HSP72 can protect cells from heat-induced apoptosis by accelerating the inactivation of stress kinase JNK

Abstract

The major heat shock protein Hsp72 prevents heat-induced apoptosis. We have previously demonstrated that transiently expressed Hsp72 exerts its anti-apoptotic effect by suppressing the activity of stress-kinase JNK, an early component of the apoptotic pathway initiated by heat shock. On the other hand, constitutive expression of Hsp72 does not lead to suppression of heat-induced JNK activation, yet still efficiently prevents apoptosis. To address this apparent contradiction, we studied the effects of constitutively expressed Hsp72 on activation of JNK and apoptosis in Rat-1 fibroblasts. We found that the level of heat-induced apoptosis directly correlated with the duration rather than the magnitude of JNK activity following heat shock. Constitutively expressed Hsp72 strongly reduced the duration of JNK while it did not suppress initial JNK activation. These effects were due to Hsp72-mediated acceleration of JNK dephosphorylation. Addition of vanadate to inhibit JNK phosphatase activity completely prevented the anti-apoptotic action of Hsp72. Therefore, suppression of heat-induced apoptosis by Hsp72 could be fully accounted for by its effects on JNK activity.

INTRODUCTION

The major inducible heat shock protein, Hsp72, plays a key role in thermotolerance. Hsp72 expression has been shown to decrease heat shock-induced apoptosis (Gabai et al 1997; Li et al 1996; Mosser and Martin 1992) and necrosis and to increase colony-forming abiliy following heat shock (Angelidis et al 1991; Li et al 1991). Members of the Hsp70 family function as molecular chaperones by facilitating protein folding, translocation across membranes and degradation (see Feige et al 1996 for review). Therefore, Hsp72 was suggested to increase cell survival by binding to unfolded polypeptides and preventing protein aggregation as well as accelerating the refolding or degradation of damaged proteins (Kabakov and Gabai 1997; Kampinga 1993). Indeed, Hsp72 expression strongly decreases heat-induced aggregation of nuclear proteins as well as heat-inactivation of firefly luciferase expressed in the cytosol of mammalian cells (Michels et al 1997; Stege et al 1994; Stege et al 1994).

The chaperone function of Hsp72 resulting in protein protection could indeed be important for prevention of heat-induced apoptosis. However, Hsp72 may also prevent apoptosis through interference with apoptotic signal transduction. In doing so, Hsp72 could act analogously to anti-apoptotic protein bcl-2, which is an effective suppressor of heat-induced apoptosis (Strasser and Anderson 1995) but not a chaperone. We have previously demonstrated that Hsp72 can suppress the activation of stress kinase JNK (Gabai et al 1997), an early step in the apoptotic signaling pathway induced by heat shock and many other stresses (UV and gamma radiation, oxidative stress, ceramide etc) (Pena et al 1997; Seimiya et al 1997; Verheij et al 1996; Xia et al 1995; Zanke et al 1996). JNK was demonstrated to phosphorylate, a tumor suppressor p53, reducing its ubiquitination and, subsequently, degradation; this facilitates p53 accumulation and leads to apoptosis (Fuchs et al 1998). JNK-mediated apoptosis has also been shown to occur in a p53-independent fashion, probably by a mechanism related to the induction of FAS ligand (Faris et al 1998a; Faris et al 1998b). Interruption of the JNK signaling pathway almost completely blocked stress-induced apoptosis in various cell lines (Meriin et al 1998; Verheij et al 1996; Zanke et al 1996). Thus, despite the initial damage that would otherwise result in apoptosis, cells can survive if the apoptotic-signaling pathway is interrupted at the JNK level.

We and others previously found that transient expression of Hsp72 suppresses heat-induced JNK activation as well as apoptosis in human lymphoid cells and fibroblasts (Gabai et al 1997; Mosser et al 1997; Volloch et al 1998). Furthermore, JNK suppression was observed also in cells pretreated with mild heat shock, a natural inducer of Hsps (Buzzard et al 1998; Gabai et al 1997; Volloch et al 1998). These data strongly indicate that the suppression of apoptosis by Hsp72 is due to the suppression of JNK. This conclusion, however, contradicts other reports that expression of Hsp72 constitutively rather than transiently does not cause suppression of heat-induced JNK activation, although apoptosis of these cells was markedly alleviated (Buzzard et al 1998; Mosser et al 1997). Based on these data, the authors suggested that in blocking heat-induced apoptotic signal transduction, Hsp72 has an additional target(s) downstream of JNK (Buzzard et al 1998; Mosser et al 1997). In the present work we used fibroblasts that constitutively express Hsp72 to clarify whether the effect of Hsp72 on JNK activation could adequately explain Hsp72-mediated protection from heat-induced apoptosis.

MATERIALS AND METHODS

Cell lines

Parental Rat-1 fibroblasts and their variants expressing Hsp72 (MVH) were kindly provided by Dr G.C. Li. Cells were grown in Dulbecco Minimal Essential Medium (DMEM) supplemented with 10% fetal bovine serum and used for experiments at 40% to 70% confluency.

Adenoviral-based expression of SEK(K/R)

The SEK-1(K/R) recombinant adenovirus has been described previously (Choukroun et al 1998). This construct expresses the kinase-inactive mutant of SEK-1 tagged with M2 FLAG epitope at its amino terminus (Choukroun et al 1998). Adenovirus was propagated in 293 cells, and high titer stocks were obtained and purified by CsCl density gradient centrifugation.

Apoptosis quantification

Apoptosis levels were measured by fluorescent microscopy using Hoechst-33342 DNA-specific dye (10 μM). Shrunken cells with condensed or fragmented nuclei were counted as apoptotic. Poly (ADP-ribose) polymerase (PARP) cleavage was monitored by Western blot analysis with anti-PARP monoclonal antibodies (C2–10, G. Poirier, Montreal, Canada).

Measurement of JNK activity

Cells were lysed in a buffer (200 μL per 35 mm dish) containing 40 mM HEPES, pH 7.5; 50 mM KCl; 1% Triton X-100; 2 mM DTT, 1 mM Na₃VO₄; 50 mM β-glycerophosphate; 50 mM NaF; 5 mM EDTA; 5 mM EGTA; 1 mM phenylmethylsulfonyfluoride; 1 mM benzamedine; and 5 μg/mL each: leupeptin, pepstatin A, aprotinin. The lysates were clarified by centrifugation in a microcentrifuge at 15 000 rpm for 5 minutes. To measure JNK activity, 5 μL extract was added to a reaction mixture (20 μL final volume), containing 40 mM HEPES, pH 7.5; 1 mM Na₃VO₄; 25 mM β-glycerophosphate; 10 mM MgCl₂; 20 μM ATP; 15 μCi of ³²P-γ-ATP; and 40 ng of recombinant c-Jun-GST. The reaction was allowed to proceed for 30 minutes at 30°C and then was stopped by addition of 10 μL SDS-PAGE loading buffer. Samples were separated by SDS-PAGE, transferred to nitrocellulose membranes and exposed to a Molecular Imager for quantitation. Subsequently, the membranes were immunoblotted with a JNK1 antibody to verify equal protein loading.

Another assay for JNK activity allowed us to measure separately the activities of 2 major isoforms (46 kDa JNK1 and 54 kDa JNK2) by immunoblotting with JNK antibody specific to the activated (phosphorylated) form of JNK (Promega, Madison, WI, USA). Secondary antibodies conjugated with peroxidase were visualized with ECL substrates (Amersham, Arlington Heights, IL, USA), and the resulting films were quantified by densitometry.

Data of typical experiments, performed in triplicate, are shown.

Measurement of JNK phosphatase activity

In vivo JNK phosphatase activity was measured as described earlier (Meriin et al 1999). Briefly, cells were exposed to heat shock and transferred back to 37°C. Then further JNK phosphorylation (activation) was completely inhibited by changing the medium to phosphate-buffered saline with 10 mM 2-deoxyglucose and 5 μM rotenone, which rapidly depleted ATP, and the samples were taken at different time points. The rate of JNK dephosphorylation under these conditions was assessed by immunoblotting with antibody that recognizes the activated (phosphorylated) form of JNK. Data of typical experiments, performed in triplicate, are shown.

RESULTS

Constitutive expression of Hsp72 in Rat-1 fibroblast blocks heat-induced apoptosis but does not affect the initial activation of JNK

We previously demonstrated that the transient expression of Hsp72 in cells exposed to heat shock leads to dramatic suppression of JNK activation, which can account for the protective (anti-apoptotic) effect of Hsp72. Therefore, it came as a surprise that constitutive expression of this heat shock protein in some cell lines did not suppress JNK activation following a heat shock, but nevertheless rendered cells thermoresistant (Buzzard et al 1998; Mosser et al 1997). To address the mechanism underlying such thermoresistance we used Rat-1 cells and their variants constitutively expressing Hsp72 (MVH) (Li et al 1991; Li et al 1992). When parental Rat-1 cells were subjected to severe heat shock (45°C, 50 minutes), about 80% of cells underwent apoptosis within 24 hours, while upon exposure of MVH cells to the same heat shock, the extent of apoptosis was only about 30% (Fig 1A). The apoptotic nature of heat-induced cell death was manifested by extensive cleavage of poly (ADP-ribose) polymerase (PARP), a caspase-3 substrate (Fig 1B). In MVH cells, PARP cleavage was suppressed dramatically (Fig 1B). Unexpectedly, when JNK activity was measured in these cells, similar levels of JNK activation were seen in parental Rat-1 and MVH cells (Fig 1C).

Fig. 1.

Constitutive expression of Hsp72 in Rat-1 fibroblasts suppresses heat-induced apoptosis without inhibition of JNK activation. Rat-1 cells and Hsp72-expressing variant (MVH) were subjected to heat shock and their apoptotic (A, B) and JNK activity (C) were measured as described in Material and Methods. (A) Apoptosis (mean ± SD) was assessed 24 hours after heat shock (45°C, 50 minutes) by Hoechst staining; (B) cleavage of PARP 16 hours after heat shock at 45°C for 30 or 45 minutes; (C) JNK activity immediately after heat shock (45°C, 50 minutes) was assessed by c-jun phosphorylation (see Material and Methods for details).

These results opened the possibility that heat-induced apoptosis may be JNK-independent in Rat-1 cells unlike several other cell lines (Verheij et al 1996; Zanke et al 1996; Meriin et al 1998). To test this possibility, we suppressed JNK activity by expression of a dominant-interfering mutant form of JNK-activating kinase SEK1, SEK1(K/R) (ref Sanchez et al 1994; Meriin et al 1998; Verheij et al 1996). To express SEK1(K/R) in Rat-1 cells, we used adenovirus-mediated gene transfer (Choukroun et al 1998). Increasing amounts of viral particles were added to the cell cultures. By 3 days after infection the cells had accumulated substantial amounts of SEK(K/R) protein. The levels of SEK(K/R) directly correlated with the amounts of adenovirus used for infection (Fig 2). When these cells were exposed to heat shock, JNK activation and heat-induced apoptosis were suppressed in SEK1(K/R) expressing cells in a dose-dependent fashion (Fig 2). Therefore, JNK activity seems to be a required component of heat-induced apoptotic pathway in Rat-1 cells.

Fig. 2.

Inhibition of JNK activation by dominant-negative SEK1 mutant suppresses heat-induced apoptosis of Rat-1 cells. Rat-1 cells were infected with increasing titers of adenovirus expressing dominant-negative SEK (SEK1K/R), and activation of JNK1, JNK2 and apoptosis after heat shock (45°C, 50 minutes) were assessed as described in Material and Methods. Expression of SEK1(K/R) was quantified with antibody to SEK1. One relative unit of dnSEK adenovirus titer was 10⁹ particles per 35 mm dish. Percent of effects relative to uninfected cells is plotted on the ordinate axis

Constitutive expression of Hsp72 in Rat-1 fibroblasts accelerates JNK inactivation following heat shock

Since JNK is necessary for heat-induced apoptosis in Rat-1 cells, we suggested that constitutively expressed Hsp72 protects cells not via reduction of the extent of initial JNK activation but by limiting the duration of heat-induced JNK activity. This possibility is consistent with several recent studies which suggest that transient JNK activation is insufficient for apoptosis triggered by gamma-radiation (Chen et al 1996), Tumor Necrosis Factor (TNF) (Guo et al 1998a, 1998b) and cisplatin (Sanchez-Perez et al 1998) while prolonged activation of JNK by these treatments leads to apoptosis.

Therefore, we tested whether heat-induced apoptosis in Rat-1 cells correlates with the duration of JNK activation. Incubation of Rat-1 cells at 45°C for 20, 30, or 50 minutes caused strong activation of JNK. Upon transfer of the cells to 37°C, JNK activity further increased for at least 1 hour, followed by the subsequent decline (Fig 3A). While these heat treatments caused similar JNK activation, the rates of decline of JNK activity differed dramatically. In cells exposed to 45°C for 20 minutes, JNK activity fell to 50% of its maximal level within 2.5 hours after heat shock, whereas, in cells heat-shocked for 30 minutes, JNK activity reached its half-maximal level only after 4 hours. After a more severe heat treatment (45°C for 50 minutes), JNK activity remained close to its maximal level for more than 5 hours after transfer to normal temperature (Fig 3A).

Fig. 3.

Heat-induced apoptosis in Rat-1 cells correlates with duration of JNK activation. Rat-1 cells were subjected to various heat shock treatments (20, 30, or 50 minutes at 45°C), then transferred to normal temperature (37°C) and their JNK activity (A) and apoptosis (B) were measured at the indicated time points. JNK activity was similar immediately after different heat shock treatments and was taken as one relative unit

When assessed after 8 hours, neither 20-minute nor 30-minute heat treatments induced apoptosis, whereas, in culture treated for 50 minutes, 40% of cells became apoptotic. When measured after 24 hours, no apoptosis was observed after 20-minute heat treatment, and only a small fraction of cells became apoptotic in cultures heat-shocked for 30 minutes (about 30%), whereas 50-minute treatment caused apoptosis in 80% of cells (Fig 3B). Hence, the onset and the extent of heat-induced apoptosis in Rat-1 cells correlates with the duration of JNK activation rather then with the extent of the initial JNK activation following heat shock.

The results led us to hypothesize that Hsp72 protects against heat-induced apoptosis by directly or indirectly facilitating JNK inactivation. Indeed, upon rather mild (30 minutes at 45°C) and severe (50 minutes at 45°C) heat shock, JNK inactivation in Hsp72-expressing cells occurred significantly faster than in parental cells (Fig 4A,B). While, in paternal Rat-1 cells JNK activity remained very close to its maximal level 5 hours after the 50-minute heat shock (projected time for decline to 50% level was about 14 hours, Fig 4B), in MVH cells JNK activity fell to 50% of its maximal level 4 hours after heat shock. Therefore, Hsp72 accelerated the decline of JNK activity so that the rate of JNK inactivation in Hsp72-expressing cells after severe heat shock (50 minutes at 45°C) resembled that seen in parental cells after milder heat shock (30 minutes at 45°C) (Fig 4B). Only 35% of MVH cells exposed to severe heat shock became apoptotic (Fig 1A), a percentage similar to that seen in parental cells exposed to milder heat shock (30 minutes at 45°C) (Fig 3B). Thus, Hsp72-mediated protection against apoptosis correlates well with the rate of JNK inactivation (Fig 4B).

Fig. 4.

Constitutive expression of Hsp72 in Rat-1 cells accelerates JNK inactivation after heat shock. Rat-1 cells and their Hsp72-expressing variant (MVH) were subjected to heat shock (45°C, 30 minutes. (A), or 30 and 50 minutes, (B) and JNK1 and JNK2 activities were measured immediately after heat shock and recovery at 37°C. Panel B presents quantification of data for JNK1

Hsp72 accelerates JNK inactivation by facilitating its dephosphorylation in heat-shocked cells

We have previously demonstrated that heat shock and other protein damaging stresses do not facilitate JNK phosphorylation by upstream kinases, but inhibit its dephosphorylation, thus leading to increase in JNK activity (Meriin et al 1999). We have also reported that transient induction of Hsp72 suppresses the inhibitory effect of heat shock on JNK dephosphorylation (Meriin et al 1999). Based on these observations we suggested that stable expression of Hsp72 can accelerate the inactivation of JNK following severe heat shock by facilitating JNK phosphatase activity. To assess the activity of JNK phosphatase, we measured the rate of JNK dephosphorylation under the conditions when upstream JNK-activating kinases are completely suppressed (Meriin et al 1999). To rapidly and completely block JNK activation by heat shock, we employed a method based on ATP depletion (see Materials and Methods) which was developed previously (Meriin et al 1999). Cells were transferred into a minimal media, and the synthesis of ATP was blocked by rotenone, an inhibitor of the respiratory chain, and 2-deoxyglucose, a competitive inhibitor of glycolysis. This treatment led to a 10-fold decrease in cytosolic ATP concentration in less than 10 minutes (Meriin et al 1999). In cells exposed to heat shock, initial ATP concentrations and rates of ATP depletion were similar to those in untreated cells (Meriin et al 1999). Dephosphorylation of JNK in unstressed cells (starting from a basal level) following ATP depletion was very rapid (not shown). Activation of JNK by UV-irradiation did not affect the rate of JNK dephosphorylation which was remarkably fast (Fig 5A). By contrast, in parental Rat-1 cells following heat shock, JNK phosphatase activity sharply declined, thus accounting for the heat-induced activation of JNK (Fig 5A). However, in heat-stressed cells expressing Hsp72, the activity of JNK phosphatase was about 3 times higher than that of parental heat-stressed cells (Fig 5B). Thus, Hsp72 accelerates JNK inactivation by facilitating activity of JNK phosphatase.

Fig. 5.

Hsp72 expression alleviates heat-induced JNK phosphatase inactivation after heat shock in Rat-1 cells. (A) Heat shock inactivates JNK phosphatase in Rat-1 cells. Rat-1 cells were subjected to UV irradiation (400 J/min²) or heat shock (45°C, 30 minutes) and JNK1 dephosphorylation was measured under ATP-depleting conditions (chase) at the indicated time points. Under the chase condition JNK phosphorylation by upstream kinases was totally suppressed, therefore the rate of JNK dephosphorylation reflects JNK phosphatase activity (see Material and Methods and ref (Meriin et al 1999 for details). (B) JNK phosphatase activity after heat shock is increased in Hsp72 expressing cells. Cells were subjected to heat shock (45°C, 30 minutes) and JNK1 dephosphorylation was measured under the chase conditions following 2 hours of recovery at 37°C. Rat-1–control cells; MVH-Hsp72 expressing variant; Rat-1/TT–thermotolerant Rat-1 (pretreated with mild heat shock at 45°C for 15 with followed by recovery at 37°C for 16 hours)

Inhibition of JNK phosphatase in Hsp72-expressing cells abolishes their thermoresistance

The results suggested a model in which the accelerated JNK inactivation accounts for the anti-apoptotic effect of Hsp72. To address this possibility we tested whether the inhibition of JNK dephosphorylation in Hsp72-expressing cells decreases cell thermoresistance. To inhibit JNK phosphatase, we used ortho-vanadate, a well-known general phosphatase inhibitor that effectively suppresses JNK phosphatase in various cell lines (Chen et al 1996; Guo et al 1998a; Meriin et al 1999). Treatment of unstressed cells with vanadate alone did not affect JNK activity (not shown) and had little effect on the proportion of apoptotic cells (Fig 6B). On the other hand, when added 1 hour after the stress, ortho-vanadate strongly inhibited JNK inactivation in heat-shocked MVH cells (Fig 6A). At the same time, vanadate completely reversed inhibitory effect of Hsp72 on heat-induced apoptosis in MVH cells (Fig 6B). This observation is consistent with the idea that accelerated JNK inactivation accounts for the resistance to apoptosis in Rat-1 cells constitutively expressing Hsp72.

Fig. 6.

Inhibition of JNK dephosphorylation in Hsp72-expressing cells by vanadate abolishes their thermoresistance. (A) Ortho-vanadate inhibits JNK dephosphorylation in Hsp72-expressing (MVH) cells. The cells were exposed to heat shock (45°C, 50 minutes) and then transferred to normal temperature (37°C). After 1 hour of incubation at 37°C, ortho-vanadate (0.5 mM) was added to a portion of the cells, while another portion was left without the drug, and JNK1 activity was measured by antibody to phospho-JNK. Ortho-vanadate itself had no significant effect on JNK activity (not shown). (B) Ortho-vanadate abolishes thermoresistance of Hsp72-expressing cells. The cells were treated as described above, and their apoptosis (mean ± SD) was assessed 7 hours after the treatments with heat shock (HS, 45°C, 50 minutes), ortho-vanadate (van, 0.5 mM), or in combination (HS + van).

The effect of Hsp72 on JNK activation depends on the cellular level of Hsp72 and the severity of heat shock

Several studied cell lines that transiently expressed Hsp72 showed suppression of the initial JNK activation (measured immediately following a heat shock), while in Rat-1 cells constitutive expression of Hsp72 did not show this effect. Hence, constitutive expression of Hsp72, in contrast to its transient expression, may result in a cellular adaptation that alters the regulation of JNK activity in response to stress. To address this possibility, we studied the effects of the transient expression of Hsp72 (and other heat shock proteins) achieved by pretreatment of Rat-1 with mild heat shock (45°C for 15 minutes) followed by 16 hours of recovery. Such pretreatment made cells thermotolerant so that apoptosis in response to severe heat shock (45°C, 50 minutes) was reduced by more than 3-fold (not shown). As seen with constitutively expressed recombinant Hsp72 (Figs 4 and 7), pretreatment with mild heat shock did not reduce the initial level of heat-induced JNK activation but accelerated JNK inactivation (Fig 7). Therefore, transient and constitutive expression of Hsp72 in Rat-1 cells have similar effects: both affect the rate of JNK inactivation without suppressing the initial JNK activation. Furthermore, as with constitutive expression of Hsp72 in MVH cells, pretreatment of cells with mild heat shock caused acceleration of JNK dephosphorylation (Fig 5B).

Fig. 7.

Transient expression of Hsp72 in Rat-1 cells accelerates JNK inactivation after heat shock. Rat-1 cells were pretreated with mild heat shock (45°C, 15 minutes) with subsequent recovery for 16 hours at 37°C. Such treatment leads to marked accumulation of Hsp72 and thermotolerance (not shown). Then control (Rat-1) and thermotolerant (TT-Rat-1), were subjected to heat shock (45°C, 30 minutes) and JNK1 activity was assessed by antibody to phosphorylated (active) JNK at the indicated time points.

The difference in the effects of Hsp72 on heat-induced JNK activity in these and prior experiments could also be due to differences in cellular levels of Hsp72 and in the severity of heat shock. Indeed, the MVH30 cell clone we selected that expresses Hsp72 at a significantly higher level than MVH cells, exhibited a reduced level of initial JNK activation following heat shock (30 minutes at 45°C) (Fig 8A). Furthermore, when cells were treated with a much milder heat shock (43°C, 30 minutes), the initial activation of JNK was lower in MVH cells compared to parental Rat-1 cells (Fig 8C). Thus, the effects of Hsp72 on JNK depend on the strength of heat shock and on the level of Hsp72 expression. Upon relatively mild heat shock conditions, Hsp72 is capable of suppression of the initial heat-induced JNK activation. Upon severe heat shock, however, the ability of Hsp72 to suppress the initial JNK activation depends on its cellular level. When this level is not sufficiently high, Hsp72 loses its ability to suppress the initial JNK activation but is still able to accelerate JNK inactivation.

Fig. 8.

Suppression of heat-induced JNK activation by Hsp72 depends on its expression level and the severity of heat shock. (A) MVH30 clone has increased level of Hsp72. Equal amounts of total cell protein were loaded as measured by Bio-Rad protein assay reagent. (B) Initial JNK activation after severe heat shock (45°C, 30 minutes) is suppressed in MVH30 cells. (C) Expression of Hsp72 in MVH cells suppresses initial JNK activation after mild heat shock (43°C, 30 minutes). Control (Rat-1), Hsp72-expressing variants (MVH and MVH30), and thermotolerant (TT) cells were exposed to heat shock and their JNK1 activity was measured by antibodies to phosphorylated (active) JNK at the time points indicated. The cells were made thermotolerant by pretreatment with mild heat shock (45°C, 15 minutes) with subsequent recovery for 16 hours at 37°C.

DISCUSSION

Previously, we and others have demonstrated that transient Hsp72 expression suppresses heat-induced activation of JNK and consequently protects cells from apoptosis (Gabai et al 1997; Mosser et al 1997). However, in several studies, constitutive Hsp72 expression protected cells from heat-induced apoptosis apparently without suppression of JNK activity. This suggested that Hsp72 has an additional target(s) downstream of JNK in the blockade of apoptosis (Buzzard et al 1998; Mosser et al 1997). The present study of the Rat-1 cell line with constitutive Hsp72 expression was undertaken in part to solve this controversy.

First, we checked whether heat-induced apoptosis in this cell line is dependent on JNK activation, since there was a report that heat-induced apoptosis of some cells (ie, thymocytes) may be JNK-independent (Nishina et al 1997). However, when heat-induced JNK activation in Rat-1 cells was suppressed by expression dominant-negative SEK1, apoptosis was markedly reduced (Fig 2). Thus, at least in Rat-1 cells, as in the most of cell lines studied (Meriin et al 1998; Verheij et al 1996; Zanke et al 1996), JNK activity is needed for heat-induced apoptosis.

Next we tested the possibility that in cell types in which Hsp72 did not affect the initial level of JNK activation, this heat shock protein reduces the duration of JNK activation following heat shock. This idea was based on recent observations suggesting that the duration of JNK activation is an important factor in induction of apoptosis. For example, prolonged JNK activation was critical for apoptosis caused by UV- or γ-irradiation (Chen et al 1996). Furthermore, inhibition of JNK dephosphorylation leading to prolonged JNK activation dramatically enhanced apoptosis caused by TNF (Guo et al 1998a, 1998b) or combination of phorbol ester with ionomycin (Chen et al 1996). Likewise, we have found that Rat-1 fibroblasts treated at 45°C for 20 minutes displayed a brief JNK activation and no apoptosis, whereas more severe treatment at 45°C for 50 minutes resulted in a prolonged JNK activation and a massive apoptosis. Importantly, the extent of the initial JNK activation under both mild and severe stresses were practically identical. Therefore, since JNK activity is critical for heat-induced apoptosis, it is the prolonged duration of JNK activity that initiates programmed cell death.

Similar to results in PEER and MEF cells (Buzzard et al 1998; Mosser et al 1997) constitutive expression of Hsp72 in Rat-1 fibroblasts provided protection from heat-induced apoptosis without suppressing initial JNK activation. We showed, however, that in Rat-1 cells, Hsp72 dramatically accelerated the inactivation of JNK following heat shock. Therefore, we suggest that Hsp72-mediated protection of Rat-1 cells from heat-induced apoptosis is due to the limitation of the duration of JNK activation. This explanation is supported by our observation that inhibition of JNK inactivation in Hsp72-expressing cells by vanadate eliminated their thermoresistance (Fig 6). Thus, reducing the duration of JNK activation appears sufficient for the prevention of apoptosis. These results strongly suggest that JNK is the major target of Hsp72 in protection from heat-induced apoptosis.

Jaattela et al found that Hsp72 can exert its anti-apoptotic functions downstream of caspase-3, since apoptosis in response to certain stimuli was prevented by Hsp72 despite cleavage of PARP, a substrate of caspase-3 (Jaattela et al 1998). This is not the case, however, with heat-induced apoptosis in Rat-1 cells, where PARP cleavage was strongly inhibited by Hsp72 (Fig 1A). It should be noted that in other cell lines (PEER, MEF) Hsp72 also exerted its effect upstream of caspase-3 (Buzzard et al 1998; Mosser et al 1997).

Previously, we have found that upon heat shock, JNK is activated not by stimulation of upstream components of the JNK signaling cascade, but via inactivation of JNK phosphatase (Meriin et al 1999). Furthermore, in several cell lines, the expression of Hsp72 suppressed initial JNK activation following heat shock by facilitating JNK dephosphorylation. Here we demonstrate that Hsp72 reduces the duration of JNK activation by a similar mechanism. Indeed, direct measurements indicate that the rate of JNK dephosphorylation following heat shock is much faster in MVH cells than in parental Rat-1 cells. These data strongly suggest that the suppression of the initial JNK activation and the limitation of the duration of JNK activation are 2 manifestations of the same activity of Hsp72 in facilitating JNK dephosphorylation.

Which effect of Hsp72 on JNK activity is seen upon heat shock depends on a balance between levels of Hsp72 and the severity of protein damage in heat-shocked cells. In other words, higher levels of Hsp72 are required for suppression of initial JNK activation than for reduction of the duration of JNK activity. For example, in MVH30 cells with high cellular levels of Hsp72, the initial JNK activation was suppressed in response to severe heat shock (45°C for 30 minutes). In contrast, MVH cells with lower levels of Hsp72 showed no suppression of the initial JNK activation in response to the same heat shock. Nevertheless, the duration of JNK activation was nevertheless shortened, and cells were thermoresistant. However, under a milder stress (43°C), Hsp72 expression in MVH cells was sufficient to reduce the initial JNK activation. As a rule, cell lines with constitutive expression of Hsp72 have lower levels of this protein than the same cells with inducible expression (probably because very high levels of Hsp72 are toxic). These cell lines usually respond to severe heat shock with accelerated JNK inactivation rather than suppression of the initial JNK activation.