. Values represent the means ± SE for three independent experiments and are expressed as percent viability. (C) Western blot analysis of EMILIN2 expression by BJ/ERM cells maintained in serum-free medium (S.F.) or in the presence of 50 ng/ml of EGF or IGF-I. The relative band quantification was obtained with the Odyssey Server software. (D) RT-PCR analysis of EMILIN2 mRNA content following transfection of BJ/ERM cells after EGF stimulation with the EMILIN2-specific siRNA (si E2). Nontargeting siRNA (si cont) and nontransfected cells (S.F.) were used as a control. The experiment was conducted over 72 h in 0.2% FBS. (E) TUNEL assay performed on BJ/ERM cells following transfection with siRNA as described above (D) Values represent the means ± SE for three independent experiments. (F) MTT assay performed on HT1080 cells after overnight incubation with the BJ/ERM conditioned medium as described above (D). Values represent the means ± SE for three independent experiments and are expressed as percent viability (*, P ≤ 0.05).

Since growth factors are implicated in the modulation of tumor initiation and progression (2), we sought to investigate whether growth factor stimulation of transformed BJ cells could affect EMILIN2 expression. Cell treatment with human recombinant EGF or with IGF-I resulted in increased EMILIN2 expression due to either enhanced synthesis or increased protein stability (Fig. 7C). BJ/ERM cells were thus prestimulated with EGF to enhance the quantity of EMILIN2. Prestimulated BJ/ERM cells were then starved to induce apoptosis, which was significantly prevented upon EMILIN2 silencing (Fig. 7D and E). To confirm that apoptosis relied on secreted EMILIN2, HT1080 cells were incubated with conditioned medium from BJ/ERM cells. As shown in Fig. 7F, medium derived from EMILIN2 siRNA-treated BJ/ERM cells was more permissive to cell survival than was EMILIN2-rich medium from control siRNA-treated cells.

EMILIN2 antineoplastic effects in a 3D context.Given the proapoptotic effects exerted by EMILIN2 in cell culture, we sought to investigate whether this molecule could also have antineoplastic activity in 3D in vitro models. To test this hypothesis, a soft agar assay was first performed on HT1080 cells overexpressing EMILIN2. EMILIN2 significantly reduced the capability of the cells to grow in semisolid medium (Fig. 8A), affecting both the number and the size of the colonies. Second, EMILIN2 also inhibited tumor cell growth in a natural matrix. In fact, both the RD-ES Ewing's sarcoma cell line challenged with recombinant EMILIN2 (Fig. 8B) as well as the SK-UT-1 sarcoma cell line overexpressing EMILIN2 (Fig. 8C) failed to grow, spread, and migrate within a Matrigel drop, whereas cells confronted with FN or mock-transfected cells easily colonized the semisolid matrix. This outcome was likely dependent on the high intra-Matrigel apoptotic levels detected by annexin V staining performed on the SK-UT-1 cells embedded in the Matrigel drop (Fig. 8D).

• Open in new tab
• Download powerpoint

FIG. 8.

EMILIN2 also exerts antineoplastic effects in 3D matrices. (A) Soft agar colony assay performed on mock (pLPC)- or EMILIN2 (pLPC-E2)-transduced HT1080 cells. After 10 days of incubation, pictures of the grown colonies were taken, and the colonies were counted; the means ± SE for three independent experiments are reported on the left. (B) RD-ES cells embedded in a Matrigel drop containing equimolar amounts (37.5 nM) of FN or EMILIN2 (E2). Pictures were taken after 10 days of incubation. (C) SK-UT-1 cells transfected with the EMILIN2 (E2) pcDNA construct or empty vector (mock) embedded in a Matrigel drop. The Matrigel edge is marked, and the inside and outside are indicated (in and out, respectively). (D) Intra-Matrigel apoptotic rate of EMILIN2 (E2)- and mock-transfected SK-UT-1 cells using annexin V staining. Nuclei are stained with Hoechst stain (*, P ≤ 0.05). Bar, 10 μm.