SCCA2 inhibits TNF-mediated apoptosis in transfected HeLa cells. 

The reactive centre loop sequence is essential for this function and TNF-induced cathepsin G is a candidate target

Authors

  • Anne F. McGettrick,

    1. Department of Biochemistry and Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland
    Search for more papers by this author
  • Ruth C. Barnes,

    1. Department of Biochemistry and Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland
    Search for more papers by this author
  • D. Margaret Worrall

    1. Department of Biochemistry and Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland
    Search for more papers by this author

  • Enzymes: caspase-3 (EC 3.4.22.36); cathepsin G (EC 3.4.21.20)

M. Worrall, Department of Biochemistry, University College Dublin, Belfield, Dublin 4, Ireland. Fax: + 353 1 2837211, Tel.: + 353 1 7061533, E-mail: mworrall@ucd.i.e

Abstract

The squamous cell carcinoma antigens, SCCA1 and SCCA2, are members of the serine protease inhibitors (serpin) superfamily and are transcribed by two tandomly arrayed genes. A number of serpins are known to inhibit apoptosis in mammalian cells. In this study we demonstrate the ability of SCCA2 to inhibit tumor necrosis factor-alpha (TNFα)-induced apoptosis. HeLa cells stably transfected with SCCA2 cDNA had increased percentage cell survival and reduced DNA fragmentation. We investigated if the reactive centre loop (RCL) was necessary to allow SCCA2 to inhibit TNFα-mediated apoptosis. The RCL amino acids (E353Q, L354G, S355A), flanking the predicted cleavage site, were mutated and the resulting SCCA2 lost both the ability to inhibit cathepsin G and to protect stably transfected cells from TNFα-induced apoptosis. The presence of SCCA2 caused a decrease in the activation of caspase-3 upon induction with TNFα but no direct inhibition of caspases by SCCA2 has been found. Expression of cathepsin G was found to be induced in HeLa cells following treatment with TNFα. This protease has recently been shown to have a role in apoptosis through cleavage of substrates, so maybe the relevant target for SCCA2 in this system.

Abbreviations
SCCA1

squamous cell carcinoma antigen 1

SCCA2

squamous cell carcinoma antigen 2

TNF

tumour necrosis factor-alpha

RCL

reactive centre loop

PI-9

proteinase inhibitor 9

PAI 2

plasminogen activator inhibitor 2

crmA

cytokine response modifier A

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide

succ-AAPF-pNA

succinyl-Ala-Ala-Pro-Phe-p-nitroanilide

Ac-DEVD-amc

acetyl-Asp-Glu-Val-Asp7-amido-4-methylcoumarin

Squamous cell carcinoma antigen (SCCA), formerly referred to as TA-4, is a tumour-associated protein isolated from human uterine cervical squamous cell carcinoma (SCC) tissue [1]. It is used as a tumour marker for SCC of the cervix and various other squamous-cell carcinomas [2]. Comparative amino-acid analysis revealed SCCA as a member of the serine protease inhibitor (serpin) family and it has been localized to chromosome 18q21.3 where it is clustered with other ovalbumin-type serpins [3]. It was initially thought that SCCA existed in two isoforms, a neutral and acidic isoform [4], but molecular studies revealed that two genes exist, SCCA1 and SCCA2/leupin[3,5]. Tissue distribution studies have shown that SCCA1 and SCCA2 are coexpressed in the suprabasal layers of the stratified squamous epithelium of the tongue, esophagus, uterine cervix, vagina, tonsil, Hassall's corpuscles of the thymus, and some areas of the skin [6]. SCCA1 and SCCA2 are not actively secreted and reside exclusively in the cytosol of squamous carcinoma cells [7]. SCCA2 appears to be at a lower level in normal tissue compared to SCCA1 but is increased more significantly in cancer tissues [8].

Serpins neutralize their target serine proteases by undergoing a conformational change, after enzymatic cleavage of the RCL, trapping the enzyme in a covalent intermediate [9]. The amino-acid sequence of the RCL plays a critical role in the ability of a serpin to inhibit a particular protease [10]. SCCA2 has 91.8% sequence identity to SCCA1 but the two proteins differ mainly in their reactive centre loop (RCL) sequence [5]. SCCA2 inhibits the serine proteases cathepsin G and mast cell chymase [11] while SCCA1 inhibits the papain-like lysosomal cysteine proteases cathepsin S, L and K [12]. Therefore, SCCA1 may limit injury from lysosomal proteases released from damaged epithelial cells.

Suminami et al. discovered that when the human head and neck squamous cell carcinoma PCI-51 cells, which do not normally express SCCA1, were transfected with SCCA1 cDNA the levels of apoptosis, induced by 7-ethyl-10-hydroxycamptothecin, tumor necrosis factor-alpha (TNFα) or interleukin (IL)-2-activated natural killer (NK) cells, decreased. Furthermore, SKGIIIa cells, which normally express SCCA1, were transfected with antisense SCCA1 cDNA and this significantly increased the susceptibility of these cells to drug-induced apoptosis. SCCA1 appears to interfere with the apoptosis pathway upstream of caspase-3 as SCCA1 expressing PCI-51 clones had decreased upregulation of caspase-3 upon induction with TNFα[13]. However, no RCL mutations were performed.

Other serpins also inhibit certain apoptotic pathways. The viral serpin crmA, when overexpressed in mammalian cells, protects the cells against apoptosis by inhibiting certain caspases [14]. Two ov-serpins, PI-9 and PAI-2, have been found to protect cells against cytotoxic T-cell and TNFα-mediated apoptosis, respectively [15,16]. In the case of the ov-serpin PAI-2, mutating the P1-Arg to an Ala caused PAI-2 to lose its ability to protect cells against TNFα-induced apoptosis [16] and the C–D interhelical region has also been found to be essential for this function [17]. We have previously reported preliminary findings that SCCA2 can protect cells from TNF-mediated apoptosis [18] and protection from radiation-induced apoptosis has also been reported [19].

In this study we demonstrate that HeLa cells, stably transfected with SCCA2, are more resistant to TNF-mediated apoptosis than untranfected or antisense controls, and that caspase-3 activity is decreased. Furthermore, SCCA2 mutated in the RCL cleavage site failed to protect, suggesting that the serpin inhibitory function is required for the anti-apoptotic mechanism.

Materials and methods

Materials

The Escherichia coli strain BL21 (DE3) was obtained from Stratagene (La Jolla, CA, USA). The bacterial cloning vector PCR2.1 and expression vector pRSETC were supplied by Invitrogen. Q-Sepharose, iminodiacetic acid (metal affinity matrix), cathepsin G, caspase-3 and their substrates succinyl-Ala-Ala-Pro-Phe-p-nitroanilide (succ-AAPF-pNA) and acetyl-Asp-Glu-Val-Asp7-amido-4-methylcoumarin (Ac-DEVD-amc), respectively, geneticin sulfate (G418), cycloheximide, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT), 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium and the horseradish peroxidase-conjugated anti-(rabbit IgG) Ig were purchased from Sigma. All oligonucleotide primers were obtained from Genosys (Cambridge, UK). Taq polymerase and MMLV reverse transcriptase were purchased from Promega. All restriction enzymes were purchased from New England Biolabs. Dulbecco's modified Eagle medium (DMEM) and fetal bovine serum were supplied by Shaw Scientific Ltd. Glutamine, penicillin and streptomycin were purchased from GibcoBRL. The IMx tumour marker kit was purchased from Abbotts Laboratories. Cytobuster™ protein extraction reagent was purchased from Novagen. BM chemiluminescence blotting substrate (POD) was purchased from Roche Molecular Biochemicals. The rabbit anti-(cathepsin G) Ig was purchased from Calbiochem.

Construction of the mutants

Mutants were generated using site directed mutagenesis. The plasmid template used was pRSETC/SCCA2, a His6-tagged fusion protein vector with the complete coding sequence of SCCA2[8]. The mutants were constructed by PCR using the following primers: 5′-GTAGTAGTAGTCCAAGGAGCATCTCCTTCA-3′ (5′ primer corresponding to nucleotides 1045–1074, but incorporating mutations at the P2, P1 and P1′ amino acids); and 5′-CCCCCGGGTACCTACGGGGATGAGAATCTG-3′ (3′ primer corresponding to the nucleotides 1154–1170 of SCCA2 incorporating a Kpn1 restriction site). PCR was performed with 30 cycles of 94 °C × 1 min, 55 °C × 1 min and 72 °C × 2 min. This mutated the P2-Glu353, P1-Leu354, P1′-Ser355 residues of SCCA2 to P2-Gln353, P1-Gly354, P1′-Ala355 and was verified by DNA sequencing. This 125 bp product was then used as a 3′ primer along with the primer 5′-GGGGGATCCATATGAATTCACTCAGTGAAG-3′ (5′ primer corresponding to nucleotides 1–19 of SCCA2 incorporating a BamH1 restriction site) to amplify the full length SCCA2 mutant (mSCCA2) cDNA. The cDNA was ligated to the BamH1 and Kpn1 cloning sites in the pRSETC bacterial expression vector.

Construction of the mammalian expression vector

The coding region of SCCA2 and mSCCA2 were amplified by PCR, from pRSETC/SCCA2 and pRESTC/mSCCA2, respectively, using the following primers: 5′-GGGGGATCCATATGAATTCACTCAGTGAAG-3′ (5′ primer corresponding to nucleotides 1–19 of SCCA2 incorporating a BamH1 restriction site); and 5′-CCCCCGGGTACCTACGGGGATGAGAATCTG-3′ (3′ primer corresponding to the nucleotides 1154–1170 of SCCA2 incorporating a Kpn1 restriction site). PCR was performed with 30 cycles of 94 °C × 1 min, 55 °C × 1 min and 72 °C × 2 min. The amplified fragment was ligated in PCR2.1. SCCA2/PCR2.1 and mSCCA2/PCR2.1 were digested with Kpn1 and ligated to the Kpn1 cloning site in the pcDNA3 mammalian expression vector. The orientation of the gene was examined using BamH1 and sense clones of SCCA2 and mSCCA2 and antisense clones of SCCA2 were selected.

Expression and purification of recombinant SCCA2 and mSCCA2 in E. coli

Recombinant SCCA2 and mSCCA2 were overexpressed from pRSETC in the E. coli strain BL21 (DE3), producing a recombinant fusion protein with a polyhistidine binding tail at the N-terminus. Cells were diluted to A600 = 0.3 and left to grow at 37 °C for 16 h. The cells were harvested by centrifugation (10 000 g × 10min) and resuspended in 15 mL sonication buffer (50 mm Tris/Cl, pH 8.0, 5 mm 2-mercaptoethanol). Following sonication the cell lysate was centrifuged (10 000 g × 10min) and the soluble fraction was applied to a Q-Sepharose anion exchange column. The column elution was performed with a gradient of 0–300-mm NaCl in 50 mm Tris/Cl pH 8.0, 5 mm 2-mercaptoethanol. The SCCA2-containing fractions were further purified by immobilized metal ion affinity chromatography using a chelating Sepharose column precharged with 5 mg·mL−1 nickel chloride. The column was equilibrated using 500 mm NaCl, 40 mm sodium phosphate pH 8.0 and column elution was performed stepwise by decreasing pH. SCCA2 remained bound to the column at pH 5.0 and was eluted with 50 mm Tris/Cl pH 8.0, 50 mm EDTA. The metal chelator was then removed by ultrafiltration.

Inhibition assays

Enzyme inhibition of cathepsin G by SCCA2 was determined by the progress curve method as described previously [20]. Under pseudo first order conditions, a constant amount of enzyme, cathepsin G, was mixed with different concentrations of the inhibitor, SCCA2, in NaCl/Pi reaction buffer (0.01 m phosphate buffer, 27 mm KCl, 137 mm NaCl, pH 7.4) at 25 °C. At specific times a sample was removed and residual enzyme activity [E] was determined by adding 20 µm of the substrate, succinyl-Ala-Ala-Pro-Phe-p-nitroanilide (Succ-AAPF-pNA), and measuring the initial rate of hydrolysis spectrophotometrically over time. The rate of enzyme consumption is determined by

inline image

The slope of ln[E]t/ln[E]o vs. time gives –kobs (pseudo first order inhibition rate constant). The first order rate of inhibition was measured at different inhibitor concentrations. The apparent second order inhibition rate constant, ks′, was determined by plotting kobs against the respective inhibitor concentration [I] and measuring the slope

inline image

To test the ability of SCCA2 to inhibit caspase-3 and caspase-7 in vitro, the proteinases were incubated with a molar excess (≈ 3- to 100-fold) of SCCA2 and incubated at 25 °C for 30 min in enzyme buffer (50 mm Hepes, pH 7.4, 100 mm NaCl, 0.1% Chaps, 1 mm EDTA, 10% glycerol, 10 mm dithiothreitol). The residual enzyme activity was determined as described previously [9]. The hydrolysis of the substrate Ac-DEVD-amc over time (velocity) was measured for both caspase-3 and caspase-7. The percentage inhibition = 100 × [1 − (velocity of inhibited enzyme reaction/velocity of the uninhibited enzyme reaction)].

Cell culture and transfection

The HeLa cell line was grown in Dulbecco's modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum, 2 mm glutamine, 100 U·mL−1 penicillin and 100 µg·mL−1 streptomycin (complete medium) in 5% carbon dioxide at 37 °C. Cells were transfected with the pcDNA3/SCCA2 sense and antisense constructs, the pcDNA3/mSCCA2 construct and a control construct of pcDNA3 with no insert, using the calcium phosphate method developed by Graham & Van der Eb [21]. pcDNA3 contains a neomycin resistance gene which, when expressed, allows the cells to grow in the presence of geneticin sulfate (G418). Stably transfected cells were therefore selected using 0.5 mg·mL−1 G418. Expression of the SCCA2 cDNA was examined using RT-PCR and Western blot analysis. Total RNA was isolated from the transfected Hela cell lines using the phenol/guanidium thiocyananate extraction method [22]. RNA (1 µg) was reverse transcribed using MMLV-reverse transcriptase and random hexanucleotide primers [23]. PCR was then performed on the cDNA using the primers 5′-GTAGTAGTAGTCCAAGGAGCATCTCCTTCA-3′ and 5′-CCCCCGGGTACCTACGGGGATGAGAATCTG-3′, which correspond to each end of the SCCA2 open reading frame, to detect the full length transcribed product.

Protein was extracted from transfected and untransfected HeLa cells using Cytobuster™ protein extraction reagent (Novagen) and precipitated using an equal volume of 20% trichloroacetic acid. The protein was incubated on ice for 15 min, washed three times in acetone and resuspended in 10 µL sterile water. These protein samples were run on a 10% SDS gel and transferred to nitrocellulose by semidry blotting. The membrane was blocked in 3% milk powder in Tris-buffered saline, pH 7.5 and incubated with the alkaline-phosphatase-conjugated monoclonal anti-SCCA Ig, derived from the IMx tumour marker kit (Abbott laboratories). The membrane was developed using 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium substrate.

Induction and assessment of apoptosis

Apoptosis was induced by treatment of cells with varying concentrations of tumour necrosis factor (TNFα) in the presence of 10 µg·mL−1 cycloheximide for 6 h. A modification of the tetrazolium dye-based MTT assay [24] was used to determine cell viability. Viable cells take up 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) and convert it to a formazan product that can be detected spectrophotometrically at 590 nm. Cells were seeded at 2 × 104 per well (200 µL) in 96-well plates and grown in complete medium overnight followed by incubation in serum-free medium with or without the test agents for 6 h. The serum-free medium was then removed by aspiration and the cells washed again in NaCl/Pi and incubated for 3 h, in 200 µL per well of DMEM containing 0.45 mg·mL−1 MTT. The cells were lysed with 100 µL per well dimethylsulfoxide and the absorbance was measured at 590 nm. Cell viability was expressed as a percentage of the values obtained for the untransfected cells.

Measurement of DNA fragmentation

A photometric sandwich-enzyme immunoassay kit [Cell Death DetectionPLUS kit (Boehringer Mannheim)] was used that determines the histone-associated mononucleosomal and oligonucleosomal DNA fragments that accumulate in the cytoplasm of apoptosed cells. The cells were seeded and treated with test agents as above and then lysed in 200 µL lysis buffer (provided in the kit). The lysed cells were centrifuged at 200 g for 10 min and 20 µL of the supernatant was removed and placed in a streptavidin-coated microtitre plate (provided with the kit) and used to measure DNA fragmentation as described in the manufacturer's instructions. Two mAbs were used in the assay: a biotin-labelled monoclonal antibody and a peroxidase-labelled monoclonal antibody from mouse. The chromogenic substrate, 2,2′-azino-di(3-ethylbenzthiazolinsulfate), was added. The absorbance of the resulting product was measured at 405 nm. The results were expressed as enrichment factors that are determined by the following formula:

inline image

Caspase-3 assay

Increases in caspase-3 activity were measured in HeLa cells transfected with SCCA2, mSCCA2 and an empty pcDNA3 vector. Cells were seeded at 5 × 105 per well (3 mL) in six-well plates and grown in complete medium overnight. The cells were then incubated with or without TNFα (10 ng·mL−1) in serum-free medium containing cycloheximide (10 µg·mL−1) for 6 h. The cells were harvested and incubated on ice in isolation buffer (25 mm Hepes pH 7.5, 5 mm MgCl2, 1 mm EDTA, 10 µg·mL−1 leupeptin, 10 µg·mL−1 pepstatin, 1 mm phenylmethanesulfonyl fluoride for 20 min. The cells were centrifuged for 20 min at 20 000 g and resuspended in incubation buffer (100 mm Hepes pH 7.5, 10% sucrose, 0.1% Chaps, 10 µg·mL−1 leupeptin, 10 µg·mL−1 pepstatin, 1 mm phenylmethanesulfonyl fluoride and 10 mm dithiothreitol). The protein concentration in each sample was estimated using the Bio-Rad assay.

The solution was incubated for 1 h at 37 °C, with and without the substrate Ac-DEVD-amc, and the fluorescence (excitation = 380 nm, emission = 460 nm) was measured. The ratio of caspase activity was determined by measuring the increase in fluorescence of TNFα(+) and TNFα(–) cells.

Induction of cathepsin G by TNFα in HeLa cells

Untransfected and transfected HeLa cells were treated as in the caspase-3 assay. The protein samples were separated on a 10% SDS gel using electrophoresis and transferred to nitrocellulose by semidry blotting. The membrane was blocked in a 1% blocking solution (Roche Molecular Biochemicals) in phosphate-buffered saline, pH 7.5 and then incubated with an anti-(cathepsin G) Ig. A secondary horseradish peroxidase-conjugated anti-(rabbit IgG) Ig was then applied and the presence of cathepsin G was detected using the BM Chemiluminescence Blotting Substrate (Roche Molecular Biochemicals).

Results

Establishment of stably transfected cells expressing SCCA2

To study the function of SCCA2 in apoptosis, it was first necessary to establish stably transfected tumour cell lines. The carcinoma cell line, HeLa, was transfected with the mammalian expression vector pcDNA3 containing cDNA for SCCA2, antisense SCCA2 and a control vector containing no insert. The expression of SCCA2 in stably transfected cells was confirmed using RT-PCR. The cells containing the SCCA2 sense or antisense cDNA transcribed detectable levels of RNA while the control cells had undetectable levels of SCCA2 RNA. The increase in protein levels was detected using Western blot analysis. SCCA2 is detected in the transfected cells, but no protein is visible in the untransfected HeLa cells, vector alone controls (Fig. 1), or in the antisense control (not shown). No differences in growth rates were observed between untransfected HeLa cells or cells containing SCCA2, antisense SCCA2 or the pcDNA3 vector alone.

Figure 1.

Expression of SCCA2 protein in transfected HeLa cells was confirmed using Western blot analysis. Lane 1, purified recombinant SCCA2; lane 2, molecular mass marker; lane 3, untransfected HeLa cells; lane 4, HeLa cells transfected with the pcDNA3 vector alone; lane 5, Hela cells transfected with wild-type SCCA2 (S1); lane 6, HeLa cells transfected with wild-type SCCA2 (S2); lane 7, HeLa cells transfected with mutated SCCA2 (Mut1); lane 8, HeLa cells transfected with mutated SCCA2 (Mut2).

SCCA2 attenuates apoptosis induced by TNF

When stably transfected cell lines were incubated with 0, 5, 10 and 50 ng·mL−1 TNFα, in the presence of 10 µg·mL−1 cycloheximide, it was found that cells expressing the SCCA2 sense clone had a higher percentage cell survival compared to the untransfected cells (Fig. 2A). Furthermore, cells expressing SCCA2 underwent less DNA fragmentation than the untransfected cells (Fig. 2B), suggesting that less apoptosis is occurring in the SCCA2-expressing clones. The cells containing the antisense SCCA2 clone underwent similar levels of DNA fragmentation and percentage cell survival as the untransfected cells (Fig. 2A,B). This suggests that the expression of SCCA2 protects cells from TNFα-mediated apoptosis. Interestingly, in the absence of cycloheximide, antisense transfectants were more sensitive to TNFα-induced apoptosis than untransfected HeLa cells (data not shown). Although not detected using the RT-PCR protocol in this study, original tissue expression studies on SCCA2/leupin did show some expression in HeLa cells [5]. The small amounts of endogenous SCCA2 expression in HeLa cells, may be blocked by the antisense SCCA2 RNA, reducing the endogenous protective effect of SCCA2 against TNFα-induced apoptosis.

Figure 2.

Cell survival following transfection with SCCA2. (A) MTT assays showed that the percentage survival of the SCCA2 sense (S1 and S2) clones is much greater than that of the antisense (A1 and A2) and untransfected clones. (B) The DNA fragmentation ELISA assay showed that less DNA fragmentation occurred in the sense (S1 and S2) clones than the antisense (A1 and A2) and untransfected clones. Asterisks indicates a statistically significant difference (P < 0.05) (using the student's t-test) between cells transfected with antisense SCCA2 and the cells transfected with sense SCCA2.

Mutation of SCCA2

The amino-acid sequence of the RCL plays an important role in determining whether or not a serpin inhibits a particular protease and the mutation of certain amino acids within this loop causes serpins to lose their inhibitory function [25]. In relation to the Schecter & Berger numbering for the archtypal serpin α-antitrypsin, the predicted P1 amino acid of SCCA2 is Leu354 [9]. Schick et al. demonstrated that the cleavage site for cathepsin G is Leu354–Ser355 [9] but cathepsin S studies with mutated SCCA2 suggest a cleavage site that aligns with SCCA1, thus placing the Glu353 at the P1 position [26]. In view of this work, and the fact that the acidic glutamate may be more likely to have importance in apoptosis, a triple mutant (Glu353→Gln, Leu354→Gly, Ser355→Ala) was generated. The full length mutated cDNA (mSCCA2) was cloned into the bacterial expression vector pRSETC and the protein, with an N-terminal polyhistidine tag, was overexpressed in E. coli. The mutated protein was purified using anion exchange and metal ion affinity chromatography (Fig. 3). We examined if these mutations affected the ability of SCCA2 to inhibit the serine protease, cathepsin G as expected from the studies of Schick et al. [9]. Cathepsin G was incubated with a molar excess of wild-type SCCA2 and mSCCA2 and pseudo first order inhibition reactions were carried out. The apparent second order inhibition rate constant for SCCA2 wild-type is k′ = 1.74 × 105 m−1·s−1. The mutation of P2, P1 and P1′ caused a 25-fold decrease (k′ = 7 × 103 m−1·s−1) in the rate of inhibition of cathepsin G (Table 1).

Figure 3.

Purification of recombinant SCCA2. Lane 1, marker; lane 2, crude bacterial extract; lane 3, Q-sepharose purified material; lane 4, immobilized metal affinity chromatography-purified material.

Table 1. Inhibition of cathepsin G by SCCA2 and mSCCA2. The apparent second order rate constant, ks′, for the inhibition of cathepsin G by SCCA2 was measured. When the P2, P1 and inline image of SCCA2 were altered, the ability of SCCA2 to inhibit cathepsin G dropped 25-fold.
RSL positionP3P2P1P1P2 k s′ (x 105)
Wild-type SCCA2ValGluLeuSerSer1.74
Mutant SCCA2ValGlnGlyAlaSer0.07

Mutated SCCA2 no longer protect cells from TNFα-induced apoptosis

HeLa cells were transfected with the mammalian expression vector pcDNA3 containing the mSCCA2 cDNA and stably transfected cell lines were selected. Again RT-PCR confirmed gene expression and Western blot analysis showed protein expression at similar levels to the wild-type SCCA2 transfectants (Fig. 1). The cells were also treated with cycloheximide and 0, 20 and 50 ng·mL−1 TNFα and DNA fragmentation and cell viability were measured. It was found that the cells expressing mSCCA2 were less viable than cells expressing wild-type SCCA2 (Fig. 4A) with their percentage cell survival similar to that of cells containing the vector alone. Cells expressing the mSCCA2 also underwent more DNA fragmentation than the cells expressing the wild-type SCCA2 (Fig. 4B). This suggests that one, or a combination of two or three, of the P2, P1 and P1′ amino acids of SCCA2 are vital for the serpin to protect cells from apoptosis induced by TNFα.

Figure 4.

Cell survival with wild-type and RCL mutated SCCA2. (A) MTT assays showed that the percentage survival of the SCCA2 wild-type sense (S1 and S2) clones is much greater than that of the SCCA2 mutant sense (Mut1 and Mut2) clones and the control (pcDNA3) clones. (B) The DNA fragmentation ELISA assay showed that less DNA fragmentation occurred in the wild-type (S1 and S2) clones than the mutant (Mut1 and Mut2) clones and control (pcDNA3) clones. Asterisks indicates a statistically significant difference (P < 0.05) (using the student's t-test) between the cells transfected with mSCCA2 and the cells transfected with wild-type SCCA2.

SCCA2 tranfectants show reduced caspase-3 activity

Caspase-3 activity was measured in HeLa cells transfected with wild-type SCCA2, mutant SCCA2 and an empty pcDNA3 vector. The substrate Ac-DEVD-amc is generally used as a caspase-3 substrate [27] but does show activity with caspase-6, -7, -8 and -10. Figure 5 shows that the increase in caspase-3 activity is significantly greater in HeLa cells transfected with the empty pcDNA3 vector than HeLa cells transfected with SCCA2. This suggests that SCCA2 is inhibiting caspase-3 activation. The cells transfected with mSCCA2 showed a similar increase in caspase-3 activity as the control cells (Fig. 5), suggesting that the RCL amino acids are vital for SCCA2 to inhibit caspase-3 activation. The ability of the wild-type recombinant SCCA2 to inhibit caspase-3 and caspase-7 was also examined but no direct inhibition was observed.

Figure 5.

Activation of caspase-3 upon incubation with TNF is reduced in HeLa cells transfected with SCCA2 (S1 and S2) in comparison to cells transfected with the empty pcDNA3 vector or with mSCCA2 (Mut1 and Mut2). Increases in caspase-3 activity were measured as Δfluorescence·h−1·µg protein−1. Asterisks indicate a statistically significant difference (P < 0.05) (using the student's t-test) between the control cells (empty vector) and the cells transfected with wild-type SCCA2.

Cathepsin G is induced by TNFα in HeLa cells

Cathepsin G is normally associated with myloid cells such as neutrophils and macrophages. We looked at its expression in untransfected HeLa cells and HeLa cells transfected with SCCA2 before and after induction with TNFα. We found that levels of this protease are significantly increased following a 6-h induction with TNFα in the presence of cycloheximide (Fig. 6).

Figure 6.

Cathepsin G production in HeLa cells is increased after incubation with TNFα in the presence of cycloheximide. Lane 1, purified cathepsin G; lane 2, uninduced HeLa cells; lane 3, HeLa cells induced with TNFα; lane 4, uninduced SCCA2 transfected cells (S1); lane 5, SCCA2 transfected HeLa cells (S1) induced with TNFα.

Discussion

A number of serpins, when overexpressed in mammalian cells, have demonstrated a protective effect against apoptosis, including SCCA1 that was recently found to inhibit apoptosis induced by various stimuli [13]. In this study we demonstrate that the expression of SCCA2 in HeLa cells confers protection from TNFα-induced cell death in the presence (Fig. 2) and absence (data not shown) of the protein synthesis inhibitor cycloheximide. SCCA2-expressing clones survived significantly longer and showed less DNA fragmentation than untransfected cells and cells containing antisense SCCA2 cDNA. In the absence of cycloheximide cells producing SCCA2 antisense RNA were more sensitive to TNFα than untransfected cells. This could be explained by antisense RNA preventing the small amount of endogenous SCCA2 from being expressed, thus further increasing susceptibility to programmed cell death.

To investigate if the inhibitory function of SCCA2 was necessary for it to protect cells against TNFα-induced apoptosis, a RCL mutant (E353Q, L354G, S355A) was generated. The ability of the SCCA2 mutant protein to inhibit cathepsin G in vitro decreased significantly as expected and expression of the mutant in stably transfected HeLa cells failed to protect the cells against TNFα-induced apoptosis, suggesting that inhibitory function is important. It is possible that the mutated protein may inhibit a different target, but as the mutant transfectants gave similar results to the untransfected cells, it is unlikely that such a ‘gain in function’, would generate a pro-apoptotic effect.

The presence of SCCA2 but not the mutated SCCA2 in HeLa cells significantly reduced the activation of caspase-3 upon induction with TNFα (Fig. 5). However no direct inhibition of caspase-3 or caspase-7 was found with the recombinant purified protein, suggesting that SCCA may inhibit a protease upstream from these caspases.

Evidence is mounting that lysosomal mediators are involved in TNFα-induced apoptosis [28]. It is known that cathepsin B and D are released from lysosomes into the cytosol during the early stages of apoptosis [29,30]. These cathepsins may activate caspases themselves or may promote the release of cytochrome c, which subsequently activates certain caspases [29]. Cathepsin G, which SCCA2 is known to inhibit, has been shown to activate caspase 7 [31]. More recently, cathepsin G has been implicated during apoptosis in NB4 leukaemic cells, through cleavage of the human brm protein [32], which is part of a complex believed to regulate chromatin conformation. Cathepsin G is normally associated with granules but became diffusely distributed during apoptosis. However, this study [32] also reported negligible levels of cathepsin G in HeLa cells. We have confirmed that the untransfected HeLa cells express little cathepsin G, but have found that TNFα induces this protease signficantly within the time course for induction of apoptosis (Fig. 6). Given that cathepsin G can cleave apoptotic substrates [32] it is likely that the mechanism by which SCCA2 protects cells is wholly or partly due to direct inhibition of cathepsin G.

In cervical cancer tissues, levels of SCCA2 expression are elevated more significantly than SCCA1, suggesting that SCCA2 may be a more important antigen in tumour development [6,8]. This increase in SCCA2 expression and subsequent protease inhibition may be an important factor in the resistance of these cells to apoptosis thus leading to tumour growth and progression.

Acknowledgement

This project was supported by the Health Research Board of Ireland.

Ancillary