Introduction

Cytotoxic T cells (CTL) play a key role in the clearance of virally infected and transformed cells.^{1, 2, 3} Killing of target cells occurs primarily by either of two mechanisms, granule exocytosis pathway or the Fas/Fas ligand pathway, leading ultimately to both DNA fragmentation and membrane disintegration.⁴ The main effector molecule of the former pathway is perforin, which is released from CTL upon target cell recognition and induces lysis of the target cell membrane and death of the cell as well as facilitating entry of granzymes into the cell.⁵ Cytotoxic T cell/target cell recognition also leads to DNA fragmentation and appearance of typical apoptotic features.⁶ The final effector mechanisms involved in apoptotic cell death include the activity of a plethora of interleukin 1beta-converting enzyme (ICE)-like proteolytic enzymes or caspases.⁷ Apoptosis induced by CTL has been shown to be mediated by the simultaneous action of perforin and a class of proteolytic enzymes called granzymes (gzm),⁸ one of the better characterized being gzmB, a serine protease that activates caspases.⁹

Neither purified perforin nor purified gzmB alone cause DNA fragmentation and perforin may be required for entry or other processing of the gzmB.^{10, 11, 12} Granzyme B is responsible for early target cell DNA fragmentation, involving processing of intracellular caspases, while gzmA mediates DNA fragmentation at later times after CTL engagement.^{13, 14} Purified perforin or fragmentin, the rat homologue of gzmB, has been shown to cause unscheduled activation of p34^cdc2 kinase and apoptosis¹⁵ implying the importance of active p34^cdc2 kinase in CTL/target killing. Furthermore, normal chromatin condensation occurring in the M phase of the cell cycle is linked to activation of p34^cdc2 activity^{16, 17} and these observations strengthen the notion that apoptosis and normal cell proliferation may have common underlying pathways. However, p34^cdc2 kinase is not involved in steroid induced death of thymocytes or apoptosis induced by certain toxins,^{18, 19} so it is not an absolute requirement for apoptosis. More recently it has been shown that Cdk2 kinase is important for apoptosis induced in thymocytes.²⁰

Fas is a molecule belonging to the TNF superfamily²¹ expressed on the cell surface of a wide variety of cells, especially on those of haematopoietic origin. Engagement of Fas by Fas ligand, expressed on CTL or NK effector cells, results in DNA fragmentation and cell death and constitutes an alternative killing mechanism for CTL, although Fas induced killing is primarily thought to be involved in immune regulation.²² Here we analyse the effect of olomoucine, a specific inhibitor of cyclin dependent kinases, and thapsigargin, which indirectly and irreversibly inhibits p34^cdc2 kinase, on P815 and L1210 target cell lysis and DNA fragmentation by alloreactive CTL to elucidate the role played by p34^cdc2 kinase in target cell killing by either the Fas or the granule exocytosis pathways.

Materials and Methods

Chemicals

Thapsigargin was purchased from Calbiochem (La Jolla, CA, USA) and olomoucine from Sigma Chemical Co. (St Louis, MO, USA). They were made up in stock solutions in DMSO and kept at –20°C. Thapsigargin stock (3 mmol/L) and olomoucine stock (60 mmol/L) solutions were diluted immediately prior to use. Anti-p34^cdc2 kinase and secondary antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Purified anti-Fas antibody (Jo2) and the FITC-conjugated form were obtained from Pharmingen (San Diego, CA, USA).

Target cells

P815 (H-2^d) mastocytoma, the leukaemia cell line L1210 (H-2^d) and L1210 cells stably transfected with mouse-Fas, L1210 Fas,²³ were grown in Eagle's minimum essential medium supplemented with 10% fetal bovine serum (FBS) at 37°C in 5% CO₂.

Animals

BALB/c (B/c; H-2^d), C57BL/6 (B6; H-2^b) and the perforin deficient (perf^–/–),²⁴ gzm B deficient (gzmB^–/–) (kindly supplied by Dr T J Ley, Washington University Medical School, Seattle, WA, USA²⁵), gzm A deficient (gzmA^–/–)²⁶ and perf^–/– X gzmA^–/– X gzmB^–/–27 backcrossed to the B6 background (H-2^b), were bred under pathogen-free conditions at the John Curtin School of Medical Research animal facility. Animals were tested for the correct genotype as described in Simon et al.¹³ Animals were always sex and age matched within individual experiments.

Generation of in vitro alloreactive CTL and in vivo NK cells

To generate alloreactive CTL in vitro 8 10⁷ responder splenocytes from B6 H-2^b or mutant mice (at 2 10⁶/mL) were cultivated with irradiated (2000 rad) stimulator splenocytes (at 1 10⁶/mL) from B/c (H-2^d) for 5–6 days. Cells were used after dead cell separation from live cells by Ficoll centrifugation.

In vivo NK effector cells were splenocytes from mice injected intraperitoneally with 10⁸ plaque forming units of Semliki forest virus (SFV) 2 days previously. Effector cells were phenotyped as described in Müllbacher and King.²⁸

Cytotoxicity assay

Cytotoxicity assays were performed in cell culture medium containing 5% FBS except for calcium depletion experiments when FBS was replaced by 2 mg/mL bovine serum albumin. The ⁵¹Cr release assay was performed as described in Müllbacher et al.²⁹ For estimation of DNA fragmentation target cells were cultured for 12 h at 0.25 10⁶/mL with [³H] thymidine (5 Ci/mL), washed once and chased in cold media for 30–60 min and then washed twice. The DNA release assay was set up as for the ⁵¹Cr release assay in 96-well round-bottomed plates with 25 000 target cells per well in replicates of four. Just prior to harvesting, 20 L of 2% triton X-100 was added to each well and the cells were incubated for a further 10 min. Plates were then spun at 300 g and 20 L supernatent harvested and counted. Percent specific DNA release was calculated as (c.p.m. experimental – c.p.m. spontaneous)/(c.p.m. total – c.p.m. spontaneous) for treated and untreated target cells. Specific lysis was calculated in a similar manner. Spontaneous release over the period of the assay was never more than 5–10% total, including target cells treated with toxins, unless otherwise stated. If it exceeded this the data were discarded. Errors are SD.

Cell lysis by Jo2 was estimated by propidium iodide exclusion. Propidium iodide at 25 g/mL was added directly to cells at 0.1–0.2 10⁶/mL, cells were incubated in the dark at 37°C for 5 min and analysed by flow cytometry. Dead cells showed at least an order of magnitude increase in propidium iodide staining. Target cells were treated with either olomoucine or thapsigargin for 2 h prior to exposure to cytotoxic cells. The toxin was present for the duration of the cytotoxic T-cell assay.

P34^cdc2 kinase activity

Olomoucine has already been shown to inhibit p34^cdc2 kinase dependent kinase activity in intact cells and cell lysates. We needed to confirm that thapsigargin also inhibited this activity in intact cells. The p34^cdc2 kinase activity was estimated as described previously in Waring et al.³⁰ Cells were left untreated or treated for 2 h with thapsigargin at 100 nmol/L in complete medium at 37°C. After pelleting the cells they were lysed in lysis buffer and p34^cdc2 kinase immunoprecipated using a specific antibody. Kinase activity was assessed by phosphorylation of histone H1 in the presence of ³²P ATP. Histone H1 phosphorylation was visualized using autoradiography following polyacrylamide electrophoresis in 17% gels.

Cell cycle analysis and apoptosis

Cells were washed once in PBS and fixed overnight in 75% ethanol, washed twice in PBS and stained with propidium iodide (40 g/mL) and RNase (10 g/mL) for 30 min at room temperature in the dark. DNA profiles were measured using a Beckton Dickinson FacScan (San Diego, CA, USA) and analysed using WinMDI, kindly provided by J Trotter (Salk Institute, La Jolla, CA, USA). Estimation of cell cycle parameters was carried out using ModFit software. Apoptotic cells appeared in the subdiploid region.

Fas expression

Cells were pelleted and washed once in PBS with 2% FBS. Cells were stained on ice with FITC-conjugated anti-Fas, Jo2 at 1:500 dilution (Pharmingen) for 30 min and washed three times in PBS with 2% serum before FACS analysis.

Intracellular calcium

[Ca²⁺]_i was measured using Indo-1 loaded cells and FACS analysis as previously described in Waring and Beaver.³¹

Results

Effects of olomoucine and thapsigargin on the cell cycle and activity of p34^cdc2 kinase

We first established that olomoucine and thapsigargin affected cell cycle progression. Olomoucine is a known reversible inhibitor of p34^cdc2 kinase in intact cells with a reported Km of 7 mol/L³² and acts by competing for the ATP binding site of the kinase. Olomoucine also inhibits a number of other cyclin dependent kinases (CDK), including p33 (cdk20/cyclin A and cyclin E kinases)³² and Cdk2.²⁰ Although not specific for p34^cdc2 its action appears to be restricted to inhibition of CDK. Thapsigargin has been reported to inhibit the p33 (Cdk2) cyclin A dependent kinase.³³ Preliminary experiments showed that P815 cells treated with these agents were arrested in the cell cycle consistent with an inhibition of cyclin dependent kinases. Figure 1 shows the effects of thapsigargin and olomoucine on the cell cycle of P815 cells up to 20 h post-treatment. Thapsigargin had the most profound effect on the cell cycle with rapid arrest of cells in G1 and corresponding decline in S phase cells. G2/M phase cells increased slightly up to 14 h and then declined.

Figure 1.

The effect of 50 mol/L olomoucine (closed symbols) and 100 nmol/L thapsigargin (open symbols) on the cell cycle of P815 cells. (, ), G1; (, ), S; (, ), G2/M. Inset: DNA fragmentation induced by 100 nmol/L thapsigargin.

Full figure and legend (19K)

Olomoucine treated cells also showed arrest in G1, which peaked at approximately 10 h post-treatment, with decline in S phase and little change in G2/M phase cells. The effect was not as pronounced as with thapsigargin-treated cells, although the percentage of cells in G1 was 55% at 10 h compared to approximately 25–30% in untreated cells. There was some entry back into the cell cycle of cells treated with 50 mol/L olomoucine after 20 h as seen by a decline in G1 and an increased proportion of cells in S phase. Higher concentrations of olomoucine (up to 150 mol/L) caused a greater proportion of cells to arrest in the G2/M phase of the cell cycle and induced some apoptosis after cell cycle arrest, but results were qualitatively the same as Figure 1. Thapsigargin showed some apoptotic DNA fragmentation as assessed by a subdiploid population of cells, but this occurs only after 12–15 h treatment (Figure 1 inset), consistent with our earlier work.³⁴ It is important to stress that very little cell death or DNA fragmentation was induced by either olomoucine or thapsigargin over the period of the Tc cell assays. There was little apoptotic DNA fragmentation above background (<10%) with olomoucine at 50–75 mol/L up to 24 h and only a 10% decrease in the viability of the cells over 15 h. Arrest in G1, induced by olomoucine, indicates the importance of p34^cdc2 kinase for G1 to S phase transition or may reflect its effects on other CDK.

Assay of p34^cdc2 kinase activity following treatment of cells showed that thapsigargin at either 100 nmol/L or 500 nmol/L irreversibly inhibits the activity of the kinase in both P815 and L1210 cells (Figure 2). Olomoucine has previously been shown to inhibit purified p34^cdc2 kinase activity³² and cyclin dependent kinases in cell lysates. As this inhibition is reversible and competitive for ATP, immunoprecipitation of the kinase and addition of fresh ATP restores activity as seen in Figure 2. This data is included for completeness. The irreversible inhibition of the p34^cdc2 kinase by thapsigargin is probably due to the inhibitory effect of increased calcium concentration on protein kinase induced phosphorylation of p34^cdc2, which is essential for activity of the latter.

Figure 2.

Activity of p34^cdc2 kinase immunoprecipated from P815 and L1210 cells after 2 h. Lane 1: untreated; lane 2 and 3: 100 nmol/L and 500 nmol/L thapsigargin, respectively; lanes 4 and 5: 150 mol/L and 500 mol/L olomoucine, respectively; lane 6: untreated; lanes 7 and 8100 nmol/L and 500 nmol/L thapsigargin, respectively.

Full figure and legend (6K)

Effect of thapsigargin and olomoucine on target cell DNA fragmentation and ⁵¹Cr release

We first assessed whether either agent had any inhibitory effect on the release of ⁵¹Cr from P815 or L1210 target cells induced by alloreactive CTL from normal B6 mice. Both P815 and L1210 cells normally express only low levels of cell surface Fas. Figure 3 shows the effects of thapsigargin and olomoucine on P815 target cells following either a 2 h or 12 h incubation with inhibitors and an 8 h CTL assay. Two hours was sufficient to irreversibly inhibit the activity of the p34^cdc2 kinase by 100 nmol/L thapsigargin (Figure 2).

Figure 3.

Effect of thapsigargin (E:T) at 100 nmol/L and olomoucine at 150 mol/L on lysis of treated P815 target cells by H-2^d-alloreactive CTL in an 8 h assay. (), untreated; (), 150 mol/L olomoucine for 2 h; (), 100 nmol/L thapsigargin 2 h; (), 150 mol/L olomoucine 12 h; (), 100 nmol/L thapsigargin 12 h.

Full figure and legend (19K)

We also tested the effects of thapsigargin on ⁵¹Cr release from target cells exposed to effector T cells from gzmA^–/–, gzmB^–/– and perf^–/– gzmA^–/– gzmB^–/– triple knock out mice (Table 1). This data shows normal cytolysis in L1210 target cells when treated with Tc cells from gzmA^–/– mice in which the perforin/gzmB pathway still operates. Cells treated with thapsigargin show no difference in their response to Tc cells from gzmA^–/– mice. As Table 1 shows, L1210 cells have little endogenous Fas because lysis by perf^–/– gzmA^–/– gzmB^–/– Tc cells is almost absent. Thus inhibition of p34^cdc2 kinase does not affect killing by the perforin/gzmB pathway. Conversely inhibition of p34^cdc2 kinase also has no effect on Fas killing as seen in Table 1. Both Figure 2 and Table 1 show that after 8 h exposure to CTL there was no inhibitory effect on specific killing as assessed by ⁵¹Cr release under conditions where p34^cdc2 kinase was inactivated.

Table 1 - The effect of 100 nmol/L thapsigargin on specific lysis of target cells induced by alloreactive effector T cells from granzyme A-, granzyme B- and perforin granzyme A granzyme B-deficient mice.

Full table

We then assessed DNA fragmentation in target cells treated with thapsigargin at concentrations known to inhibit p34^cdc2 kinase activity. For this assay we used effector cells from both B6 and perf^–/– mice. The latter should induce DNA fragmentation primarily via the Fas pathway because perforin is required for both granzyme B and A induced early DNA fragmentation.^{3, 13} DNA fragmentation was identical in treated and untreated target cells (Figure 4). Maximal DNA fragmentation was seen after 4 h. Pretreatment of targets with 150 mol/L olomoucine gave similar results with no effect on cell killing (data not shown). Thus neither lysis (⁵¹Cr release) nor DNA fragmentation (³H thymidine release) is affected if p34^cdc2 kinase activity is inhibited in target cells. Furthermore perforin independent and therefore presumably Fas-induced killing by Tc cells does not appear to be affected by inhibition of p34^cdc2 kinase activity.

Figure 4.

Effect of 100 nmol/L thapsigargin (E:T) on specific DNA release from treated P815 target cells by H-2^d alloreactive CTL (B6 and perforin^–/–) over 4 h. (), H-2^d reactive CTL from B6 mice, no target cell treatment; (), H-2^d reactive CTL from perforin^(–/–) mice, no target cell treatment; (), H-2^d reactive CTL from B6 mice, 100 nmol/L thapsigargin for 2 h; (), H-2^d reactive CTL from perforin deficient mice, 100 nmol/L thapsigargin.

Full figure and legend (16K)

In order to explore further the effects of the inhibitors on Fas-induced DNA fragmentation, we examined the effect of thapsigargin treatment on DNA release from L1210 and L1210 Fas-expressing cells. We also used effector cells from normal mice as well as gzmB^–/– knockout mice. As discussed, L1210 cells express low or undetectable levels of Fas, and DNA release at 4 h was low with effector cells derived from either perf^–/– or gzmB^–/– mice because both the granule exocytosis and the Fas pathways are inoperative (Table 2). However, Fas-transfected cells showed significant DNA release when exposed to CTL from either perf^–/– or gzmB^–/– mice, which was little affected by prior treatment with thapsigargin. This confirms that Fas-mediated DNA fragmentation is independent of p34^cdc2 kinase activity. The above data thus shows that the perforin/gzmA or B pathway and the Fas pathway are independent of the activity of p34^cdc2 kinase.

Table 2 - The effect of 100 nmol/L thapsigargin on specific DNA release from L1210 target cells induced by perforin- and granzyme B-deficient alloreactive effector T cells.

Full table

The above experiments were carried out using a 2 h incubation and up to 8 h cytotoxic assays because this is sufficient time to cause maximal killing in untreated targets by normal effector cells, thus allowing any possible inhibitory effects on killing to be detected. However, as already shown, effector cells from perf^–/– mice normally show low levels of lysis in cells that display low Fas levels on their surface, such as P815 or L1210 cells. When we examined the effect of a 12 h olomoucine treatment at 150 mol/L on ⁵¹Cr release from P815 cells by perf^–/– CTL, we unexpectedly found that ⁵¹Cr release was dramatically increased. In panel A of Figure 5, CTL from B6 mice (denoted 'normal') resulted in approximately 60% ⁵¹Cr release of untreated P815 cells, which is completely inhibited by calcium chelation. Calcium chelation with EGTA prevents granule exocytosis thus inhibiting perforin/granzyme induced ⁵¹Cr release, but has little or no effect on Fas-mediated killing.³ In addition P815 cells normally express only low levels of Fas. ⁵¹Cr release of P815 cells treated with 150 mol/L olomoucine changed little (panel A), but now ⁵¹Cr release was not completely abrogated by calcium chelation in the 4 h assay. This effect was more dramatic in an 8 h assay (panel B) where at an effector to target ratio of 9:1, calcium chelation only reduced ⁵¹Cr release in target cells treated with olomoucine from 95% to 70%, whereas calcium chelation again completely abrogated killing of untreated target cells. Panel C shows the same experimental set up, but using CTL derived from perf^–/– mice. No lysis is seen in normal P815 target cells. However, there is an increase in ⁵¹Cr release of up to 20% in olomoucine-treated cells, which is unaffected by calcium chelation. In an 8 h ⁵¹Cr release assay there is significant lysis induced by perf^–/– CTL, which could be inhibited by calcium chelation. This was unexpected and suggests a third pathway for killing, involving a perforin-independent, but calcium-dependent, pathway. Significantly, olomoucine treated P815 target cells showed greatly enhanced ⁵¹Cr release, which again is only partly inhibited by calcium chelation. We did not observe these effects with thapsigargin-treated cells.

Figure 5.

Granule exocytosis independent increase in specific lysis following P815 cell treatment with 150 mol/L olomoucine for 12 h in a 4 h and 8 h assay with H-2^d alloreactive CTL from B6 [normal; (a), 4 h; (b), 8 h] and perforin^(–/–) [ (c), 4 h; (d), 8 h] mice. (), untreated; (), +EGTA; (), +150 mol/L olomoucine; (), +150 mol/L olomoucine + EGTA.

Full figure and legend (24K)

We also examined the effect of prolonged incubation of cells with olomoucine and thapsigargin on DNA fragmentation. While lysis induced by alloreactive CTL from B6 mice was unaffected (Figure 3) and is actually increased when perf^–/– CTL are used as effectors (Figure 5), DNA fragmentation was actually reduced after 12 h incubation with olomoucine and thapsigargin, at which time the cells had significantly slowed in the cell cycle. This was true for effectors from perf^–/– and B6 mice, particularly at the highest effector/target ratio (Table 3). The effect was most obvious with target cells treated with olomoucine and alloreactive CTL from normal mice where there was up to 50% decrease in released DNA.

Table 3 - The effect of 12 h treatment of P815 target cells with thapsigargin and olomoucine on specific DNA release induced by normal and perforin-deficient alloreactive effector T cells.

Full table

Increased Fas expression after treatment with olomoucine

We then sought an explanation for enhanced lysis of P815 cells by perf^–/– CTL after prolonged treatment with olomoucine. When P815 cells were treated with 150 mol/L olomoucine for 12 h we could detect enhanced levels of cell surface Fas expression (Figure 6). The kinetics of Fas appearance were consistent with new protein synthesis (Figure 6; inset) and the presence of 20 mol/L cycloheximide completely abrogates the effect of olomoucine on Fas expression. We found no effect of thapsigargin on either Fas expression or CTL-mediated cell lysis by effector cells from perf^–/– mice. However, thapsigargin is known to inhibit protein synthesis and we determined that 100 nmol/L thapsigargin was sufficient to inhibit protein synthesis in P815 cells by 80% (data not shown). As cycloheximide prevents Fas expression in olomoucine-treated cells, this would account for the lack of Fas expression in thapsigargin-treated cells. In addition, no increased Fas expression was observed in normal L1210 cells after olomoucine treatment. Thus, the effect appears to be cell type specific.

Figure 6.

Increased Fas expression on P815 cells after 12 h treatment with 150 mol/L olomoucine (solid line), with added 10 mol/L cycloheximide (dotted line) and control (light line). Inset shows time course of increased Fas expression. (), 150 m olomoucine; (), untreated.

Full figure and legend (20K)

Fas antibody induced lysis of P815 cells treated with olomoucine: Fas expression is cell cycle dependent

In order to confirm an increased expression of functional Fas on P815 cells we used Jo2, a monoclonal antibody to Fas, which has been shown to induce death in cells with functionally active Fas receptor. We examined the effect of Jo2 on the viability of cells treated with olomoucine for 12 h followed by treatment with 500 ng/mL Jo2. Figure 7 shows that only those cells treated with olomoucine and Jo2 showed significantly increased loss in viability 4–5 h post Jo2 treatment. There was a 10% increase in loss of viability of cells treated with olomoucine over the first 12 h. There was no effect of Jo2 when cells were not exposed to olomoucine. Treatment of P815 cells with olomoucine at 150 mol/L caused maximal Fas expression accompanied by some apoptosis. However, at lower concentrations there was less apoptosis even after 24 h. Thus, P815 cells treated with 75 mol/L olomoucine are arrested in G1 and G2/M at 12 h, but begin to re-enter the cell cycle at 24 h and are fully recovered by 48 h. Fas expression parallels this cell cycle arrest indicating that Fas is only expressed when cyclin-dependent kinases are inhibited (Figure 8). In this figure, apoptotic cells were gated out based on the decreased forward scatter in the DNA histograms (but not the Fas histograms) in order to facilitate cell cycle analysis. However, the levels of apoptotic DNA fragmentation in this experiment were low, being 17% at 12 h, 5% at 24 h and <5% at 48 h, consistent with the observed recovery of the cells. As a majority or possibly all cells expressed Fas at 12 h, a large proportion of the Fas- expressing cells must have subsequently lost Fas when they re-entered the cell cycle and not simply died.

Figure 7.

Lysis of P815 cells treated with 150 mol/L olomoucine for 12 h then 500 ng/mL Jo2. (), untreated; (), Jo2 (anti-Fas) only; (), olomoucine only; (), olomoucine + Jo2 (anti-Fas). Cell viability was assessed using propidium iodide as described in the methods section.

Full figure and legend (14K)

Figure 8.

Cell cycle dependence of Fas expression on P815 cells treated with 75 mol/L olomoucine. At 12 h post-treatment with olomoucine there is maximum Fas expression as shown by a comparison of treated and untreated cells as for Figure 6 (top panel). Cells are arrested in both G1 and G2/M phases of the cell cycle. At 24 h there is a clear reduction in Fas positive cells and some entry into the cell cycle as seen by an increase in S phase. Fas levels decline dramatically at 48 h post-treatment and display a normal DNA profile.

Full figure and legend (36K)

Increased NK killing of P815 treated with olomoucine and DNA damaging agents

One function of NK cells is thought to be immune surveillance in the early recognition of tumour cells,³⁵ although P815 cells normally function as very poor targets for NK cells.^{36, 37} Killing via the Fas pathway has been shown to be important in NK-cell induced killing.³⁸ We therefore examined whether olomoucine treatment affected the killing of P815 cells by NK cells. Figure 9 illustrates the increase in specific lysis induced in P815 cells by NK cells following treatment of the former with olomoucine. Results for YAC-1 cells, a typical NK target cell, are shown for comparison.

Figure 9.

Increased NK dependent killing of P815 cells treated with olomoucine for 16 h. (), untreated P815; (), P815 treated with 50 mol/L olomoucine; (), P815 cells treated with 75 mol/L olomoucine; () YAC cells.

Full figure and legend (19K)

Discussion

p34^cdc2 kinase has an essential role in the S phase to G2/M transition in normal cycling cells, particularly chromatin condensation.³⁹ A number of parallels have been drawn between this process and that of apoptotic cell death in which similar chromatin condensation occurs, but usually accompanied by internucleosomal DNA fragmentation.⁴⁰ Mitotic catastrophe in yeast, for example, due to dysfunctional kinase activity, results in abortive attempts to enter the cell cycle.⁴¹ Premature activation of p34^cdc2 kinase has been reported as an important event in CTL killing mediated by perforin/gzmB.¹⁵ In these studies activity of p34^cdc2 kinase in YAC-1 cells increased up to six-fold after treatment with perforin and gzmB. The same authors showed further evidence for a role of p34^cdc2 kinase in apoptotic DNA fragmentation using a temperature sensitive mutation in p34^cdc2 kinase.

At 100 nmol/L thapsigargin, which mobilizes intracellular calcium by inhibiting the endoplasmic reticulum Ca²⁺ ATPase,⁴² raises intracellular calcium levels in P815 cells to 0.5–1 mol/L within minutes and is sustained for at least 90 mins (data not shown). We have also observed that in P815 cells treated with thapsigargin at 100 nmol/L, there is a general reduction in levels of protein phosphorylation (P Waring, unpubl. data, 1995), which may account for the inhibition of p34^cdc2 kinase because specific phosphorylation of this kinase on threonine residue 161 is required for activity.⁴³ Importantly, there is little lysis or DNA fragmentation in thapsigargin-treated cells up to 16 h until exposed to CTL.

We show here that thapsigargin is a potent irreversible inhibitor of p34^cdc2 kinase at 100 nmol/L, with at least a 10-fold reduction in kinase activity by 2 h post-treatment. Despite this we observed unchanged cell lysis and DNA fragmentation in two different thapsigargin-treated target cells exposed to alloreactive CTL compared to control target cells. We found similar results using the reversible inhibitor of the kinase olomoucine. This insensitivity to p34^cdc2 kinase activity extended to target cells treated with CTL derived from perf^–/–, gzmA^–/–, gzmB^–/– and perf^–/– gzmA^–/– gzmB^–/– KO mice and target cells expressing low or high levels of Fas. Thus, alloreactive killing via the Fas or the perforin/gzmB or perforin/gzmA pathway is independent of p34^cdc2 kinase activity. Although thapsigargin can exert a number of effects on cells due to its influence on intracellular calcium levels and is not a specific inhibitor of p34^cdc2 kinase, the unquestionably irreversible nature of this inhibition and the lack of any other obvious effects over the course of the assays make it a useful tool in this study.

Alloreactive CTL from perf^–/– mice lacking functional perforin resulted in little or no lysis of P815 target cells after 4 h. However, significant levels of DNA fragmentation do occur, as shown in Figure 4. We ascribe this to normal endogenous low levels of cell surface Fas expression on P815 cells or a third killing pathway. When P815 cells were treated for 12 h with olomoucine perf^–/– effector CTL were able to efficiently lyse these target cells and this was preceded by increased expression of cell-surface Fas. Thus, increased lysis by perf^–/– CTL is due to increased levels of cell-surface Fas on P815 cells. Overall killing by alloreactive CTL from normal mice was unaffected as assessed by ⁵¹Cr release, but DNA fragmentation was reduced in treated cells. This is consistent with decreased levels of DNA fragmentation induced by CTL in target cells in a G0 type state.⁴⁴ As this occurs 10–12 h after inhibition of p34^cdc2 kinase, it cannot be attributed to a direct inhibition of the kinase. Lysis in P815 cells at 8 h induced by perf^–/– CTL (Figure 5d) may be indirectly due to, and a consequence of, the early DNA fragmentation (Figure 4). However, raised Fas levels due to olomoucine cause increased lysis over that due indirectly to DNA fragmentation (Figure 5d), suggesting that significantly increased Fas levels may result in ⁵¹Cr release independent of DNA fragmentation. This is consistent with earlier reports that Fas-dependent cell lysis occurs just as effectively in enucleated cells.⁴⁵ Increased Fas-dependent lysis occurs when DNA fragmentation is in fact suppressed at 12 h. Thus, increased cell surface Fas and consequent increased lysis after cell cycle arrest may provide a compensatory mechanism in response to less apoptotic DNA fragmenation due to cells entering G0 and emphasizes the independence of the two processes.

Increased lysis due to Fas expression was confirmed by exposing olomoucine treated P815 cells to the soluble anti-Fas antibody Jo2. This antibody has been shown to kill cells expressing functional Fas on the cell surface, although not all Fas-expressing cells can be lysed by Fas ligand or anti-Fas antibody and lysis can be inhibited by Flice inhibiting protein in the Fas receptor death complex.⁴⁶ Thus, L1210 Fas- transfected cells that express high cell surface Fas are not lysed in our hands with 500 ng/mL Jo2 (data not shown). Olomoucine also failed to affect Fas levels on normal L1210 cells, indicating that this phenomenon is cell-type dependent

It was recently shown that cytotoxic drugs sensitize human cancer cells to Fas ligand induced death⁴⁷ and we have now shown that the cyclin-dependent kinase inhibitor olomoucine also results in increased expression of functional Fas in P815 cells. Furthermore, we have demonstrated that olomoucine also sensitizes NK insensitive P815 cells to NK-induced lysis, indicating that NK cells may also play a part in the in vivo action of this type of compound. Related to this, bleomycin and cisplatin, antitumour chemotherapeutic agents, also upregulate cell surface Fas in P815 cells, which can result in increased NK cell killing (data not shown). Importantly, increased expression of Fas induced by olomoucine in P815 cells is cell cycle dependent and Fas is lost when cells recover CDK activity and re-enter the cell cycle. These results suggests that Fas expression may be under the control of cyclin-dependent kinases as well as the P53 protein as has been suggested,^{48, 49} and may represent an alert signal for removal of cells that have arrested in the cycle for whatever reason. It has also recently been shown that irradiated cells express increased levels of cell surface Fas⁴⁹ and our data supports the suggestion that monoclonal antibodies to Fas may be a useful adjunct in the treatment of tumours with antiproliferative agents, particularly if the tumour cell does not express high levels of functional Fas. The important caveat, however, is that cells that recover from DNA damage and re-enter the cell cycle may subsequently lose cell surface Fas and become susceptible to killing via the Fas pathway.

References

1. Blanden RV. Mechanisms of recovery from a generalized viral infection: mousepox. I. The effects of anti-thymocyte serum. J. Exp. Med. 1970; 132: 1035–54. | Article | PubMed | ISI | ChemPort |
2. Karupiah GR, Buller ML, van Rooijen N, Duarte CJ, Chen JH. Different roles for CD4⁺ and CD8⁺ T lymphocytes and macrophage subsets in the control of a generalized virus infection. J. Virol. 1996; 70: 8301–9. | PubMed | ChemPort |
3. Kägi D, Ledermann B, Bürki K, Zinkernagel RM, Hengartner H. Molecular mechanisms of lymphocyte-mediated cytotoxicity and their role in immunological protection and pathogenesis in vivo. Annu. Rev. Immunol. 1996; 14: 207–32. | Article | PubMed | ISI | ChemPort |
4. Kägi D, Vignaux F, Ledermann BeA. Fas and perforin pathways as major mechanisms of T cell-mediated cytotoxicity. Science 1994; 265: 528–30. | Article | PubMed | ISI | ChemPort |
5. Liu CC, Walsh CM, Young JDE. Perforin: structure and function. Immunol. Today 1995; 16: 194–201. | Article | PubMed | ChemPort |
6. Duke RC, Chervenak R, Cohen JJ. Endogenous endonuclease induced DNA fragmentation: an early event in cell mediated cytolysis. Proc. Natl Acad. Sci. USA 1983; 80: 6361–5. | Article | PubMed | ChemPort |
7. Vaux DL, Strasser A. The molecular biology of apoptosis. Proc. Natl Acad. Sci. USA 1996; 93: 2293–44.
8. Henkart PA. Lymphocyte-mediated cytotoxicity: Two pathways and multiple effector molecules. Immunity 1994; 1: 343–6. | Article | PubMed | ISI | ChemPort |
9. Smyth MJ, Trapani JA. Granzymes: exogenous proteinases that induce target cell apoptosis. Immunol. Today 1995; 16: 292–06.
10. Froelich CJ, Orth K, Turbov J et al. New paradigm for lymphocyte granule-mediated cytotoxicity. Target cells bind and internalize granzyme B, but an endosomolytic agent is necessary for cytosolic delivery and subsequent apoptosis. J. Biol. Chem. 1996; 271: 29 073–9.
11. Trapani JA, Jans DA, Jans PJ, Smyth MJ, Browne KA, Sutton VR. Efficient nuclear targeting of granzyme B and the nuclear consequences of apoptosis induced by granzyme B and perforin are caspase-dependent, but cell death is caspase-independent. J. Biol. Chem. 1998; 273: 27 934–8.
12. Müllbacher A, Tha Hla R, Tran T et al. Granzymes are the essential downstream effector molecules for the control of primary virus infection by cytolytic leukocytes. Proc. Natl Acad. Sci. USA 1999; 96: 13 950–5.
13. Simon MM, Hausmann M, Tran T et al. In vitro- and ex vivo- derived cytolytic leukocytes from granzyme AxB doubleknockout mice are defective in granule-mediated apoptosis but not lysis of target cells. J. Exp. Med. 1997; 186: 1781–6. | Article | PubMed | ISI | ChemPort |
14. Anel A, Gamen S, Alava MA, Schmitte-Verhulst AM, Pineiro A, Naval J. Inhibition of CPP32-like proteases prevents granzyme B and Fas but not granzyme A based cytotoxicity exerted by CTL clones. J. Immunol. 1997; 158: 1999–2006. | PubMed | ChemPort |
15. Shi L, Nishioka WK, Th'Ng JPH, Bradbury EM, Litchfield DW, Greenberg AH. Premature p34cdc2 activation required for apoptosis. Science 1994; 263: 1143–5. | Article | PubMed | ISI | ChemPort |
16. Bradbury E. Reversible histone modification and the cell cycle. Bioessays 1992; 14: 9–16. | Article | PubMed | ISI | ChemPort |
17. Th'Ng JPH, Guo XW, Bradbury EM. Molecular and cellular approaches to the control of proliferation and differentiation. In: Stein GS, Stein JB (eds). Academic Press Inc.: New York, 1992; 381–407.
18. Martin SJ, McGahon AJ, Nishioka WK et al. P34(cdc2) kinase and apoptosis. Science 1995; 269: 106–7. | Article | PubMed | ISI | ChemPort |
19. Norbury C, MacFarlane M, Fearnhead H, Cohen GM. Cdc2 activation is not required for thymocyte apoptosis. Biochem. Res. Comm. 1994; 202: 1400–6. | ChemPort |
20. Hakem A, Sasaki T, Kozieradzki I, Penninger JM. The cyclin-dependent kinase Cdk2 regulates thymocyte apoptosis. J. Exp. Med. 1999; 189: 957–68. | Article | PubMed | ISI | ChemPort |
21. Nagata S, Golstein P. The Fas Death Factor. Science 1995; 267: 1449–56. | Article | PubMed | ISI | ChemPort |
22. Nagata S. Apoptosis by death factor. Cell 1997; 88: 355–65. | Article | PubMed | ISI | ChemPort |
23. Rouvier R, Luciani M-F, Golstein P. Fas involvement in Ca²⁺-independent T-cell-mediated cytotoxicity. J. Exp. Med. 1993; 177: 195–200. | Article | PubMed | ISI | ChemPort |
24. Kägi D, Ledermann B, Bürki K et al. Cytotoxicity mediated by T cells and natural killer cells is greatly impaired in perforin-deficient mice. Nature 1994; 369: 31–7. | Article | PubMed | ISI | ChemPort |
25. Heusel JW, Wesselschmidt RL, Shresta S, Russell JH, Ley TJ. Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA fragmentation and apoptosis in allogeneic target cells. Cell 1994; 76: 977–87. | Article | PubMed | ISI | ChemPort |
26. Ebnet K, Hausmann M, Lehmann-Grube F et al. Granzyme A-deficient mice retain potent cell-mediated cytotoxicity. EMBO J. 1995; 14: 4230–9. | PubMed | ISI | ChemPort |
27. Simon M, Waring P, Lobigs M et al. Cytotoxic T cells specifically induce Fas on target cells thereby facilitating exocytosis-independent induction of apoptosis. J. Immunol. (in press).
28. Müllbacher A, King NJC. Target cell lysis by Natural killer cells is influenced by BETA-2-microglobulin expression. Scand. J. Immunol. 1989; 30: 21–9. | PubMed |
29. Müllbacher A, Hill A, Blanden R, Cowden W, King N, Tha Hla R. Alloreactive cytotoxic T cells recognise MHC Class I antigen without peptide specificity. J. Immunol. 1991; 147: 1765–72. | PubMed | ChemPort |
30. Waring P, Khan T, Sjaarda AW. Apoptosis induced by gliotoxin is preceded by phosphorylation of histone H3 and enhanced sensitivity of chromatin to nuclease digestion. J. Biol. Chem. 1997; 272: 17 929–36.
31. Waring P, Beaver J. Cyclosporin A rescues thymocytes from apoptisis induced by low concentrations of thapsigargin: effects on mitochondrial function. Exp. Cell. Res. 1996; 227: 264–76. | Article | PubMed | ISI | ChemPort |
32. Vesely J, Havlicek L, Strnad M et al. Inhibition of cyclin dependent kinases by purine analogues. Eur. J. Biochem. 1994; 224: 771–86. | Article | PubMed | ISI | ChemPort |
33. Takuwa N, Zhou W, Kumada M, Takuwa Y. Involvement of intact inositol-1,4,5-triphosphate sensitive Ca2+ stores in cell cycle progression at the G1/S boundary in serum starved fibroblasts. FEBS Lett. 1995; 360: 173–6. | Article | PubMed | ChemPort |
34. Beaver J, Waring P. Thapsigargin induces mitochondrial dysfunction and apoptosis in the mastocytoma P815 cell line and in mouse thymocytes. Cell Death Differ. 1996; 3: 415–24. | ChemPort |
35. Haller O, Hanson M, Kissling R, Wigzell H. Role of non- conventional natural killer cells in resistance against syngeneic tumour cells in vivo. Nature 1979; 270: 609.
36. Teng JM, Liu XR, Mills GB, Dupont B. CD28 mediated cytotoxicity by the human leukemic NK cell line YT involves tyrosine phosphorylation, activation of phosphatidylinositol 30kinsae. J. Immunol. 1996; 156: 3222–32. | PubMed | ISI | ChemPort |
37. Pouliot N, Maghni K, Blanchette F et al. Natural killer and lectin-dependent cytotoxic activities of Kurloff cells: target cell selectivity, conjugate formation and Ca²⁺ dependence. Inflammation 1996; 20: 647–71. | Article | PubMed | ChemPort |
38. Eischen CM, Leibson PJ. Role for NK-cell associated Fas ligand in cell mediated cytotoxicity and apoptosis. Res. Immunol. 1997; 148: 164–9. | PubMed | ChemPort |
39. Forsburg SL, Nurse PQ. Cell cycle regulation in the yeasts S. cerevisiae and S. pombe. Annu. Rev. Cell Biol. 1991; 7: 227–56. | Article | PubMed | ISI | ChemPort |
40. Greenberg AH, Litchfield DW. Granzymes and apoptosis: targetting the cell cycle. Curr. Top. Microbiol. Immunol. 1995; 19: 95–119.
41. Lundgren K, Walworth N, Booher R, Dembski M, Kirschner M, Beach D. Mik 1 and wee 1 cooperate in the inhibitory tyrosine phosphorylation of cdc2. Cell 1991; 64: 1111–22. | Article | PubMed | ISI | ChemPort |
42. Thastrup O, Cullen PJ, Drobak B, Hanley MR, Dawson AP. Thapsigargin, a tumor promoter, discharges intracellular Ca²⁺ stores by specific inhibition of endoplasmic reticulum Ca²⁺ ATPase. Proc. Natl Acad. Sci. USA 1990; 87: 2466–70. | Article | PubMed | ChemPort |
43. Lorca T, Labbe JC, Devault A et al. Dephosphorylation of cdc2 on threonine 161 is required for cdc2 kinase inactivation and normal anaphase. EMBO J. 1992; 11: 2381–90. | PubMed | ISI | ChemPort |
44. Nishioka WK, Welsh RM. Susceptibility to cytotoxic lymphocyte induced apoptosis is a function of the proliferative status of the cell. J. Exp. Med. 1994; 179: 769–74. | Article | PubMed | ChemPort |
45. Nakajima H, Goldstein P, Henkart PA. The target cell nucleus is not required for cell mediated granzyme or Fas based cytotoxicity. J. Exp. Med. 1995; 181: 1905–9. | Article | PubMed | ChemPort |
46. Irmler M, Thome M, Hahne M et al. Inhibition of death receptor signals by cellular FLIP. Nature 1997; 388: 190–5. | Article | PubMed | ISI | ChemPort |
47. Micheau O, Solary E, Hamman A, Martin F, Dimancheboitrel MT. Sensitivity of cancer cells treated with cytotoxic drugs to Fas mediated cytotoxicity. J. Natl Cancer Inst. 1997; 89: 783–9. | Article | PubMed | ChemPort |
48. Krammer PH. The tumor strikes back — new data on the expression of the CD95 (APO-1Fas) receptor/ligand system may cause paradigm changes in our view on drug treatment and tumor immunology. Cell Death Differ. 1997; 4: 362–4. | Article | PubMed | ChemPort |
49. Reap EA, Roof K, Maynor K, Borrero M, Booker J, Cohen PL. Radiation and stress induced apoptosis: a role for Fas/Fas ligand interactions. Proc. Natl Acad. Sci. USA 1997; 94: 5750–5. | Article | PubMed | ChemPort |

Acknowledgements

The authors would like to Mr Allan Sjaarda and Mr Ron Tha Hla for expert technical assistance and Dr Markus Simon for critical comments on the manuscript.