(Received for publication, March 14, 1997, and in revised form, April 17, 1997)
From the Boston Biomedical Research Institute,
Boston, Massachusetts 02114, the § Medical Radiology
Research Center, 249020 Obninsk, Russia,
Biotechnology Research
Institute, Montreal H4P 2R2, Quebec, Canada, and the ** Dana Farber
Cancer Institute, Boston, Massachusetts 02115
Harmful conditions including heat shock, oxidative stress, UV, and so forth cause programmed cell death, whose triggering requires activation of the Jun N-terminal kinase, JNK. High levels of Hsp72, a heat-inducible member of Hsp70 family, protect cells against a variety of stresses by a mechanism that is unclear at present. Here we report that elevated levels of Hsp72 inhibit a signal transduction pathway leading to programmed cell death by preventing stress-induced activation of JNK. Stress-induced activation of another stress-kinase, p38 (HOG1), is also blocked when the level of Hsp72 is increased. Similarly, addition of a purified recombinant Hsp72 to a crude cell lysate reduced p38 kinase activation, while depletion of the whole family of Hsp70 proteins with a monoclonal antibody enhanced such activation. In addition, we have found that accumulation of abnormal proteins in cells upon incubation with amino acid analogs causes activation of JNK and p38 kinases, which can be prevented by overproduction of Hsp72. Taken together, these data suggest that, in regulation of JNK and p38 kinases, Hsp70 serves as a "sensor" of the build-up of abnormal proteins after heat shock and other stresses. The inhibitory effect of an increased level of Hsp70 on JNK appears to be a major contributor to acquired thermotolerance in mammalian cells.
Exposure of mammalian cells to severe heat shock, strong oxidants, UV irradiation, and other stressful conditions causes activation of a family of homologous stress-activated protein kinases including JNK1 and p38 (1-3). This activation proceeds through a signal transduction pathway that involves the small GTP-binding proteins, MEK kinase, MEKK1, and dual-specificity kinases MKK3, MKK4 (SEK1), and MKK6, which in turn phosphorylate and activate JNK and p38 (4-8). JNK was recently shown to be an essential component of a signal transduction pathway which leads to programmed cell death in response to certain stimuli, including stressful conditions (9-12). In fact, overproduction of a dominant negative mutant of JNK activating kinase, SEK1, inhibits programmed cell death in response to heat shock, UV irradiation, oxidative stress (9), and certain other inducers of apoptosis (10-12).
Another protein, which when overproduced enhances cell survival following exposure to a variety of stressful conditions, is Hsp72, a heat-inducible member of Hsp70 family. Indeed, overproduction of Hsp72 either by expression under the regulation of a strong promoter on a plasmid, or physiologically, by exposure of cells to a mild heat shock, leads to a dramatic protection against severe heat shock, UV irradiation, H2O2, and other harmful conditions and factors (13-16).
Members of the Hsp70 protein family function as molecular chaperones in refolding of denatured polypeptides (17, 18). In fact, overproduction of Hsp72 was shown to reduce stress-induced denaturation and aggregation of certain proteins (19, 20) that has led to the common assumption that refolding and antiaggregating activities of Hsp72 determine its role in cellular protection against stresses (21, 22). However, under some conditions the protective action of Hsp72 appears to be unrelated to its chaperoning activity. For example, tumor necrosis factor (TNF) causes cell death by the activation of a signal transduction pathway leading to apoptosis (23). This apoptotic process can be prevented by overproduction of Hsp72 (24), which therefore seems to interfere with the apoptotic program. Similarly TNF, exposure to high temperatures, and some other stresses cause activation of the apoptosis-triggering signal transduction pathways that include activation of JNK (9-12). The protective action of Hsp72 in these circumstances may also involve direct interference with the apoptotic program. It should be noted, however, that upon exposure to extremely high temperatures, when cell death is not due to apoptosis, but to necrosis, the antiaggregating and protein refolding activities of Hsp72 may become critical for cell protection (20).
At what step can Hsp72 interfere with the apoptotic program? We hypothesized that overproduction of Hsp72 inhibits activation of JNK by harmful conditions. This idea was tested on the U937 leukemia cell line, in which JNK activation appeared to be essential for the activation of apoptosis by heat shock, UV, and oxidative stress (9), and on the PEER lymphoid cell line, in which we overproduced Hsp72 under the control of tetracycline-regulated promotor.
U937 and PEER human lymphoid tumor cells were grown in RPMI 1640 medium with 10% (U937) or 20% (PEER) fetal bovine serum and were used for experiments while in the mid-log phase (3-7 × 105 cells/ml). PEER cells were stably transfected with plasmids encoding Hsp72 under the control of a tetracycline-regulated transactivator. Hsp72 was inserted as a first cistron in a dicistronic expression vector that contains the green fluorescent protein (GFP) gene. In this construct synthesis of Hsp72 is repressed when grown in the presence of tetracycline, but removal of this antibiotic leads to rapid and efficient induction of Hsp72. Cells were incubated in the absence of tetracycline for 48 h in order to induce expression of Hsp72 and GFP. Cells that express GFP were selected with a cell sorter. These cells were viable and continued to grow normally when tetracycline was added to repress Hsp72 synthesis.
Cell Lysate Preparation2 × 107 PEER cells were collected by centrifugation and washed once with phosphate-buffered saline. Cells were resuspended in 0.4 ml of a hypotonic buffer that contained 10 mM Hepes, 10 mM KCl, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 0.2 mM EDTA (pH 7.4). After 15 min of swelling on ice, cells were disrupted on ice by 60 strokes in Dounce homogenizer, and a 0.2 volume of a buffer containing 50 mM Hepes and 250 mM KCl (pH 7.4) was added. Nuclei and cell debris were removed by centrifugation at 500 × g for 5 min at 4 °C, and the lysate was returned to ice. Typically cell lysates prepared by this procedure contained 4-7 mg/ml protein.
Apoptosis AssaysFor determination of cell viability, 0.5 mg/ml of MTT (Sigma) was added to 100 ml of cell suspension (0.5 × 106 cells/ml in 96-well plates) for 4 h, and the formazan formed was dissolved in acidic 2-propanol; optical density was measured using an enzyme-linked immunosorbent assay reader at 590 nm. The OD of formazan formed by control (untreated) cells was taken as 100%. PARP degradation was determined by Western blotting with anti-PARP polyclonal antibodies.
JNK and p38 Kinase AssaysJNK was assayed using glutathione S-transferase-c-Jun protein as a substrate either in an in-gel assay or after immunoprecipitation with anti-JNK antibodies (Santa Cruz Biotechnology) as described previously (25). p38 kinase activity was determined by Western blotting with polyclonal antibodies (New England Biolabs) specifically reacting with the phosphorylated (active) form of p38.
We first investigated the effect of enhanced levels of Hsp72 on
activation of stress kinase and apoptosis in U937 cells. To raise the
level of Hsp72, U937 cells were subjected to a mild heat shock (20 min
at 43 °C), followed by a 6-h recovery. Such mild heat shock caused
only a weak transient activation of JNK and did not activate an
apoptotic program (not shown). The second, severe, heat shock (60 min
at 43 °C) was applied, and it caused death of about 70% of the cell
population (Fig. 1A), as measured by the MTT
assay (ability of cells' mitochondria to reduce tetrazolium salts) or
by counting the number of surviving cells (not shown). Pretreatment
with the mild heat shock (preheating) reduced cell death after the
second heat shock by more than 90% (Fig. 1A). Cell death
was preceded by an extensive cleavage of PARP (Fig. 1B) and
nuclear condensation (not shown), thus indicating that it was of an
apoptotic, rather than of a necrotic, nature. Both the PARP cleavage
(Fig. 1B) and the nuclear condensation were strongly
inhibited in the preheated cells.
Preheating of cells also led to almost complete inhibition of activation of JNK in response to the second challenge with heat shock (Fig. 1C). While JNK became activated 30 min following the heat shock and maintained its activity for 4 h, in preheated cells JNK activity was practically undetectable at any point during this time period (not shown). It is noteworthy that activation of another stress-activated kinase, p38, was also inhibited under these conditions by more than 70% (Fig. 1D). This kinase was shown to participate in induction of IL-6, activation of phospholipase A2, regulation of phosphorylation of Hsp27, and other processes (26-30). Preheating of cells also strongly reduced activation of JNK and p38 kinases in response to another stress, a challenge with 9% ethanol (Fig. 1, C and D), and inhibited ethanol-induced apoptosis (Fig. 1, A and B). Since inactivation of JNK by overexpression of a dominant negative mutant of the upstream kinase was previously shown to protect cells against heat shock and some other stresses (9), we conclude that the protective action of the mild heat shock is associated with an suppression of the JNK activation. Thus, prevention of activation of JNK in preheated cells may be a significant factor in the phenomenon of acquired thermotolerance and tolerance to ethanol. These data complement the recent observation that JNK activation in response to mild heat shock is strongly reduced in a stable thermoresistant cell line (31) that overproduced the whole set of heat shock proteins including Hsp72 (32).
Since the protective effects of the mild heat shock may be associated
not only with Hsp72 but also with some other heat-inducible polypeptides, we also studied the effects of stresses on a stably transfected PEER cell line that overexpresses Hsp72 under control of a
tetracycline-regulated transactivator. We first reproduced the effects
of mild heat shock on this cell line, i.e. demonstrated that
preheating of the cells in the presence of tetracycline (without expression of Hsp72 from a plasmid) protects against heat shock and
ethanol treatments and reduces activation of JNK and p38 (Fig. 1,
C and D). To study the effect of Hsp72 produced
from the plasmid under normal conditions, without mild heat shock (Fig.
2A), cells were incubated in the absence of
tetracycline for 24 h, followed by selection of cells containing
elevated levels of Hsp72 (see "Experimental Procedures"). In these
cells the level of Hsp72 was about 3-4 times higher than that of PEER
cells after the mild heat shock (not shown) and was similar to the
level of Hsp72 in U937 cells following the mild heat shock. The
selected cells with elevated levels of Hsp72 were challenged with a
severe heat shock (60 min at 43 °C) or 9% ethanol. Activation of
both JNK and p38 kinases was strongly inhibited in these cells compared
with the cells with the normal level of Hsp72 grown in the presence of tetracycline (Fig. 2, B and C). A strong
protection against apoptosis upon heat shock and ethanol treatments was
also observed in the Hsp72 overproducing cells (Fig.
2D).
These results support the hypothesis that Hsp72, when overproduced, inhibits activation of JNK and p38 kinases and thus suppresses cell suicide. It should be mentioned that, in the stably transfected cells that chronically overexpress Hsp72 (13), we observed protection from apoptosis without inhibition of JNK activity.2,3 Therefore, it is possible that another mechanism of prevention of apoptosis that does not involve inhibition of JNK operates in cell lines that chronically overexpress Hsp72. On the other hand, the effect described here of the regulated acute overexpression of Hsp72 from the plasmid is identical to that of elevated levels of Hsp72 induced by the mild heat shock, indicating the physiological relevance of the observed phenomenon.
It is noteworthy that the regulated overexpression of Hsp72 in the PEER cell line not only suppressed the responses to heat shock and ethanol but also strongly inhibited activation of both JNK (not shown) and p38 (Fig. 2E) kinases caused by some other stresses, for example, osmotic shock, H2O2, and UV irradiation. Therefore, an elevated level of Hsp72 appears to interfere with transduction of signals from a variety of stressful conditions. Moreover, accumulation of Hsp72 strongly reduced activation of stress-induced kinases in response to TNF and IL-1,4 which is consistent with recent findings that the response of JNK (and potentially of p38 kinase) to UV and osmotic shock is mediated by the activation of TNF, IL-1, and epidermal growth factor receptors (33).
Does a high level of Hsp72 directly affect activation of JNK and p38
kinases or does it through induction or repression of synthesis of some
other proteins? To address this question, effects of addition of a
purified recombinant Hsp72 on the activation of p38 kinase in cell
lysate were studied. In a crude lysate p38 kinase becomes activated
upon incubation at 20 °C for 15 min (Fig. 3A). Under exposure of the lysate to
43 °C, p38 kinase was activated even further (Fig. 3B).
No additional p38 activation was observed in the lysates subjected to
UV irradiation (Fig. 3B). Addition of purified Hsp72 (but
not bovine serum albumin, not shown) repressed the activation by about
40-70% (Fig. 3A). Importantly, the addition of a
monoclonal anti-Hsp70 antibody (MA3-007, Affinity Bioreagents) raised
against the conserved ATPase domain, which cross-reacts with all major
species of Hsp70, had an effect opposite to that triggered by the
addition of Hsp72. It led to a 1.5-3-fold stimulation of p38 kinase
activity (Fig. 3B). Addition of control monoclonal anti--galactosidase antibody did not cause activation of p38 (not
shown). These data indicate that the level of Hsp70 in a cell may
directly regulate stress-activated kinases. The observed variability in
the extent of stimulation of p38 by MA3-007 antibody may be due to the
presence of endogenous ADP. Indeed, addition of ADP strongly reduced
the stimulatory effect of the anti-Hsp70 antibody (not shown).
Experiments with activation of p38 kinase in a cell lysate by addition
of anti-Hsp70 antibody suggested that depletion or just capturing of
Hsp70 may result in activation of stress kinases in cells. Members of
the Hsp70 family bind misfolded polypeptides and therefore could be
captured upon a build-up of abnormal misfolded proteins in the cytosol.
Therefore, to test whether accumulation of abnormal proteins cause
activation of stress kinases, cells were incubated with amino acid
analogs, which incorporate into newly synthesized polypeptides and
prevent their proper folding. In fact, incubation of cells with the
proline analog, L-azetidine carboxylate, as well as with
the arginine analog, canavanine, stimulated both JNK and p38 kinases,
and the overproduction of Hsp72 strongly reduced such activation (Fig.
4, A and B). A similar stimulation
of stress-activated kinases was observed upon incubation of the cells
with puromycin, which causes an accumulation of truncated and therefore
improperly folded N-terminal fragments of proteins (not shown).
Therefore, build-up of abnormal proteins activates stress kinase, while
Hsp72 and probably other members of Hsp70 family suppress such
activation.
These data suggest a mechanism of activation of stress kinases by heat
shock and some other stressful conditions. Such stresses may cause
protein damage and denaturation and may lead to a build-up of abnormal
proteins. In turn these abnormal proteins deplete free Hsp70, which
leads to activation of stress-activated kinases (Fig.
5). The role of Hsp70 in the activation of stress
kinases may be analogous to its role in activation of heat shock gene transcription (34, 35). Accordingly, we suggest that in unstressed cells Hsp70 associates with an upstream component of the kinase cascade
and keeps it in an inactive form. Depletion of a free cellular Hsp70 by
accumulated abnormal proteins after heat shock (21, 22) would cause
dissociation of the bound Hsp70 from the upstream component of the
kinase cascade, thus leading to activation of downstream kinases.
Therefore, Hsp70 may function as a "stress sensor" that triggers
not only transcription of the heat shock genes but also activation of
stress kinases.
The results described in this report may also explain why overproduction of Hsp72 prevents activation of the phospholipase A2 by TNF (36) and prevents induction of synthesis of IL-6 by IL-1 (15). Indeed, Hsp72 may simply prevent activation by TNF or IL-1 of the p38 kinase, which was shown to activate the phospholipase A2 (27) and which is responsible for IL-6 induction (26). In summary, Hsp70 seems to play a very general role in signal transduction pathways, which lead to adaptation of cells and organisms to stressful conditions.
We thank Dr. Vladimir Volloch and Dr. Alfred Goldberg for helpful discussion.