Heat Shock Protein 72 Modulates Pathways of Stress-induced Apoptosis*

The resistance to stress-induced apoptosis conferred by the thermotolerant state or by exogenous expression of HSP72 was measured in mouse embryo fibroblasts. The induction of thermotolerance protects cells from heat, tumor necrosis factor a (TNF a ), and ceramide-induced apoptosis but not from ionizing radiation. Because the development of thermotolerance is associated with increased levels of heat shock proteins, we determined whether constitutive expression of one of the major inducible heat shock proteins, HSP72, could also protect cells from stress-induced apoptosis. Cells expressing constitutive HSP72 were shown to have significantly reduced levels of apoptosis after heat, TNF a , and ceramide but not after ionizing radiation. Activation of stress-activated protein kinase/c-Jun N-terminal kinase (SAPK/JNK) was found to be strongly inhibited in thermotolerant cells after heat shock but not after other stresses. Cells that constitutively express HSP72 did not demonstrate decreased SAPK/JNK activation after any of these stresses. Thus, factors other than HSP72 that are induced in the thermotolerant state are able to reduce activation of SAPK/JNK after heat stress. Notably, the level of activation of SAPK/JNK did not correlate with the amount of apoptosis detected after different stresses. Constitutive HSP72 expression inhibited poly-(ADP-ribose) polymerase cleavage in cells after heat shock and TNF a but not after ceramide or ionizing radiation. StressGen Biotechnologies). The cleavage of poly(ADP-ribose) polymerase (PARP) from a 116-kDa protein to an 85-kDa fragment was detected using a mouse monoclonal antibody, C-2–10 (a kind gift from Dr. S. Kaufmann), which recognizes both the full-length and the 85-kDa fragment. Samples containing equal amounts of protein were subjected to SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose. The membranes were incu- bated with C-2–10, SPA-810, or N15 antibody, followed by horseradish peroxidase-conjugated rabbit anti-mouse IgG and ECL chemilumines- cence for PARP or alkaline phosphatase-conjugated goat anti-mouse IgG and color staining for HSP72. As a loading control, membranes were immunostained with an anti- a -tubulin antibody (clone B-5–1-2, Sigma) and detected by a color reaction using an alkaline phosphatase-conjugated secondary antibody. SAPK/JNK Assay— SAPK/JNK activity was assayed in vitro using a bacterially expressed glutathione S -transferase (GST)-c-Jun (1–141) fusion protein as a substrate (26). GST-c-Jun was bound to glutathione-Sepharose 4B beads and maintained as a 10% slurry in phosphate- buffered saline/sodium azide. Cells were washed in cold phosphate-buffered saline and lysed in PLC buffer (50 m M Hepes, pH 7.5, 10% glycerol, 1% Triton X-100, 150 m M NaCl, 10 m M MgCl 2 , 1 m M EGTA, 5 m M dithiothreitol, 1 m g/ml aprotinin, 1 m g/ml leupeptin, 1 m g/ml pep- statin, 15 m M p -NPP, 1 m M phenylmethylsulfonyl fluoride) and incubated on ice for 10 min. The lysate was cleared

The regulation of normal development involves a combination of cell division, differentiation, and death. Although much emphasis was previously placed on mechanisms of cell cycle progression, mechanisms of cell death have only recently begun to be understood. Programmed cell death, or apoptosis, is an active process resulting in characteristic morphological changes to the cell including condensed regions of nuclear material, internucleosomal DNA cleavage, and membrane blebbing (1). The fragmented chromatin, still membranebound, is phagocytosed by neighboring cells, thus avoiding an inflammatory response. Apoptosis is responsible for the removal of unwanted cells of many lineages and is thought to be an important safeguard against hyperplasia, which can be an early event in neoplasia (2).
Environmental stresses such as heat, radiation, and hypoxia; growth factors and ligands for surface receptors; and many drugs or chemical agents can induce apoptosis. Nevertheless, cells undergoing apoptosis exhibit a similar morphology, suggesting that these divergent apoptotic stimuli converge to trigger a common pathway of cell death. The common pathway involves a family of proteases known as the interleukin-1␤converting enzyme (ICE)-like 1 proteases or caspases, which are activated in a proteolytic cascade to cleave specific substrates (3,4). More recently, it has become evident that the transmission of signals from external stresses is accompanied by the activation of two families of kinases, the stress-activated protein kinases (SAPKs), also known as the c-Jun N-terminal kinases (JNKs), and the p38/HOG-1 kinases (5). Another possible signaling molecule in apoptotic pathways is the second messenger, ceramide. Increases in ceramide levels mediated by the activation of sphingomyelinase and consequent hydrolysis of sphingomyelin have been observed after exposure of cells to heat, radiation, TNF␣, and peroxide (6). In addition, exogenously added ceramide induces apoptosis (7). Thus, whereas the early signals appear to be specific to a particular stress or group of stresses, the subsequent events, such as the activation of the caspases and possibly the SAPKs and ceramide, are common events.
Heat shock proteins (HSPs) are a group of inducible proteins, some of which are constitutively expressed and increase in response to stress, whereas others are expressed only after stress (8). The constitutively expressed proteins act as chaperones for other cellular proteins, binding to nascent polypeptides to prevent premature folding and to translocate proteins into organelles (9). The induction of increased levels of the stress proteins is associated with the development of thermotolerance, a transient resistance to heat induced by prior exposure to mild heat or other stress agents (10,11). It is apparent that induced stress proteins can act to protect cells from stressinduced damage by preventing protein denaturation and/or by repairing such damage (12).
One major group of HSP is the HSP70 family that comprises a multi-gene family with at least 11 genes in humans (13) and 8 identified so far in the mouse (14). We and others have shown previously that induction of thermotolerance with a parallel induction of HSP inhibits heat-induced apoptosis in several cell lines (15,16). The consequences of exogenous expression of HSP72 in cells have been investigated by several groups. Expression of human HSP72 protects cells from heat stress (17)(18)(19), from some drugs (20), from the cytotoxic effects of TNF␣ (21), from monocyte killing (22), and from nitric oxide (23). In some of these studies, however, it was not demonstrated whether the cells died by apoptosis after these stresses.
As part of our investigation of pathways of stress-induced apoptosis, we have measured the effects of either induction of thermotolerance or constitutive expression of HSP72 in blocking apoptosis induced by external stresses. The protection provided by HSP72 against diverse apoptosis-inducing stresses that we have reported here and other data in the literature lead us to propose that the heat shock gene family represents another class of apoptosis-inhibiting genes.

EXPERIMENTAL PROCEDURES
Cell Cultures-Mouse embryo fibroblasts (MEF) derived from Balb/c mice were minimally transformed with E1A and Ras as described previously (24). To generate cells with constitutive HSP72 expression, clones were infected with the pMVH/human HSP72 proviral vector containing the 2.3-kilobase human HSP72 cDNA under the control of the cytomegalovirus promoter (25), a kind gift from Dr. G. Li. Cultures were single cell cloned, and HSP72 levels were determined by Western blotting. Except where specified, the high HSP72 expressing clone 2 cell line was used in experiments.
MEF were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum at 37°C in a humidified 5% CO 2 atmosphere. Immediately prior to each experiment, the medium was replaced with fresh pH equilibrated medium. All experiments were performed on 50% confluent cell cultures. For heat shock treatments, cells were immersed in a circulating water bath calibrated to 43 Ϯ 0.1°C for 10 -60 min. Lyophilized mouse TNF␣ (Sigma) was reconstituted in deionized H 2 O at 170 units/l and diluted into culture medium at concentrations ranging from 50 to 500 units/ml. Human TNF␣, a kind gift from Dr. A. Strasser, was diluted into culture medium at concentrations ranging from 100 to 1000 ng/ml. C 2 -ceramide (N-acetyl-D-sphingosine, Sigma) was initially dissolved in ethanol at 100 mM and aliquoted into serumfree Dulbecco's modified Eagle's medium before being diluted into culture medium. Concentrations of ceramide ranged from 25 to 100 M. Control cultures were supplemented with an equivalent amount of ethanol. Cells were subjected to ionizing radiation (5-15 Gy) from a cesium source (0.77 Gy/min). After treatment, cells were allowed to recover at 37°C for various times.
Apoptosis Assays-Nuclear morphology was assessed by fluorescence microscopy after propidium iodide staining. Briefly, flasks containing approximately 3 ϫ 10 6 cells were trypsinized, centrifuged at 1000 rpm, and resuspended in 200 l of phosphate-buffered saline containing 10 g/ml propidium iodide, 0.1% Triton X-100, and 0.1% sodium acetate. Samples were incubated on ice for 30 min in the dark before being scored for apoptotic morphology using a fluorescence microscope. A cell was scored as apoptotic if it displayed one or more of the following: nuclear margination, chromatin condensation, or formation of apoptotic bodies. At least 200 cells were counted in each experiment, and the data shown are the means and standard deviation of at least three independent experiments.
Western Analysis-HSP72 expression in cells was detected using a mouse monoclonal antibody specific for human HSP72 (N15, a kind gift from Dr. W. Welch) or with an antibody that detects both human and mouse HSP72 (SPA-810, StressGen Biotechnologies). The cleavage of poly(ADP-ribose) polymerase (PARP) from a 116-kDa protein to an 85-kDa fragment was detected using a mouse monoclonal antibody, C-2-10 (a kind gift from Dr. S. Kaufmann), which recognizes both the full-length and the 85-kDa fragment. Samples containing equal amounts of protein were subjected to SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose. The membranes were incubated with C-2-10, SPA-810, or N15 antibody, followed by horseradish peroxidase-conjugated rabbit anti-mouse IgG and ECL chemiluminescence for PARP or alkaline phosphatase-conjugated goat anti-mouse IgG and color staining for HSP72. As a loading control, membranes were immunostained with an anti-␣-tubulin antibody (clone B-5-1-2, Sigma) and detected by a color reaction using an alkaline phosphatase-conjugated secondary antibody.

Thermotolerance Protects Cells from Apoptosis after Heat
Shock, Ceramide, and TNF␣ but Not after Ionizing Radiation-MEF were treated with a mild heat shock (44°C for 10 min) to induce thermotolerance and allowed to recover at 37°C for 6 h before being subjected to more severe heat shock, ceramide, TNF␣, or ionizing radiation. Cells were scored for apoptotic morphology under a fluorescence microscope following permeabilization and staining with propidium iodide. As shown in Fig. 1A, cells in a thermotolerant state showed significant decreases in apoptosis after heat shock, ceramide, and TNF␣ but not after ionizing radiation. The thermotolerant state protected cells from apoptosis after a range of heat doses (Fig. 1B) or TNF␣ doses (Fig. 1C). The induction of thermotolerance induces high levels of HSP72 (Fig. 2).
Constitutive Expression of HSP72 Protects Cells from Apoptosis after Heat Shock, Ceramide, and TNF␣ but Not after Ionizing Radiation-Because the development of thermotolerance closely correlates with the induction of HSP72, it was decided to investigate whether constitutive expression of HSP72 alone could protect MEF from stress-induced apoptosis. MEF were infected with HSP72 and single cell cloned, and a number of clones expressing human HSP72 were isolated. The highest HSP72 expressing clone, clone 2, has less HSP72 than is measured in thermotolerant MEF but more than is expressed by clone 3 (Fig. 2). Control and HSP72 expressing cells were treated either with heat shock, ceramide, TNF␣, or ionizing radiation as described above, stained with propidium iodide, and scored for apoptotic morphology. As shown in Fig. 3, cells expressing constitutive HSP72 are protected from heat shock-, ceramide-, and TNF␣-induced apoptosis but not from ionizing radiation. Although we found that HSP72 does protect the cells from heat stress, we found greater protection from TNF␣-and ceramide-induced apoptosis. Also shown in Fig. 3 is the response of clone 2 cells to heat or TNF␣ after the induction of thermotolerance. The level of HSP72 increases further after induction of thermotolerance in clone 2 but is similar to the level induced in the parental MEF (Fig. 2). Cells expressing both high levels of HSP72 and thermotolerance are more resistant to heat stress than those expressing HSP72 alone (Fig.  3A), whereas the response to TNF␣ (Fig. 3B) is similar to that of HSP72 expressing cells. Thus the protection from heat-induced apoptosis afforded by the thermotolerant state cannot be attributed solely to the induction of HSP72, whereas protection from TNF␣-induced apoptosis in thermotolerant cells can be fully accounted for by the expression of HSP72.
The level of expression of HSP72 determined the extent of resistance to apoptosis. As shown in Fig. 3 (A and B), clone 3, which expresses less HSP72 than clone 2 (Fig. 2), has less resistance to heat-and TNF␣-induced apoptosis. In other retrovirally infected clones expressing less HSP72 than seen in clone 3, no protection against stress-induced apoptosis was detected (data not shown). In summary, protection was only seen in cells expressing moderate to high levels of HSP72, indicating that the process of infection with the retrovirus alone has not altered the response of the cells to stress.
The strong protection against TNF␣-induced apoptosis afforded by the expression of HSP72 led us to consider other members of the TNF receptor superfamily. To check whether the p55 or p75 TNF receptor was active in MEF, we measured the extent of apoptosis induced by human TNF␣. MEF were sensitive to human TNF␣, and the toxicity was blocked by HSP72 (Fig. 3E). This indicates that apoptosis is mediated through the p55 TNF receptor since the p75 TNF receptor does not bind human TNF␣. We also tested for sensitivity of MEF to apoptosis induced by binding of antibody or ligand to Fas or binding of Trail to the Trail receptor, other receptors in the TNF receptor family. MEF do not express Fas in amounts detectable by antibody binding, nor do they show signs of toxicity after incubation in either anti-Fas antibody, Fas ligand, or Trail ligand (data not shown).
Thermotolerance Reduces SAPK/JNK Activity after Heat Shock but Not after Ceramide, TNF␣, or Ionizing Radiation-The activation of the SAPK/JNK cascade detected after a number of cytotoxic insults has been proposed to be required for apoptosis (6). To determine whether the thermotolerant state was protecting MEF from stress-induced apoptosis through the inhibition of this pathway, SAPK/JNK activity was measured by an in vitro kinase assay using a GST-c-Jun (1-141) fusion protein as a substrate. Time courses were performed after each stress to determine optimum times for detecting SAPK/JNK activity. Maximal activity was detected at 30 min after heat shock and TNF␣. MEF show a 10 -15-fold activation of SAPK/ JNK after heat shock and 2-3-fold activation after TNF␣. Data shown for ceramide and ionizing radiation were obtained 1 h after stress; however, little SAPK/JNK activation was detected in MEF after ceramide and ionizing radiation over a 24-h period. Using these conditions, SAPK/JNK activity was measured in control and thermotolerant cells. Significant reduction of SAPK/JNK activity in thermotolerant cells after heat shock was observed, but no changes in activity were detected after other stresses (Fig. 4A).
Constitutive Expression of HSP72 Does Not Reduce SAPK/ JNK Activation after Stress-Because the thermotolerant state was shown to reduce SAPK/JNK activation after heat shock but not after other stresses, the effect of HSP72 expression on SAPK/JNK activity was measured. Control and HSP72 expressing cells were subjected to heat shock, ceramide, TNF␣, or ionizing radiation and SAPK/JNK activity measured. As shown in Fig. 4B, HSP72 did not reduce SAPK/JNK activation after any of these stresses, despite protecting the cells from apoptosis.
Thermotolerance Inhibits PARP Cleavage in Cells after Heat Shock, TNF␣, and Ceramide but Not after Ionizing Radiation-A common event in the apoptotic pathway is the activation of a caspase cascade. PARP, a DNA repair enzyme, has been identified as a substrate for caspase-3. Caspase-3 activity was measured by following the cleavage of PARP from a 116-kDa protein to an 85-kDa fragment. PARP cleavage is first FIG. 1. Effect of thermotolerance on apoptosis after heat, TNF␣, ceramide, and ionizing radiation. Thermotolerance (TT) was induced in MEF after a mild heat shock at 44°C for 10 min followed by 6 h of recovery at 37°C. A, cells were subsequently treated with either a severe heat shock (HS) at 44°C for 30 min, TNF␣ (500 units/ ml), ceramide (Cer) (100 M), or ionizing radiation (Irr) (10 Gy) and scored for apoptotic morphology after 16 h (heat, TNF␣, or ceramide) or 48 h (ionizing radiation). Using a Student's t test, the protection against heat, TNF␣, and ceramide was significant, with a p value less than 0.001. Dose response curves of apoptosis in parental (open circle) and thermotolerant (closed circle) MEF are shown after heat shock (B) and TNF␣ (C).

FIG. 2. Levels of expression of HSP72 in the parental MEF, HSP72 infected clones, thermotolerant MEF, and thermotolerant HSP72 clone 2 cells.
MEF were given a mild heat shock at 44°C for 10 min and allowed to recover for 6 h to induce thermotolerance. Proteins from parental MEF (Control), clones with constitutive HSP72 expression (Clone 2 and Clone 3), thermotolerant cells (TT), and thermotolerant clone 2 cells (Clone 2 TT) were separated by SDS-polyacrylamide gel electrophoresis and immunoblotted with a mouse monoclonal antibody that detects both mouse and human HSP72 (SPA-810) followed by an alkaline phosphatase-conjugated goat anti-mouse IgG. detected in MEF cells 6 h after stress but is more extensive at 16 h after heat, ceramide, and TNF␣ or 24 h after ionizing radiation (data not shown). We examined the possibility that the thermotolerant state may inhibit PARP cleavage either directly or by regulating some upstream event in the caspase cascade. MEF were treated with heat shock, ceramide, TNF␣, or ionizing radiation and allowed to recover for 16 or 24 h. Cell extracts were subjected to immunoblotting with a PARP-specific antibody that recognizes both the 116-and 85-kDa fragments. As shown in Fig. 5A, the induction of thermotolerance can reduce PARP cleavage after heat shock, TNF␣, and ceramide but not after ionizing radiation. Measurement of ␣-tubulin levels demonstrates equal protein loading in each lane.

Constitutive HSP72 Expression Inhibits PARP Cleavage in Cells after Heat Shock and TNF␣ but Not after Ceramide or
Ionizing Radiation-To determine whether HSP72 could mimic the effect of thermotolerance on PARP cleavage, MEF were treated with heat shock, ceramide, TNF␣, or ionizing radiation, and immunoblotting was performed on cell extracts. As found for the thermotolerant state, constitutive HSP72 expression reduced PARP cleavage in cells after heat shock and TNF␣ but not after exposure to ceramide or ionizing radiation (Fig. 5B). DISCUSSION Many forms of cellular stress lead to the activation of two related signaling pathways mediated by the stress-activated protein kinases, SAPK/JNK and p38/HOG-1 (5). The SAPK/ JNKs, which are activated by a kinase known as Sek1, comprise a family of eight or more isoforms that phosphorylate the c-Jun and JunD components of the AP1 transcription factor as well as Elk-1 and ATF-2. ATF-2 is also a substrate for p38/ HOG-1, as is MAPKAP kinase-2, which in turn phosphorylates the heat shock protein, HSP27, at the same sites at which it is phosphorylated in response to stress (5). Many stresses capable of activating SAPK/JNK, including heat shock, TNF␣, ceramide, UV light, ionizing radiation, osmotic stress, and anti-cancer agents, also result in apoptosis (reviewed in Ref. 5). Despite the observation that SAPK/JNK activation and apoptosis are often co-incident, there is strong debate about the requirement of SAPK/JNK activation for apoptosis. Stable transfection of dominant negative mutants of either Sek1 or c-Jun (TAM-67) has been shown to inhibit apoptosis after ionizing radiation, ceramide, UVC light, heat shock, and hydrogen peroxide (6). In contrast, studies performed in Sek1 (Ϫ/Ϫ) ES cells demonstrated a normal apoptotic response after ionizing radiation, UV light, osmotic stress, serum deprivation, anisomycin, heat shock, dexamethasone, and anti-cancer agents, although thymocytes from Sek1 (Ϫ/Ϫ) mice showed impaired Fas-induced apoptosis (27). Controversy surrounds the involvement of SAPK/JNK activation in TNF␣-induced apoptosis, with studies claiming that a catalytically inactive MAPKKK (ASK1), which no longer activates SAPK/JNK and p38/HOG-1 (28), can inhibit TNF␣-induced apoptosis, as can either Sek1 or TAM67 dominant negative mutants (6). In contrast, other studies have shown that SAPK/JNK activation occurs through a pathway that is not required for TNF␣-induced apoptosis (29,30).
In our study, we have shown that the thermotolerant state or constitutive expression of HSP72 can protect cells from apoptosis induced by heat, TNF␣, and ceramide but not from ionizing radiation, thus demonstrating the existence of multiple pathways of stress-induced apoptosis. It is interesting to note that HSP72, which is widely believed to protect cells from heat stress, prevents heat-induced apoptosis to a lesser extent than TNF␣-or ceramide-induced apoptosis. Further, the protection seen in thermotolerant cells after TNF␣ and ceramide can be accounted for by the induction of HSP72, whereas the induction of thermotolerance provides a significantly higher level of protection against heat-induced apoptosis than can be accounted for by HSP72 alone. An explanation for this may lie in the fact that heat is a nonspecific stress that causes damage to many proteins and organelles in the cell, and thus multiple targets need to be protected from heat stress. HSP72 may be able to protect some of these targets, but the induction of other stress proteins may be required for more complete protection. In contrast, TNF␣ initiates apoptosis through a specific pathway that begins with the activation of TNF receptors. HSP72 could act at one stage of this pathway to minimize the extent to which apoptosis is triggered by TNF␣.
The possibility that HSP72 may regulate stress-induced apoptosis through inhibition of the SAPK/JNK pathway was investigated, but expression of HSP72 was found to not alter the extent of activation of SAPK/JNK. In addition, we found that the extent of SAPK/JNK activation after the different stresses did not correlate with the extent of apoptosis observed, possibly indicating that SAPK/JNK activation is not required for the induction of apoptosis in MEF. Alternatively, SAPK/JNK activation may be necessary but is not sufficient for apoptosis to occur. We therefore conclude either that MEF can undergo stress-induced apoptosis using a SAPK/JNK-independent Cell extracts were subjected to immunoblotting with a PARP-specific antibody that recognizes both the 116-and 85-kDa fragments. Shown are Western blots visualized by ECL chemiluminescence. Membranes were also probed with a mouse anti-tubulin antibody and visualized with an alkaline phosphatase color reaction as a control for protein loading.
pathway that can be regulated by HSP72 or that HSP72 acts downstream of SAPK/JNK activation. Interestingly, SAPK/ JNK was inhibited in thermotolerant cells after heat shock but not in cells expressing constitutive HSP72. This supports the finding that thermotolerance provides a substantially higher level of protection against heat-induced apoptosis than HSP72 alone, suggesting that factors other than HSP72 that are induced in the thermotolerant state are capable of suppressing SAPK/JNK activation.
An expanding family of cysteine proteases (caspases), of which ICE is the prototype, has been shown to play a key role in mammalian cell apoptosis. The caspases can be divided into three main groups based on sequence similarity: ICE, ICErel 11/Tx/ICH-2 and ICErel 111/Ty; CPP32/Yama/apopain/prICE, ICE-LAP3/MCH3 and MCH2; and finally, Nedd2/ICH-1 (4). Overexpression of any one of these enzymes induces apoptosis in transfected cells (31,32). Evidence is emerging that multiple caspases may function sequentially after induction of apoptosis in a cell (33). A number of substrates have now been identified, including other caspases, proteins involved in DNA repair (PARP and DNA-protein kinase) and some structural proteins (lamin, actin, and vimentin) (4), but the importance of some of these substrates in apoptosis is not clear. We investigated the possibility that HSP72 could inhibit stress-induced apoptosis through the regulation of the caspase pathway. Caspase-3 activity can be measured by following the cleavage of PARP from a 116-kDa protein to an 85-kDa fragment. We have shown that constitutive HSP72 expression can prevent PARP cleavage after heat shock or TNF␣ but not after ceramide or ionizing radiation. The lack of protection of PARP by HSP72 after ceramide exposure is surprising because HSP72 does protect against ceramide-induced apoptosis. The data suggest that HSP72 acts to block the activation or the activity of one of the caspases and/or to protect the substrates of these caspases from proteolytic degradation. The possibility that HSP72 acts as a decoy substrate for the caspases was investigated by Mosser et al. (19), but they found that HSP72 was unable to inhibit the caspase-3-mediated cleavage of PARP in an in vitro assay.
Of interest in this study was the marked protection by HSP72 from TNF␣-induced apoptosis, mediated through the p55 TNF receptor. While triggering apoptosis in many cells, TNF␣ also activates the transcription factor NF-B that blocks apoptosis (30). How HSP72 may interact with this pathway is unknown, but in WEHI-S cells, the expression of HSP72 does not alter the TNF-induced activation of NF-B, nor does it alter TNF receptor levels (34). Instead, HSP72 inhibits TNF-induced activation of phospholipase A 2 , which releases arachidonic acid from membrane phospholipids (34). It has been shown previously that arachidonic acid is released in response to TNF in sensitive cells but not in TNF-resistant cells (35).
Two other families of genes that block apoptosis have been identified. The inhibitor of apoptosis protein family, first recognized as the candidate gene for spinal muscular atrophy, can block apoptosis induced by a number of stresses (36 -38). The best characterized inhibitors of apoptosis are Bcl-2 and several of its homologues, which can block apoptosis in response to a large number of, but not all, stimuli (39). The mechanism by which Bcl-2 and its homologues block apoptosis is not clear; however it has been suggested that Bcl-X L prevents disruption of the mitochondrial membrane potential that otherwise permits the release of apoptosis-inducing proteins into the cytosol (40). Bcl-2 blocks the release of cytochrome c from the mitochondria that occurs following apoptotic stimuli but prior to membrane depolarization. This in turn prevents the interaction between cytochrome c and Apaf-1 and subsequent caspase activation and apoptosis (41,42). It is feasible that HSP72 may protect cells from apoptosis through a similar mechanism; either through the inhibition of cytochrome c release from the mitochondria, or, more likely, by binding directly to cytosolic cytochrome c and preventing activation of the caspases. HSP72 is known to bind to peptides derived from cytochrome c (43).
Heat shock proteins have been recognized for some years to protect cells from heat damage and more recently from some other cytotoxic insults. As well as the protective effects of HSP72 shown here and in other studies (17)(18)(19)(20)(21)(22)(23) the small heat shock protein, HSP27, has been shown to block apoptosis induced by Fas/Apo-1 and staurosporine (44) and by some anticancer drugs (20,45). Heat shock proteins are often expressed at high levels in tumor cells, leading to suggestions that they can protect cells either from immune attack or from the type of therapy being administered. For example, several recent studies have revealed that tumor specimens from patients whose cancer therapy has resulted in a short disease-free survival contain high levels of one or more of the stress proteins (46 -49). Emerging clones of tumor cells would be subjected to host initiated stresses such as TNF and may show prolonged survival if they express high levels of HSP72. Thus, we believe there is evidence to suggest that high levels of heat shock proteins may confer a survival advantage on an emerging or metastasizing clone of neoplastic cells by protecting it from immune surveillance and other host-derived stresses. We therefore propose that the heat shock gene family represents a third class of anti-apoptosis genes.