Pivotal role of reactive oxygen species as intracellular mediators of hyperthermia-induced apoptosis.

The effects of cellular antioxidant capacity on hyperthermia (HT)-induced apoptosis and production of antiapoptotic heat shock proteins (HSPs) were investigated in HL-60 cells and in HL-60AR cells that are characterized by an elevated endogenous catalase activity. Exposure of both cell lines to 43 degrees C for 1 h initiated apoptosis. Apoptosis peaked at 3-6 h after heat exposure in the HL-60 cells. Whereas HL-60AR cells were partially protected against HT-induced apoptosis at these early time points, maximal levels of apoptosis were detected later, i.e. 12-18 h after heat exposure. This differential induction of apoptosis was directly correlated to the induction of the antiapoptotic HSP27 and HSP70. In particular, in the HL-60 cells HSP27 was significantly induced at 12-18 h after exposure to 43 degrees C when apoptosis dropped. In contrast, coinciding with the late onset of apoptosis in HL-60AR cells at that time HL-60AR cells lacked a similar HSP response. In line with the higher antioxidant capacity HL-60AR cells accumulated reactive oxygen species to a lesser degree than HL-60 cells after heat treatment. Protection from HT-induced apoptosis as well as diminished heat-induced HSP27 expression was also observed after cotreatment of HL-60 cells with 43 degrees C and catalase but not with superoxide dismutase. These data emphasize the pivotal role of reactive oxygen species for HT induced pro- and antiapoptotic pathways.

The effects of cellular antioxidant capacity on hyperthermia (HT)-induced apoptosis and production of antiapoptotic heat shock proteins (HSPs) were investigated in HL-60 cells and in HL-60AR cells that are characterized by an elevated endogenous catalase activity. Exposure of both cell lines to 43°C for 1 h initiated apoptosis. Apoptosis peaked at 3-6 h after heat exposure in the HL-60 cells. Whereas HL-60AR cells were partially protected against HT-induced apoptosis at these early time points, maximal levels of apoptosis were detected later, i.e. 12-18 h after heat exposure. This differential induction of apoptosis was directly correlated to the induction of the antiapoptotic HSP27 and HSP70. In particular, in the HL-60 cells HSP27 was significantly induced at 12-18 h after exposure to 43°C when apoptosis dropped. In contrast, coinciding with the late onset of apoptosis in HL-60AR cells at that time HL-60AR cells lacked a similar HSP response. In line with the higher antioxidant capacity HL-60AR cells accumulated reactive oxygen species to a lesser degree than HL-60 cells after heat treatment. Protection from HT-induced apoptosis as well as diminished heat-induced HSP27 expression was also observed after cotreatment of HL-60 cells with 43°C and catalase but not with superoxide dismutase. These data emphasize the pivotal role of reactive oxygen species for HT induced pro-and antiapoptotic pathways.
Hyperthermia (HT) 1 has a potential as an antineoplastic treatment modality when combined with radiation or chemotherapy (1). Thus, the biological effects of elevated temperatures have been studied extensively. Cell death is apparent after application of a critical temperature load (2,3). Previously, we and others were able to demonstrate that HT-induced cell death can at least partly be attributed to the induction of apoptosis (2)(3)(4). Despite this effect, exposure of malignant cells to elevated temperatures also elicits a well regulated cellular defense including the production of HSPs (5,6). In particular, induction of HSP27 and HSP70 has been shown to exert the protective role of HSP by inhibition of apoptosis (7)(8)(9)(10)(11). The exact mechanism(s) of heat-induced pro-and antiapoptotic pathways, however, are currently not entirely clear.
Several studies indicate the involvement of heat-mediated oxidative stress in HT-induced cytotoxicity. After exposure to heat-increased levels of superoxide anions, hydrogen peroxide and nitric oxide as well as increased lipid peroxidation products have been found in various cell lines and in tumor tissue (12)(13)(14)(15)(16). The origin of the ROS, however, remains to be determined. Noteworthy, ROS are also known to induce HSPs that are critical for cellular thermoresistance and the development of thermotolerance (2,17,18). Thus, heat-induced oxidative stress seems to play a pivotal role, i.e. the induction of apoptotic cell death and induction of antiapoptotic HSPs. Via manipulation of the activity of antioxidant enzymes one should be enabled to determine the importance of ROS for heat-mediated apoptosis and the induction of cellular defense. SOD, CAT, and glutathione peroxidase are the main cellular ROS degrading enzyme systems. SOD converts the superoxide radical (O 2 . ) into hydrogen peroxide, which is metabolized by CAT and glutathione peroxidase. Therefore, the aim of this study was to investigate whether the antioxidative capacity is affecting HT-mediated apoptosis and HSP production. To this end we investigated HT-induced apoptosis in the parental leukemia cell line HL-60 as well as in HL-60AR cells that are characterized by elevated endogenous CAT activity (19,20) and the modulation of HT-induced apoptosis and HSP induction by exogenous treatment with CAT or SOD.

MATERIALS AND METHODS
Chemicals and Antibodies-The mouse anti-Bcl-2, mouse anti-Bcl-X S/L , and rabbit anti-Bax antibodies as well as the appropriate horseradish peroxidase-labeled secondary antibodies were obtained from Santa Cruz (Santa Cruz, CA) and used according to the manufacturer's recommendations. The mouse monoclonal anti-HSP27 and anti-HSP70 antibodies were purchased from Stressgen (Victoria, Canada). CAT, SOD, and all other chemicals were obtained from Sigma.
Cell Lines and Culture Conditions-The human leukemia cell lines HL-60 and HL-60AR were a kind gift from Dr. K. Bhalla (Medical University of South Carolina, Charleston, SC). Endogenous CAT enzyme activity in HL-60 and HL-60AR cells was determined to be 53.3 Ϯ 17.8 and 176.3 Ϯ 55.8 mol min Ϫ1 mg protein Ϫ1 .The cells were grown in RPMI 1640 with 10% fetal bovine serum in a humidified 5% CO 2 in air atmosphere at 37°C. For heat exposure cells were seeded in 25-cm 2 tissue culture flasks, allowed to grow for 24 h, and then exposed to 43°C for 1 h. After heat treatment, cells were returned to 37°C. Cell viability was determined using the trypan blue exclusion test. Before each experiment cells were determined to have a viability of Ͼ90%. Time zero was considered the point at which the cells were removed from the water bath after heat treatment. All experiments were repeated at least three times. Data shown in the graphs for various parameters represent the means Ϯ S.D., and data sets were compared using the Student's t test.
Temperature Treatment-The 25-cm 2 tissue culture flasks were placed into plastic freezer bags and filled with 5% CO 2 /air before sealing. The flasks were placed into a gently shaking water bath maintained at 43°C Ϯ 0.1°C for 1 h. The medium in the flasks equilibrated * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
to the water bath temperature in about 5 min. The temperature in the water bath was controlled using a thermistor temperature probe, which was regularly calibrated to a tenth degree in the range of 20 -50°C. After the heat treatment the flasks were removed from the freezer bags and returned to the 37°C incubator.
Quantitation of DNA Fragmentation-Extent of DNA fragmentation was determined by a modified method described in Ref. 21. At different time points after heat exposure cells were collected and resuspended in 100 l of lysis buffer (5 mM Tris-HCl, pH 7.4, 1 mM EDTA, 0.5% Triton-X). After incubation for 20 min at 4°C, the suspension was centrifuged at 14,000 rpm for 20 min, and subsequently the fragmented DNA was recovered from the supernatant. The pellets were sonicated for 15 s in 100 l of lysis buffer. The amount of DNA in both fractions was determined by a fluorometric method using 10 g/ml DAPI as described in Ref. 22. The fluorescence intensity was measured using 360-nm excitation wavelength and 460-nm emission wavelength (Cyt-oFluor TM 2350 Millipore, Eschborn, Germany). The percentage of DNA fragmentation was defined as the ratio of the amount of fragmented DNA to the total amount of DNA.
Preparation of Cell Lysates and Determination of DEVD-AMC Cleavage-Cells were exposed to 43°C for 1 h. At frequent intervals thereafter, cells were collected by centrifugation and resuspended in lysis buffer (100 mM Hepes, pH 7.5, 10% sucrose, 0.1% CHAPS, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride). Cell lysate (30 g of protein/ sample) was incubated with 12 M of the fluorogenic peptide substrate DEVD-AMC in a 96-well microtiter plate at room temperature. The cleavage of DEVD-AMC was monitored by AMC liberation in a fluoroscan plate reader (CytoFluor TM 2350, Millipore) using 360-nm excitation and 460-nm emission wavelength. Fluorescence was measured every 60 s during a 60-min period, and fluorescence units were converted to pmol amounts of AMC using a standard curve generated with free AMC (Sigma). Control experiments confirmed that the release of substrate was linear with DEVD-AMC and protein concentration. Addition of the competitive inhibitor DEVD-CHO (50 nM) to the samples blocked DEVD-AMC cleavage.
Quantitation of Apoptotic Cells-At various time points after heat treatment, cells were treated with 3.7% paraformaldehyde for 10 min followed by fixation onto glass slides with 70% ethanol. After rehydration glass slides were stained with Hoechst dye (1.5 g/ml) for 5 min, washed with phosphate-buffered saline, and mounted in 2.5% Dabco (1,4-diazabicylo[2.2.2]octane). Nuclei were visualized using a Zeiss fluorescence microscope. Apoptotic nuclei were expressed as a percentage of total nuclei.
Determination of Intracellular Generation of ROS-The production of intracellular ROS was estimated fluorometrically using the oxidation sensitive fluorescent probe 2Ј,7Ј-dichlorofluorescein diacetate (DCF; Sigma). Before heat exposure, cells were washed with phosphate-buffered saline and incubated with 5 M DCF for 30 min at 37°C. Subsequently cells were washed and resuspended in DCF free RPMI 1640. Cells were then exposed to 37 or 43°C for 1 h. At frequent time intervals thereafter, samples were analyzed using a CytoFluor TM 2350, Millipore fluorometer. Fluorescence of DCF was detected at an excitation and emission wavelengths of 485 and 530 nm, respectively.
RNA Extraction and Northern Blot Analysis-Total RNA was isolated according to the method described in Ref. 23. 15 g of total RNA were subjected to electrophoresis in denaturing 1% agarose gels containing 0.7 mol/liter formaldehyde. RNA was transferred onto nylon membranes (Nytran Plus, Schleicher & Schü ll) with a vacuum blotting apparatus (Amersham Pharmacia Biotech). Filters were cross-linked with UV light dried at 80°C for 2 h and prehybridized for 4 h at 42°C in 45% formamide, 5ϫ SSC, 5ϫ Denhardt's solution, 0.1% SDS, and 100 g/ml sonicated denatured salmon testis DNA. Hybridizations were performed in fresh solution supplemented with the radioactive probe (0.5 to 2 ϫ 10 6 cpm/ml) for 2 days at 42°C. Hybridization probes were polymerase chain reaction generated fragments. Polymerase chain re-action fragments were [␣-32 P]dCTP-labeled with a commercially available kit (MBI Fermentas, St. Leon-Rot, Germany). After hybridization, the filters were washed twice for 15 min in 2ϫ SSC/0.1%SDS at 50°C and then twice in 0.1ϫ SSC/0.1% SDS at 60°C. Filters were sealed in plastic bags and analyzed by exposure to imaging plates for a Bio Imaging Analyzer (BAS 1000; Fuji, Dü sseldorf, Germany).

Hyperthermia-induced Apoptosis in HL-60 and HL-60AR
Cells-Exposure of HL-60 to 43°C for 1 h resulted in significant apoptotic cell death as demonstrated by DNA fragmentation, caspase 3 activation, and quantitation of apoptotic nuclei after Hoechst staining (Fig. 1). To test for the involvement of ROS in HT-induced apoptosis, HL-60 cells were incubated with CAT (500 units/ml) or SOD (50 units/ml) during heat exposure. As shown in Fig. 2 treatment of HL-60 cells with heat in the presence of CAT but not SOD partially protected from HTinduced apoptosis 3 h after heat exposure. Moreover, modulation of apoptosis by antioxidant enzymes was demonstrated by heat exposure of the parental HL-60 cells and the HL-60AR cells, for which an increased endogenous CAT enzyme activity has been described (also see "Materials and Methods"). In the HL-60 cells, apoptosis peaked at 3-6 h and disappeared by 18 -24 h after treatment, whereas HL-60AR cells showed a late onset of apoptosis. Peak apoptosis was not observed until 12-18 h after exposure to 43°C for 1 h (Fig. 1).
Hyperthermia-mediated Generation of Intracellular ROS-Determination of intracellular ROS in HL-60 and HL-60AR cells after exposure to 43°C for 1 h confirmed the accumulation of ROS in response to heat as described earlier by others (Ref. 14 and Fig. 3). However, hyperthermia induced a significantly higher accumulation of ROS in HL-60 cells compared with HL-60AR cells. This is in line with the higher antioxidative capacity of the HL-60AR cells via their increased endogenous CAT enzyme activity (20).
Expression of Members of the Bcl-2 Family as Well as HSP27 and HSP70 after Heat Treatment-Members of the Bcl-2 protein family play an important role for pro-but also antiapoptotic pathways. Interestingly, basal Bcl-2 protein expression was significantly higher in the HL-60AR cells than in the parental HL-60 cells. In contrast, basal expression of Bcl-X L and Bax was not altered in the HL-60AR cells. Treatment of HL-60 and HL-60AR cells with 43°C did not affect the protein levels of the antiapoptotic Bcl-2 and Bcl-X L as well as the proapoptotic Bax at any time (Fig. 4A). In contrast, exposure of both cell lines to 43°C induced HSP27 and HSP70 mRNA and protein levels (Fig. 4), but the response markedly differed between HL-60 and HL-60AR cells. Following exposure to 43°C for 1 h HSP70 mRNA levels in HL-60 cells peaked at 3 h after treatment (Fig. 4B). In contrast, in HL-60AR cells the induction of HSP70 mRNA was delayed for 3 h and reached its maximum 6 h after treatment. Similar results were obtained for HSP27. HL-60 cells presented with detectable basal HSP27 mRNA and protein levels and a rapid induction after exposure to heat. Both the induction of HSP27 and HSP70 in the HL-60 cells coincided with the drop of hyperthermia-induced apoptosis. In contrast, in the HL-60AR cells faint basal HSP27 mRNA and protein levels were only moderately induced after exposure to heat. To further determine the effects of the cellular redox status on HSP27 expression, HL-60 cells were treated with CAT or SOD in addition to HT. As demonstrated in Fig. 5, heat-induced HSP27 induction was diminished by co-treatment of HL-60 cells with heat and CAT but not with SOD. DISCUSSION In the present study we found that HT exerted its cytotoxic effect via induction of apoptosis in the human leukemia cells HL-60. These data are consistent with previous reports using HL-60 cells and other human leukemia cell lines (2)(3)(4). We further demonstrate that the induction of apoptotic cell death by heat was modulated by the antioxidant capacity of the cells. CAT, glutathione peroxidase, and SOD are the main cellular ROS degrading enzyme systems. SOD converts the superoxide radical into hydrogen peroxide, which is metabolized by CAT and glutathione peroxidase. Treatment of HL-60 cells with CAT conferred partial protection against the induction of heatinduced apoptosis, whereas SOD did not affect the heat-induced apoptotic cell death. Thus, with respect to HT-induced apoptosis, hydrogen peroxide but not superoxide appears to be of greater importance. In line with this assumption, different effects of hydrogen peroxide and superoxide anions on apoptosis are reported from studies with vascular smooth muscle cells and neutrophils (24).
ROS degrading enzymes like CAT and SOD are poorly cell permeable, and nonphysiological concentrations are needed to affect intracellular ROS concentrations. Although previous reports demonstrate that exogenous delivery of ROS degrading enzymes is sufficient to affect intracellular ROS concentrations (25)(26)(27), data from cells with endogenously elevated CAT activity would be much more convincing. Therefore, to further test the hypothesis of whether hydrogen peroxide is involved in HT-induced apoptosis, we made use of the closely related parental leukemia cell line HL-60 and its subline HL-60AR. HL-60 AR cells have increased endogenous CAT enzyme levels and activity (19). In line with the experiments in which exogenous ROS degrading enzymes were used to modulate intracellular ROS levels, the induction of apoptosis was significantly different in HL-60 and the HL-60AR cells. Although apoptosis HL-60 and HL-60AR cells were exposed to 37 or 43°C for 1 h. At frequent time intervals thereafter, intracellular generation of ROS was determined using the fluorescent probe DCF as described under "Materials and Methods." was seen as early as 3-6 h after exposure to 43°C for 1 h in the HL-60 cells, HL-60AR cells seemed to be partially protected against HT-induced apoptotic cell death at this time point. The partial resistance of HL-60AR cells at this time point was correlated to significantly diminished HT-induced ROS levels in HL-60AR cells, possibly because of the higher antioxidative capacity of the HL-60AR cells. The origin of the ROS has not yet been identified and needs further investigation. ROS are produced when oxygen is consumed in the electron transport chain reaction, and mitochondria can be a major source of ROS; ROS production by mitochondria during apoptosis has already been considered by several investigators (28 -31).
However, assuming that HL-60 and HL-60AR cells differ in their levels of hydrogen peroxide through higher CAT activity in HL-60AR cells, this difference does not explain the observation that 12-18 h after heat exposure when apoptosis dropped in the HL-60 cells maximal levels of apoptotic cell death were detected in HL-60AR cells. To gain insight into the mechanism for this late onset of apoptosis, different proteins with regulatory function in apoptotic pathways were investigated. Important general regulators of apoptotic cell death are the Bcl-2 proteins and the HSPs, i.e. HSP27 and HSP70 (32)(33)(34)(35). Whereas Bcl-2 proteins affect the apoptotic pathway via interaction with the mitochondrial membrane potential, cytochrome c, and finally the caspases, HSPs seem to exert their apoptotic effect downstream of caspase 3-like proteases (11,36,37). Basal Bcl-2 expression was significantly higher in the HL-60AR cells compared with the HL-60 cells. After exposure to heat, however, protein expression of Bcl-2, Bcl-X L , and Bax was not significantly altered. Thus, it seems that these members of the Bcl-2 family are not involved in the HT-mediated induction of apoptosis. These data are in agreement with earlier reports of our group and others (3,4). In contrast, HT-dependent induction of HSP27 and HSP70 directly correlated with the apoptotic effect in HL-60 and HL-60AR cells. Whereas HSP27 and HSP70 mRNA and protein expression was significantly increased after exposure to heat in the HL-60 cells, the induction of HSP27 in particular was markedly diminished in the HL-60AR cells. This differential HSP expression in the HL-60 and HL-60AR cells implies that the HT-induced increase of HSPs is dependent on the accumulation of ROS. Strong support for this assumption was the observation that HSP27 expression was reduced by CAT. Taken together and in agreement with earlier reports (2,17,38), our data support the notion that the induction of HSP27 is dependent on the cellular redox status. However, because the influence of the increased CAT activity was more pronounced for HSP27 compared with HSP70, one has to assume a higher sensitivity of HSP27 by the cellular redox status. A differential expression of HSP27 and HSP70 is in line with reports of others and may be due to specific transcriptional regulatory elements (2).
Transcriptional regulation of heat shock genes requires the activation of the HSF-1, which binds to the consensus heat shock element located in the promoter region of the heat shock genes. Activation of HSF-1 requires phosphorylation, trimerization, and nuclear translocation. Recent studies suggest that hydrogen peroxide increases nuclear translocation and DNA binding activity of HSF-1 (39,40), which is in line with our observation reported here and makes a redox mechanism in heat-induced signal transduction pathways during apoptosis very likely.
Understanding of the regulation of HSP induction may be important for anticancer strategies, because it has been demonstrated that HSP27 can limit the efficacy of antitumor agents and enhance tumorigenicity (41,42). Induction of HSPs in the HL-60 cells coincided with the drop of HT-induced apoptosis. HSPs may be induced as a cellular defense, which would support the concept of an inverse relationship of HSP induction and HT-induced cytotoxicity (43). As a consequence of the above hypothesis, the lack of protection via diminished HSP27 and, to lesser degree, HSP70 induction in the HL-60AR cells is sensitizing these cells to undergo apoptosis at later time points.
Our data underscore the role of ROS in HT-induced apoptosis. However, the role of HT-induced accumulation of ROS appears to be ambiguous. Although a decrease in HT-mediated FIG. 4. Changes in protein (a) and mRNA (b) expression after exposure of HL-60 and HL-60AR cells to 43°C for 1 h. HL-60 and HL-60AR cells were treated with 43°C for 1 h. a, expression of Bcl-2, Bcl-X L , Bax, HSP27, and HSP70 protein expression as a function of temperature treatment in HL-60 and HL-60AR cells. Cellular protein was extracted from HL-60 and HL-60AR cells up to 24 h after exposure to 43°C for 1 h. b, expression of HSP27 and HSP70 mRNA expression as a function of temperature treatment in HL-60 and HL-60AR cells. HSP27 and HSP70 mRNA were detected as described under "Materials and Methods." FIG. 5. Modulation of heat-induced HSP27 and HSP 70 expression by antioxidant enzymes. HL-60 cells were exposed to 43°C in the absence or presence of CAT (500 -1500 units/ml) or SOD (50 -150 units/ml) for 1 h. After heat treatment, cells were washed and resuspended in fresh medium in the absence of ROS modulating molecules. 6 h later cellular protein was extracted and analyzed for HSP27 and HSP70 expression. For comparison, extracts from HL-60AR cells 6 h after exposure to 43°C were loaded into the second lane from the left. Equal protein loading per lane was checked by probing for ␤-actin. ROS generation via a higher endogenous antioxidative capacity confers partial resistance against HT-induced apoptosis at first, it might also affect the cellular defense, i.e. HSP production. Thus, ROS represent a common denominator of both proand antiapoptotic HT-induced pathways.