Sensitization of interferon-γ induced apoptosis in human osteosarcoma cells by extracellular S100A4

The human osteosarcoma cell line OHS and its anti-S100A4 ribozyme transfected counterpart II-11b have been described previously [8]. Cell lines were cultivated in RPMI-1640 (Bio Whittaker, Verviers, Belgium) containing 10% fetal bovine serum (FBS; Biochrome KG, Berlin, Germany), 1 mM Hepes, and 2 mM Glutamax (GIBCO BRL, Life Technologies, Paisley, UK). For all experiments, subconfluent cultures were trypsinated and seeded at 1.5 × 10⁴ cells/cm². After overnight incubation, the culture medium was replaced with medium in presence or absence of IFN-γ and recombinant S100 proteins, and harvested as indicated in the text. Where indicated, protease inhibitors or antioxidants were added to the cell culture simultaneously as IFN-γ.

Human recombinant IFN-γ, L-NMMA, zVAD-fmk, zYVAD-fmk, zVDVAD-fmk, zDEVD-fmk, zVEID-fmk, zIETD-fmk, zLEHD-fmk, zFA-fmk and E64 were purchased from Calbiochem (San Diego, CA, USA). N-acetylcysteine was from Sigma Chemical Co (St Louis, MO, USA).

Histidine-tagged mouse recombinant S100A4 was cloned into the pQE30 vector (kindly provided by E. Lukanidin, Institute of Cancer Biology, Copenhagen, Denmark), expressed in E. coli, purified on a Ni²⁺-column and dialyzed against phosphate-buffered saline. A satisfactory degree of purity was confirmed by SDS polyacrylamide gel electrophoresis (SDS-PAGE) where a single band was visualized by silver staining. Verification of the identity of the protein was performed using matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI). Human recombinant S100A10 and S100A13 tagged with maltose-binding protein were kindly provided by C. S. Skjerpen (Dep. of Biochemistry, The Norwegian Radium Hospital, Oslo, Norway).

Cell viability was estimated by measuring metabolic activity using CellTiter 96^® AQ_ueous One Solution Reagent (MTS) (Promega, Madison, WI, USA) according to the manufacturer's manual. The absorbance at 490 nm was recorded using a Wallac 1420 Victor² Multilabel counter (Wallac Oy, Turku, Finland).

Protein lysates were prepared in 50 mM Tris-HCl (pH 7.5) containing 150 mM NaCl and 0.1% NP-40 with 1 mM PMSF and 2 μg/ml each of pepstatin, aprotinin (Sigma Chemical Co, St Louis, MO, USA) and leupeptin (Roche Diagnostics, Mannheim, Germany). Total protein lysate from each sample was separated on 8–12% SDS-polyacrylamide gel electrophoresis, and transferred onto Immobilon-P membranes (Millipore, Bedford, MA, USA) according to the manufacturer's manual. As a loading and transfer control, the membranes were stained with 0.1% amidoblack. After blockage of non-specific binding sites with 10% non-fat dry milk in Tris-buffered saline with 0.25% Tween (TBST), blots were incubated for 1 h at room temperature with rabbit polyclonal anti-poly (ADP-ribose) polymerase (PARP; diluted 1:1000, Roche Diagnostics, Mannheim, Germany), mouse monoclonal anti-lamin B (0.3 μg/ml; Oncogene Research Products, Boston, MA, USA), mouse monoclonal anti-α-tubulin (0.3 μg/ml; Oncogene Research Products, Boston, MA, USA), rabbit polyclonal anti-caspase-1 (1:100; Santa Cruz Biotechnology, Santa Cruz, CA, USA), goat polyclonal anti-caspase-3 (1:2000; R&D Systems, Minneapolis, MN, USA), mouse monoclonal anti-caspase-8 (1:5000; Alexis Biochemicals, Montreal, Canada), mouse monoclonal anti-caspase-9 (1:1000; R&D Systems, Minneapolis, MN, USA), rabbit polyclonal anti-caspase-12 (1:2000; Oncogene Research Products, San Diego, CA, USA), mouse monoclonal anti-NFκB p65 (1:1000, Cell Signaling Technology Inc. Beverly, MA, USA) or mouse monoclonal anti-phospho-IκBα (Ser32/36) (1:2000, Cell Signaling Technology Inc. Beverly, MA, USA). After washing, the blots were incubated for 1 h at room temperature with a horseradish peroxidase-conjugated secondary antibody (DAKO, Glostrup, Denmark) diluted 1:5000. Anti-lamin B, anti-α-tubulin, anti-PARP, anti-caspase-1, anti-caspase-8, anti-NFκB p65 and anti-phospho-IκBα were diluted in TBST with 5% non-fat dry milk, anti-caspase-3 and anti-caspase-12 were diluted in TBS containing 0.05% Tween and 2% non-fat dry milk and anti-caspase-9 was diluted in TBS with 0.05% Tween, 1% non-fat dry milk and 1% BSA. Signals were visualized using the ECL chemiluminescence substrate (Amersham Pharmacia Biotech, Buckinghamshire, UK).

Cells were resuspended in lysis buffer (10 mM Tris-HCl pH 7.4, 10 mM NaCl, 10 mM EDTA, 0.5 % SDS and 0.5 μg/ml proteinase K) and incubated for 1 hour at 50°C. High molecular weight DNA was precipitated by adding 1 M NaCl and subsequently incubated overnight at 4°C. Samples were centrifuged for 30 min, 2700 g at 4°C, and supernatants collected. 95 % ethanol was added, and DNA precipitated overnight at -20°C. Samples were centrifuged for 10 min, 10 000 rpm at 4°C, and pellets were washed in 70 % ethanol and re-centrifuged. Thereafter, pellets were air-dried and dissolved in 10 mM Tris-HCl (pH 7.0), 15 mM NaCl, 1 mM EDTA and 0.2 mg/ml RNAse A, and incubated for 1 h at room temperature. DNA was separated by electrophoresis in 1.0 % agarose gel containing ethidium bromide and photographed under ultraviolet illumination.

Cell culture medium from OHS and II-11b cells were harvested at 24, 48 and 72 hours, and immediately filtered using a 0.45 μm filter to discard cells in suspension. From each sample, 1.5 ml of medium was subjected to immunoprecipitation using 15 μg of rabbit polyclonal anti-S100A4 (DAKO, Glostrup, Denmark) or 15 μg rabbit polyclonal anti-p300 (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Primary antibody binding was carried out in cell culture medium at 4°C overnight. Thereafter, antibody was precipitated using Protein A Sepharose beads for 1 h at 4°C and eluted in loading buffer. Samples were subjected to Western blot analysis as described above using rabbit polyclonal anti-S100A4 diluted 1:300 in TBST.

The NF-κB reporter construct containing three NF-κB response elements (i.e., sites identical to the κB site from the Igκ light chain promoter) driving luciferase expression has been described previously [25]. 1.0 × 10⁶ II-11b cells were transiently transfected with 10 μg of the NF-κB reporter construct using electroporation (240 V). Thereafter, cells were seeded in 96-well plates at a density of 4.5 × 10⁴ cells/cm² and incubated in the absence or presence of rS100A4 and zFA-fmk. Cells were harvested 48 h later and assayed for luciferase activity using the Luciferase Assay System (Promega, Madison, WI, USA) according to the manufacturer's manual.

All statistical analyses were performed using Student's t-test. P-values less than 0.05 were considered to be statistically significant.

We have previously observed that the human osteosarcoma cell line OHS is more sensitive to IFN-γ induced apoptosis than the S100A4-ribozyme transfected counterpart II-11b [24]. To examine the effects of IFN-γ on cell viability in more detail, time- and dose-response experiments were performed (Fig. 1A and 1B). In OHS cells, addition of 1 u/ml IFN-γ had no effect on cell viability, while an increasing amount of cell death was observed with higher concentrations. On the other hand, II-11b cells only showed a marked reduction in cell number after treatment with 1000 u/ml IFN-γ for 72 hours.

Previously, we have shown that treatment of OHS cells with IFN-γ induced a significant increase in the fraction of apoptotic cells, while the cell cycle distribution remained unchanged [24]. Thus, it seemed plausible that the reduced cell viability could be due to induction of apoptosis, and to confirm this hypothesis, we investigated some of the known hallmarks of the apoptotic process. As seen in Fig. 2A, IFN-γ induced DNA fragmentation in OHS cells. Additionally, treatment with IFN-γ induced cleavage of poly (ADP-ribose) polymerase (PARP) and lamin B (Fig. 2B). Taken together, these results clearly demonstrated that IFN-γ-treatment induces apoptosis in OHS cells.

When performing the OHS cell culture experiments, we observed that, when keeping the concentration of IFN-γ constant, decreasing the volume of cell culture medium or increasing the number of cells resulted in increased cell death (data not shown). This observation suggested that the cells might secrete a soluble factor with impact on the level of apoptosis induced by IFN-γ. Since the II-11b cells are derived from the OHS cells by transfection with an S100A4-specific ribozyme, this finding prompted us to investigate whether S100A4 is secreted from these cells, and whether extracellular S100A4 could cooperate with IFN-γ in induction of apoptosis. Fig. 3 shows an immunoprecipitation of cell culture medium from OHS and II-11b cells grown for 24, 48 and 72 hours. This clearly demonstrated that S100A4 is present in cell culture medium from OHS cells, and to a much lesser extent from II-11b cells. A small amount of S100A4 was also detected in the cell culture medium control. This possibly reflects bovine S100A4 present in fetal calf serum added to the medium, but the exact nature of this signal has not been further investigated. To exclude cell lysis as the cause of extracellular S100A4, lactate dehydrogenase activity was measured in the conditioned culture medium. No significant increase in enzyme activity was detected, supporting the hypothesis that S100A4 is actively secreted (data not shown). Importantly, cell culture medium from OHS cells cultured for 72 hours immunoprecipitated with a rabbit polyclonal anti-p300 antibody did not give any S100A4-signal, indicating that the detection of S100A4 was not due to unspecific binding.

In order to investigate whether extracellular S100A4 could increase apoptosis induced by IFN-γ, recombinant S100A4 (rS100A4) was added to the cell culture medium of OHS and II-11b cells in addition to IFN-γ. Addition of rS100A4 alone had no effect on cell viability at concentrations ranging from 400 pg/ml to 20 μg/ml (Fig. 4 and data not shown). On the other hand, when 20 μg/ml of rS100A4 was added together with 100 u/ml IFN-γ to II-11b cells, a significant decrease in cell viability was observed (Fig. 4B). In fact, when II-11b cells were incubated with 20 μg/ml rS100A4, the level of apoptosis induced by IFN-γ was similar as in OHS cells. Addition of rS100A4 to OHS cells also increased the IFN-γ-mediated cell death, thus seemingly adding to the effects of the endogenously produced extracellular S100A4 (Fig 4A). Accordingly, a significant decrease in cell viability was seen at lower concentrations of added rS100A4 in OHS cells than in II-11b cells. Furthermore, we made attempts to neutralize the secreted S100A4 protein by adding anti-S100A4 antibodies to the cell culture medium of IFN-γ treated OHS cells, but were unable to detect any decrease in the amount of cell death (data not shown). Possibly, the antibodies used (DAKO, Glostup, Denmark, as well as three in-house produced monoclonal antibodies [26]) were not able to block the activity of extracellular S100A4. S100 proteins share a high degree of sequence homology; therefore it was of interest to investigate whether other S100 proteins were able to sensitize tumor cells to IFN-γ-mediated cell death. Addition of 20 μg/ml recombinant S100A10 or S100A13 had no effect on apoptosis induction by IFN-γ in II-11b cells. Thus, it seems unlikely that the observed effects are a general feature of S100 proteins, but whether other S100 proteins than those tested here could possess IFN-γ-sensitizing properties requires further investigation. The possibility exists that rS100A4 and IFN-γ induce a different type of cell death in II-11b cells than the IFN-γ-mediated apoptosis in OHS cells. DNA fragmentation (Fig. 5A) as well as PARP and Lamin B cleavage (Fig. 5A) were demonstrated also in II-11b cells, suggesting that addition of recombinant S100A4 actually mimics the biological effects of endogenous S100A4 in these experiments.

Caspase-mediated apoptosis is the most important program of cell death. In order to investigate whether IFN-γ activated the caspase cascade in our cell system, we have followed two different approaches: detection of activated caspases by Western blotting and treatment of cell cultures with caspase inhibitors. Using specific antibodies detecting both the proenzyme and the active form of caspase-1, -3, -8, -9 and -12, we were not able to detect any activation of caspase-1, -3, -8 or -12 (Fig. 6 and data not shown). However, the caspase-9 antibody detected a weak, but distinct, ~37 kDa band in IFN-γ treated cells at 48 and 72 h and also in control cells at 72 h. This probably corresponds to one of the active forms of caspase-9. Additionally, we observed a lower level of expression of the proenzyme of caspase-3 in IFN-γ treated cells at 48 and 72 h (Fig. 6). This could possibly indicate a suppression of caspase-3 expression by IFN-γ or activation of caspase-3 with resulting cleavage of the proenzyme. On the other hand, no active forms of caspase-3 were detected, even when exposing the film for several hours. To ensure that the antibodies were able to detect the mature forms, immunotoxin-treated breast cancer cells were included as a positive control [27]. Furthermore, the pan-caspase inhibitor zVAD-fmk inhibited IFN-γ-induced cell death, albeit only at high concentrations (Fig. 7). zVAD-fmk suppressed IFN-γ-mediated apoptosis significantly at 125 μM, whereas 50 μM inhibited cell death non-significantly, and 10 μM zVAD-fmk had no effect on apoptosis inhibition. The latter concentration has been reported to be sufficient to inhibit most caspases [28], therefore it seems likely that the observed effect of zVAD-fmk could be due to inhibition of other proteases or signaling events than caspases. In addition, inhibition of caspase-1 (zYVAD-fmk), caspase-2 (zVDVAD-fmk), caspase-3 and -7 (zDEVD-fmk), caspase-6 (zVEID-fmk), caspase-8 (zIETD-fmk) and caspase-9 (zLEHD-fmk) had no effect on IFN-γ-induced cell death at concentrations up to 50 μM. Taken together, we have not observed any activation of caspases-1, -3, -8 or -12, whereas a low-level activation of caspase-9 was detected. Inhibitors against a range of caspases including caspase-9 did not suppress induction of cell death; hence, from these experiments we concluded that the cell death program induced by IFN-γ in OHS cells most likely is caspase-independent.

The caspase inhibitors used in the experiments above were solved in dimethyl sulfoxide (DMSO), and therefore DMSO was included as a negative control. Surprisingly, a significant inhibition of IFN-γ-induced apoptosis was observed using 0.25 % DMSO (Fig. 7). One of the known biological properties of DMSO is scavenging of hydroxyl radicals, indicating that the generation of reactive oxygen species (ROS) could be implicated in the observed induction of apoptosis. Furthermore, treatment with the commonly used antioxidant N-acetylcysteine significantly suppressed the IFN-γ-induced cell death (Fig. 7), whereas the nitric oxide synthase inhibitor L-NMMA had no effect (data not shown). Notably, zVAD-fmk was recently shown to possess potent antioxidant effects at high concentrations [29] possibly explaining its inhibitory effects shown above.

The pan-caspase inhibitor zVAD-fmk has previously been shown to bind the cysteine protease cathepsin B and to suppress cathepsin B activity in vitro [30]. Since cathepsins are able to act as mediators of programmed cell death [17], zVAD-fmk could possibly inhibit IFN-γ-induced apoptosis through inhibition of cathepsin B. However, 50–250 μM E64 (a cysteine protease inhibitor) or 1–10 μM zFA-fmk (a cathepsin B inhibitor) failed to suppress induction of the observed cell death (data not shown), arguing against cathepsin B as a mediator of IFN-γ-induced apoptosis. On the other hand, high concentrations (125 μM) of zFA-fmk completely inhibited induction of apoptosis (Fig. 7). Cathepsin B activity is most likely suppressed at concentrations below 10 μM [28]; thus, we believe that the observed inhibition by zFA-fmk is due to non-specific effects. Indeed, zFA-fmk was recently shown to suppress transactivation by NF-κB at concentrations similar to those used in these experiments [30]. Interestingly, addition of rS100A4 to II-11b cells lead to a marked induction of NF-κB activity in transient transfection assays using an NF-κB reporter construct, and this induction was blocked by zFA-fmk (Fig. 8A). Furthermore, IκBα was phosphorylated upon addition of extracellular S100A4 (Fig. 8B), confirming activation of the NF-κB pathway. Taken together, these results demonstrate that extracellular S100A4 activates NF-κB, and suggest that the apoptosis induced by S100A4 and IFN-γ could be mediated through an NF-κB-dependent pathway, but whether these two events are actually causally connected remains to be investigated.