Inhibition of Epidermal Growth Factor-induced Interleukin-1β-converting Enzyme Expression Reduces Proliferation in the Pancreatic Carcinoma Cell Line AsPC-1¹ Free

ICE³ was first characterized by its function in processing the cytokine IL-1β, known to play a key role in inflammation and other pathophysiological processes (1, 2). ICE (also named caspase-1) is a member of cysteine proteases that cleave their substrates after aspartic acid and are all synthesized as pro-enzymes that are activated by autocleavage or cleavage by other members of ICE-family proteases. The active human ICE protein, generated by proteolytic cleavage of its precursor (p45), consists of two subunits, p20 and p10. Although the active site Cys285 is located in the p20 subunit, both p20 and p10 are essential for activity (3). The crystal structure of ICE indicate that the catalytically active form is a tetramer consisting of a (p20-p10)₂ homodimer (4, 5).

Another function of ICE is discussed in the regulation of programmed cell death (apoptosis; 6). It has been reported that overexpression of ICE in rat fibroblasts leads to apoptosis (7) and that coexpression of the crmA protein, a potent inhibitor of ICE, blocks ICE-induced apoptosis (8). Studies by Yuan et al. (9) on the developmental cell death in the nematode worm Caenorhabditis elegans have shown that the ced-3 gene plays a crucial role in executing cell death. The cloning of the ced-3 gene has revealed a large sequence homology to mammalian ICE, providing evidence of evolutionary conservation of the apoptotic pathways (9). However, in recent studies, we could demonstrate that ICE is overexpressed in the adenocarcinomas of the pancreas and correlated significantly with the overexpression of EGF-receptor, EGF, and cyclin D1, pointing to possible aspects of ICE in proliferative processes in human pancreatic carcinoma cells (10). To assess the function of ICE in pancreatic cancer we investigated ICE expression in the human pancreatic carcinoma cell line AsPC-1 after stimulation with the mitogen EGF.

AsPC-1 cells were grown in monolayers at 37°C in DMEM supplemented with 10% FCS, penicillin/streptomycin, and glutamine (Biochrom, Berlin, Germany). Cells were seeded at a density of about 1.5 × 10⁶ cells in 100-mm (diameter) plates, starved for 24–48 h, and then incubated with or without EGF (25 ng/ml in FCS-free DMEM) for different times. To investigate functional properties of ICE, specific cell-permeable inhibitors of ICE—the tetrapeptide Ac-YVAD.cmk and the polypeptide Ac-YVAD.CHO (Calbiochem)—were used. Cells were incubated with EGF (25 ng/ml) and Ac-YVAD.cmk (100 μm, in DMSO) or Ac-YVAD.CHO (25 μm in DMSO) or DMSO alone in DMEM containing 0.5% FCS. After incubation, cells were collected by mechanical scraping and used for Western blot or cell cycle analyses. All of the results were validated in at least three independent experiments.

mRNA was isolated using a guanidinium thiocyanate method and oligo(dT)-cellulose column chromatography (QuickPrep, Micro mRNA Purification kit; Pharmacia Biotech). The mRNA was dissolved in the elution buffer (30 μl) provided in the kit. cDNA was prepared by reverse transcription of mRNA using SuperScript RT RNase-H-reverse transcriptase (Life Technologies Inc.). A 399-bp ICE fragment was amplified using the following primers: 5′-GGAAATTACCTTAATATGCAAGAC-3′ (sense) and 5′-CATGAACACCAGGA-ACGTGCTGTC-3′ (antisense). A 5-μl aliquot of cDNA was subjected to 40 amplification cycles as follows: at 94°C for 60 s, 55°C for 30 s, and 72°C for 120 s. The PCR products were electrophoresed through a 1% agarose gel containing ethidium bromide and visualized by UV light. β-actin was used as an internal standard to confirm equal loading in each experiment.

Cells were lysed in 10 mm Tris (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1% Triton X-100, and 1 μg (each) of aprotinin, leupeptin, and pepstatin/ml. Ten μg of protein from each sample were loaded on a denaturing 15% polyacrylamid gel. After electrophoresis, the resolved proteins were transferred onto a nitrocellulose membrane (Schleicher & Schuell) using a semidry transblot apparatus (Phase; Lübeck, Germany). Nonspecific protein interactions were blocked by the preincubation of the membranes with 10% dry milk in PBS overnight at 4°C. After the incubation of the membranes with antibodies, specific binding was detected using the enhanced chemiluminescence system (Amersham). The monocytic cell line THP.1 served as a positive control for ICE expression. To exclude the detection of nonspecific protein bands, a Western blot analysis of the same cell lysates was performed by skipping incubation with the first antibody. Equal loading was confirmed by actin Western blot analysis. The rabbit polyclonal sera of anti-ICE(p10) (C-20), anti-bcl-2 (ΔC21), and the goat polyclonal serum of anti-lamin A/C (N-18) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA) and anti-actin (Ser-Gly-Pro-Ser-Ile-Val-His-Arg-Lys-Cys-Phe) from Sigma. Anti-IL-1α and anti-IL-1β polyclonal sera were obtained from PeproTechEC (London, England).

For cell cycle analysis, cell nuclei were stained with propidium iodide using the Cell Cycle Plus (Becton & Dickinson). Analyses were performed by flow cytometry carried out on a FACScan using the Cell Quest program (Becton & Dickinson). For the determination of cell size, trypsinated unstained cells were used. Cell cycle experiments were performed in a total of 5 independent experiments.

To detect morphological changes in the nuclear chromatin of cells undergoing apoptosis, cells were stained with the DNA-binding fluorochrome bis-benzimide (Hoechst 33258; Sigma). Cells were fixed with 3.7% paraformaldehyde in PBS for 15 min and washed twice with PBS. Staining of the nuclei was performed by incubation of the cells with bis-benzimide (15 μg/ml) for 15 min followed by washing with PBS. A 5-μl aliquot of cells was placed on a glass slide, and the average number of nuclei per field was scored for the incidence of condensed chromatin fragments in apoptotic cells under a fluorescence microscope.

Results are expressed as mean values ± SD. For statistical analysis, Student’s paired t test was used. Significance was defined as P < 0.05.

The incubation of AsPC-1 cells with EGF (25 ng/ml) for 5, 7.5, 10, 12.5, 16, and 24 h caused a time-dependent increase of ICE expression. The expression of ICE mRNA was determined by RT-PCR amplifying a 399-bp ICE fragment. The maximum mRNA expression was observed after 16 h of EGF incubation (Fig. 1 A).

The expression of the ICE protein was detected in Western blot analysis (Fig. 1 B) as a protein band migrating at M_r 45,000 corresponding to the molecular weight of the inactive precursor form of ICE. The time course showed a slight increase in ICE expression at 5 h and 7.5 h, and a maximum of EGF-induced ICE expression was observed after approximately 16 h of incubation. At that time point, the proteolytically active subunits p20 and p10 of ICE were also detectable. Twenty-four h after EGF incubation the level of active ICE decreased. A protein band with M_r of approximately 30,000, described as an active form of ICE (11), also seemed to be time-dependent, and, at higher concentrations, this signal was present as a double band.

The activity of ICE was also confirmed by proteolysis of lamin, which is known to be a substrate of ICE (12). Lamin was processed in the same time-dependent manner as the expression of ICE. Sixteen to twenty-four h after EGF incubation, fragments of lamin with M_r 47,000 and 37,000 were detected, whereas the amount of the uncleaved lamin decreased (Fig. 2). Western blotting using antibodies directed against IL-1α or IL-1β revealed increases in the M_r 15,000 and 17,000 forms of active IL-1α and IL-1β, respectively, after EGF stimulation (Fig. 2).

To examine whether expression of ICE induces apoptosis in the AsPC-1 cell line, we performed cell cycle analysis of cells incubated with EGF (25 ng/ml) for 24 and 72 h in a total of five experiments. The analyses revealed that 6.8 ± 1.9% and 13.2 ± 1.8% of the stimulated AsPC-1 cells became apoptotic after 24 h and 72 h of incubation, respectively. In unstimulated cells, apoptosis was found in 7.2 ± 1.9% and 14.6 ± 1.3% of the cells after 24 and 72 h, respectively, demonstrating no significant difference in the stimulated cells.

Using bis-benzimide staining for detection of condensed nuclear chromatin indicating apoptosis, we observed no difference between stimulated or unstimulated cells (data not shown). These data coincide with the observation that incubation with EGF induced a time-dependent expression of the anti-apoptotic molecule bcl-2 (Fig. 2). Similar to the time-dependence of ICE expression, maximum expression levels of bcl-2 in AsPC-1 were detected after 16–24 h after EGF stimulation.

Proliferation of AsPC-1 cells was significantly increased after incubation with EGF for different times between 0 h and 48 h compared with unstimulated cells. Cell cycle analysis revealed that 16–18 h after stimulation, cells began to transit into the S phase. Maximum of proliferation was observed after 21–30 h, and cell division was completed after 36–48 h (Fig. 3,A). EGF-stimulated AsPC-1 cells showed a significant increased proliferation by the factor 1.5 (±0.3) compared with the control. Interestingly, the maximum of ICE expression and activity (Figs. 1and 2) was observed after 16 h of incubation coinciding with the beginning of the transition into the S phase.

As demonstrated above, stimulation with EGF induced cell proliferation and revealed an increase of ICE expression without leading to apoptosis in AsPC-1 cells. To examine whether EGF-induced ICE expression itself has an influence on cell proliferation, we performed cell cycle analysis in the presence and in the absence of the specific ICE inhibitor, Ac-YVAD.cmk. AsPC-1 cells were incubated with EGF (25 ng/ml) and Ac-YVAD.cmk (100 μm) or with EGF and DMSO alone for 24 h. In the presence of the ICE-inhibitor, the S-phase fraction was significantly reduced by 10.4% (±2.8) as compared with the proliferation of cells incubated with EGF and DMSO alone (Fig. 3 B; P < 0.01, t test). In the presence of Ac-YVAD.cmk, a slight increase of apoptosis (8.2% ± 3.6) was also observed in the EGF-stimulated cells compared with stimulated cells without inhibitor (6.8% ± 1.9). However, this difference did not reach statistical significance. To verify these results, the polypeptide Ac-YVAD.CHO, another cell-permeable specific inhibitor of ICE, was used. In the presence of Ac-YVAD.CHO (25 μm), the proliferation of AsPC-1 cells was also significantly reduced (15.6% ± 8.8; P < 0.01).

Another effect of Ac-YVAD.cmk was observed regarding cell size. In the presence of the ICE-inhibitor, an increase of cell size could be observed (10.4 ± 3%), detected as an increase of intensity of the FSC light measured in the flow cytometer (Fig. 3 C).

We investigated the expression of ICE in the pancreatic carcinoma cell line AsPC-1 after stimulation with EGF. Incubation of these cells with EGF revealed a time-dependent increase of the expression of active ICE. PCR analysis demonstrated the induction of ICE mRNA-synthesis after EGF exposure, which suggests that the increase of the ICE protein expression is not due to protein stabilization. The molecular weight of the detected proteins shown by Western blot corresponded to that of the bands revealed in our positive control cell line THP.1, which confirmed the reports of other investigators (3, 11). Protein bands migrating at M_r 20,000 and 10,000—indicative for the expression of active ICE—appeared 16–24 h after EGF stimulation (Fig. 1,B). A protein band with M_r of approximately 30,000, which was also detectable in a similar time course and is described as an active form (11), corresponds to that shown in human pancreatic carcinoma tissues as recently published (10). Activity of the detected ICE after EGF stimulation was also confirmed by the comparable time-dependent proteolytical cleavage of lamin A/C (Fig. 2), one of the substrates of ICE (12). Furthermore, ICE activity was demonstrated by the processing of the IL-1β molecule showing maximal active IL-1β levels at 16 h after stimulation.

In view of the fact that IL-1α was also processed, our data provide additional evidence for a functional activation of ICE because ICE is known to be involved in the release of the M_r 15,000 form of this molecule (13).

In many systems, ICE was reported to be involved in the process of the programmed cell death (8, 14). Interestingly, the expression of ICE in our EGF-stimulated pancreatic carcinoma cells did not lead to apoptosis. Chin et al. (15) recently reported that EGF stimulation activated STAT proteins causing the expression of ICE and inducing apoptosis in A431 cells but not in HeLa cells, whereas the EGF receptor autophosphorylation and mitogen-activated protein kinase activation were similar in both of the cell lines. The finding that EGFs do not stimulate but inhibit cell growth and, furthermore, induce apoptosis has also been observed in some other mammalian cells (16, 17). However, in our system, the activation of ICE after EGF stimulation did not induce programmed cell death but even permitted proliferation. Furthermore, the inhibition of ICE by Ac-YVAD.cmk and Ac-YVAD.CHO inhibited significantly the transition into the S phase in the stimulated AsPC-1 cells, which suggested that ICE plays an active role in proliferation in our examined pancreatic carcinoma cells. The apoptotic potential of the ICE molecule may be neutralized by the up-regulation of the antiapoptotic bcl-2 protein which is known to have suppressive influence on ICE-induced apoptosis (18, 19). An anti-apoptotic function of ICE was also suggested in human neutrophils and in acute myelogenous leukemia, in which ICE has also been reported to inhibit apoptosis (20, 21).

Recently, we have shown immunohistochemically that ICE is also expressed in human pancreatic carcinoma tissue (10). There, all the ductal cells, supposed to proliferate in the tumor surrounding chronic pancreatitis tissue, showed a predominantly unclear staining in these areas. A nuclear expression of ICE has also been reported in human neuroblastoma cells (22). Rao et al. (12) showed that lamin A/C is proteolysed by ICE during apoptosis, facilitating nuclear events. In our studies, we observed beginning degradation of lamin 16 h after EGF stimulation, just before the cells transit into the S phase. It may be assumed that lamin degradation also facilitates nuclear processes in proliferation and that the nuclear lamina needs controlled proteolysis in the completion of cell division. The finding that the cell size of AsPC-1 cells increases under the influence of the ICE inhibitor Ac-YVAD.cmk may be attributed to this hypothesis.

In conclusion, we have shown an up-regulation of active ICE in the EGF-induced proliferation of the pancreatic carcinoma cell line AsPC-1. Because we did not observe apoptosis, it may be suggested that the decision as to whether ICE induces proliferation or apoptosis could be dependent on (a) the cell compartment in which it accumulates in an active form, (b) the concentration of the active ICE, or (c) the regulation of anti-apoptotic genes such as bcl-2.

We thank Alexandra Kröner and Thea Hamma for expert technical assistance.