Heat shock proteins and Bcl-2 expression and function in relation to the differential hyperthermic sensitivity between leukemic and normal hematopoietic cells.

A major problem in autologous stem cell transplantation is the occurrence of relapse by residual neoplastic cells from the graft. The selective toxicity of hyperthermia toward malignant hematopoietic progenitors compared with normal bone marrow cells has been utilized in purging protocols. The underlying mechanism for this selective toxicity has remained unclear. By using normal and leukemic cell line models, we searched for molecular mechanisms underlying this selective toxicity. We found that the differential heat sensitivity could not be explained by differences in the expression or inducibility of Hsp and also not by the overall chaperone capacity of the cells. Despite an apparent similarity in initial heat-induced damage, the leukemic cells underwent heat-induced apoptosis more readily than normal hematopoietic cells. The differences in apoptosis initiation were found at or upstream of cytochrome c release from the mitochondria. Sensitivity to staurosporine-induced apoptosis was similar in all cell lines tested, indicating that the apoptotic pathways were equally functional. The higher sensitivity to heat-induced apoptosis correlated with the level of Bcl-2 protein expression. Moreover, stable overexpression of Bcl-2 protected the most heat sensitive leukemic cells against heat-induced apoptosis. Our data indicate that leukemic cells have a specifically lower threshold for heat damage to initiate and execute apoptosis, which is due to an imbalance in the expression of the Bcl-2 family proteins in favor of the proapoptotic family members.

Dynamic changes in the localization of thermally unfolded nuclear proteins associated with chaperone-dependent protection.

Molecular chaperones are involved in the protection of cells against protein damage through their ability to hold, disaggregate, and refold damaged proteins or their ability to facilitate degradation of damaged proteins. Little is known about how these processes are spatially coordinated in cells. Using a heat-sensitive nuclear model protein luciferase fused to the traceable, heat-stable enhanced green fluorescent protein (N-luc-EGFP), we now show that heat inactivation and insolubilization of luciferase were associated with accumulation of N-luc-EGFP at multiple foci throughout the nucleus. Coexpression of Hsp70, one of the major mammalian chaperones, reduced the formation of these small foci during heat shock. Instead, the heat-unfolded N-luc-EGFP accumulated in large, insoluble foci. Immunofluorescence analysis revealed that these foci colocalized with the nucleoli. Time-lapse analysis demonstrated that protein translocation to the nucleolus, in contrast to the accumulation at small foci, was fully reversible upon return to the normal growth temperature. This reversibility was associated with an increase in the level of active and soluble luciferase. Expression of a carboxyl-terminal deletion mutant of Hsp70(1-543), which lacked chaperone activity, had no effect on the localization of N-luc-EGFP, which suggests that the Hsp70 chaperone activity is required for the translocation events. Our data show that Hsp70 not only is involved in holding and refolding of heat-unfolded nuclear proteins but also drives them to the nucleolus during stress. This might prevent random aggregation of thermolabile proteins within the nucleus, thereby allowing their refolding at the permissive conditions and preventing indirect damage to other nuclear components.

Modulation of in vivo HSP70 chaperone activity by Hip and Bag-1.

The chaperone activity of Hsp70 is influenced by the activities of both positive and negative regulatory proteins. In this study, we provide first time evidence for the stimulating effect of the Hsp70-interacting protein Hip on the chaperone activity in the mammalian cytosol. Overexpressing Hip enhances the refolding of the heat-inactivated reporter enzyme luciferase expressed in hamster lung fibroblasts. Also, it protects luciferase from irreversible denaturation under conditions of ATP depletion. We demonstrate that these stimulating actions depend on both the presence of the central Hsp70-binding site and the amino-terminal homo-oligomerization domain of Hip. The carboxyl terminus (amino acids 257-368) comprising the 7 GGMP repeats (Hsc70-like domain) and the Sti1p-like domain are dispensable for the Hip-mediated stimulation of the cellular chaperone activity. Bag-1, which inhibits the Hsp70 chaperone activity both in vitro and in vivo, was found to compete with the stimulatory action of Hip. In cells overexpressing both Hip and Bag-1, the inhibitory effects of Bag-1 were found to be dominant. Our results reveal that in vivo a complex level of regulation of the cellular chaperone activity exists that not only depends on the concentration of Hsp70 but also on the concentration, affinity, and intracellular localization of positive and negative coregulators. As the Hsp70 chaperone machine is also protective in the absence of ATP, our data also demonstrate that cycling between an ATP/ADP-bound state is not absolutely required for the Hsp70 chaperone machine to be active in vivo.

Bag1 functions in vivo as a negative regulator of Hsp70 chaperone activity.

Studies on the Hsp70 chaperone machine in eukaryotes have shown that Hsp70 and Hsp40/Hdj1 family proteins are sufficient to prevent protein misfolding and aggregation and to promote refolding of denatured polypeptides. Additional protein cofactors include Hip and Bag1, identified in protein interaction assays, which bind to and modulate Hsp70 chaperone activity in vitro. Bag1, originally identified as an antiapoptotic protein, forms a stoichiometric complex with Hsp70 and inhibits completely Hsp70-dependent in vitro protein refolding of an unfolded polypeptide. Given its proposed involvement in multiple cell signaling events as a regulator of Raf1, Bcl2, or androgen receptor, we wondered whether Bag1 functions in vivo as a negative regulator of Hsp70. In this study, we demonstrate that Bag1, expressed in mammalian tissue culture cells, has pronounced effects on one of the principal activities of Hsp70, as a molecular chaperone essential for stabilization and refolding of a thermally inactivated protein. The levels of Hsp70 and Bag1 were modulated either by transient transfection or conditional expression in stably transfected lines to achieve levels within the range detected in different mammalian tissue culture cell lines. For example, a twofold increase in the concentration of Bag1 reduced Hsp70-dependent refolding of denatured luciferase by a factor of 2. This effect was titratable, and higher levels of wild-type but not a mutant form of Bag1 further inhibited Hsp70 refolding by up to a factor of 5. The negative effects of Bag1 were also observed in a biochemical analysis of Bag1- or Hsp70-overexpressing cells. The ability of Hsp70 to maintain thermally denatured firefly luciferase in a soluble state was reversed by Bag1, thus providing an explanation for the in vivo chaperone-inhibitory effects of Bag1. Similar effects on Hsp70 were observed with other cytoplasmic isoforms of Bag1 which have in common the carboxyl-terminal Hsp70-binding domain and differ by variable-length amino-terminal extensions. These results provide the first formal evidence that Bag1 functions in vivo as a regulator of Hsp70 and suggest an intriguing complexity for Hsp70-regulatory events.

In vivo chaperone activity of heat shock protein 70 and thermotolerance.

Heat shock protein 70 (Hsp70) is thought to play a critical role in the thermotolerance of mammalian cells, presumably due to its chaperone activity. We examined the chaperone activity and cellular heat resistance of a clonal cell line in which overexpression of Hsp70 was transiently induced by means of the tetracycline-regulated gene expression system. This single-cell-line approach circumvents problems associated with clonal variation and indirect effects resulting from constitutive overexpression of Hsp70. The in vivo chaperone function of Hsp70 was quantitatively investigated by using firefly luciferase as a reporter protein. Chaperone activity was found to strictly correlate to the level of Hsp70 expression. In addition, we observed an Hsp70 concentration dependent increase in the cellular heat resistance. In order to study the contribution of the Hsp70 chaperone activity, heat resistance of cells that expressed tetracycline-regulated Hsp70 was compared to thermotolerant cells expressing the same level of Hsp70 plus all of the other heat shock proteins. Overexpression of Hsp70 alone was sufficient to induce a similar recovery of cytoplasmic luciferase activity, as does expression of all Hsps in thermotolerant cells. However, when the luciferase reporter protein was directed to the nucleus, expression of Hsp70 alone was not sufficient to yield the level of recovery observed in thermotolerant cells. In addition, cells expressing the same level of Hsp70 found in heat-induced thermotolerant cells containing additional Hsps showed increased resistance to thermal killing but were more sensitive than thermotolerant cells. These results suggest that the inducible form of Hsp70 contributes to the stress-tolerant state by increasing the chaperone activity in the cytoplasm. However, its expression alone is apparently insufficient for protection of other subcellular compartments to yield clonal heat resistance to the level observed in thermotolerant cells.

Altered association of transcriptionally active DNA with the nuclear-matrix after heat shock.

PURPOSE: Exposure of human cells to heat leads to denaturation and aggregation of proteins. Within the nucleus, it has been suggested that protein aggregation is linked to the selective inhibition by hyperthermia of nucleotide excision repair in transcriptionally active genes. In this study it was investigated in detail whether and how the inhibition of repair of transcriptionally active genes might be related to alterations in their association with the nuclear-matrix. MATERIAL AND METHODS: Different protocols for nuclear-matrix isolation (high salt and lithium 3',5'-diiodosalycilate [LIS] extraction of nuclei) were used to compare DNA loop organization and positioning of transcriptionally active genes in both heated and non-heated cells. RESULTS: DNaseI digestion of total genomic DNA in Cu2+ -stabilized LIS-extracted nuclei revealed that heat shock perturbed the formation of nuclear-matrix attachment sites. Specific labelling of active genes indicated that the number of nuclear-matrix attachment sites in transcriptionally active DNA was increased due to the heat shock. At the level of individual genes, heat treatment led to stabilization of the 5' matrix attachment site (MAR) in the transcriptionally active adenosine deaminase (ADA) housekeeping gene. Moreover, heat shock resulted in the formation of an additional MAR at the 3' end of the ADA gene. The inactive 754 locus was unassociated, irrespective of a heat shock. CONCLUSIONS: The reported changes in chromatin structure might underlie the selective inhibition of repair in transcriptionally active genes and consequently may be mechanistically linked to the sensitization of heated cells to ionizing radiation.

Resistance to heat radiosensitization and protein damage in thermotolerant and thermoresistant cells.

Recently, randomized phase III trials have indicated that hyperthermia combined with radiation leads to significantly better tumour control of certain malignancies than does radio-therapy alone. Yet, the full capacity of such combined treatments might not have been optimally exploited as in vitro data indicate that repeated beating of cells can result in either the development of a transient heat resistance (thermotolerance) and/or the selection/induction of a stable heat resistant cell population. Although the mechanism of thermotolerance and its effect on thermo-radiotherapy has been studied extensively, little data are available on the mechanism of stable heat resistance and its impact on combined heat and radiation treatments. In the current study, a comprehensive analysis was made of the differences and similarities between thermotolerance (TT) and stable heat resistance (TR) in terms of the mechanism of resistance to the direct toxic action of heat and in terms of the impact on the extent of thermal radiosensitization. Using heat resistant mutants previously derived from a murine radiation-induced fibrosarcoma (RIF-1), it was observed that these cells were resistant to protein denaturation and aggregation in the cytoplasmic/membrane compartment (measured by ESR (electron spin resonance) analysis and by in situ thermal denaturation of the foreign firefly luciferase targeted to the cytoplasm) but not in the nuclear compartment (measured by TX-100 insoluble nuclear proteins and by in situ thermal denaturation of luciferase targeted to the nucleus). RIF-1-TT cells, in contrast, were resistant for a 1 end-points tested. The lack of protection of nuclear heat damage in the RIF-TR cells could not be explained by a failure of one or more of the HSP70 isoforms to enter the nuclei of these cells. In relation to the absence or presence of heat resistance in the nucleus, the extent of heat radiosensitization was reduced in RIF-1-TT but not RIF-TR cells. This implies that resistance for heat killing is not necessarily accompanied by a reduction in the ability of heat to enhance the cellular radiosensitivity. The data indicate that the mechanism leading to permanent resistance after repeated heating and the mechanism causing thermotolerance may share common features but are in part different.

Chromatin structure and cellular radiosensitivity: a comparison of two human tumour cell lines.

The role of variation in susceptibility to DNA damage induction was studied as a determinant for cellular radiosensitivity. Comparison of the radiosensitive HX142 and radioresistant RT112 cell lines previously revealed higher susceptibility to X-ray-induced DNA damage in the sensitive cell line using non-denaturing elution, but not when using alkaline unwinding. The present data also show that no difference in the amount of initial damage is seen when pulsed-field gel electrophoresis (PFGE) or comet analysis are used for DNA damage assessment. However, using the halo assay or a modified version of PFGE in which the higher DNA architecture remained partially intact, the radiosensitive cells showed steeper dose-response curves for initial DNA damage than the radioresistant cells. Analysis of the protein composition, of DNA-nucleoid structures revealed substantial differences when isolated from HX142 or RT112 cells. From our data, it is concluded that HX142 and RT112 differ in their structural organization of chromatin. As no differences in the kinetics of DNA damage rejoining were found, it is hypothesized that the same amount of lesions have a different impact in the two cell lines in that the 'presentation' of DNA damage alters the ratio of repairable to non-repairable DNA damage.

Radiation induced DNA damage and damage repair in three human tumour cell lines.

Three human tumour cell lines (HX142, RT112 and MGH-U1) with different radiosensitivities were tested for differences in the rate and/or extent of DNA unwinding in alkali as well as for differences in the induction of DNA double strand breaks by means of the pulsed field gel electrophoresis, after X-irradiation. Unlike that which has been found using the non-denaturing filter elution technique (NDE, McMillan et al., 1990), no differences in initial DNA damage (the extent of alkaline unwinding and the induction of double strand breaks) were found for the three cell lines. These data suggest that rather than a different number of DNA lesions per Da per Gy between these cell lines, structural differences in chromatin structure (related to radiosensitivity) might impair the detectability of lesions in some assays like the NDE. The nature of such structure differences remains unclear. However, the differences did not affect alkaline unwinding profiles, as all three cell lines showed identical rates of DNA unwinding after exposure to X-rays. Furthermore, the three cell lines did not show significant differences in the kinetics of DNA strand break rejoining nor in the amounts of damage remaining after 24 h repair. The results obtained in this study, together with other findings, suggest that the three cell lines may differ in their 'presentation' of DNA damage.

Thermotolerance and nuclear protein aggregation: protection against initial damage or better recovery?

Heat-induced nuclear protein aggregation and subsequent disaggregation were measured in nonpreheated and preheated (thermotolerant) HeLa S3 cells. The effect of thermotolerance on the formation of and recovery from heat-induced nuclear protein aggregates was related to the cellular levels of hsp27, hsp60, hsp70, hsc70, and hsp90. Cells heated at different time points after the thermotolerance trigger showed various levels of protection against heat-induced nuclear protein aggregation. This protection, however, did not parallel the development and decay of thermotolerance on cell survival. The protection was maximal when the thermotolerance level already had started to decay. The level of protection against nuclear protein aggregation did however parallel the cellular level of hsp70 indicating that hsp70 may be involved in this process. At all stages during the development and decay, thermotolerant cells showed a more rapid recovery (disaggregation) from the heat-induced nuclear protein aggregates than non-thermotolerant cells. The rates of disaggregation during development and decay of thermotolerance paralleled the cellular levels of hsp27 suggesting that hsp27 is somehow involved in this recovery process from heat-induced nuclear protein aggregates. The total cellular levels of none of the individual hsp's completely correlate with development and decay of thermotolerance, indicating that overexpression of any of these hsp's alone does not determine the level of thermotolerance. Clonogenic cell survival paralleled the rates of disaggregation, leading to the notion that recovery processes are the most important determinant for the thermotolerant state of HeLa S3 cells. The best correlation with clonogenic survival was found when both initial aggregation and subsequent disaggregation were taken into account, suggesting that the combined action of various hsp's in these two processes have to be included in thermotolerance development and decay.

Thermal protein denaturation and protein aggregation in cells made thermotolerant by various chemicals: role of heat shock proteins.

Thermotolerance (TT) induced by sodium arsenite (A-TT: 100 microM, 1 h, 37 degrees C) was compared to heat-induced thermotolerance (H-TT: 15 min, 44 degrees C) using HeLa S3 cells. All four pretreatments led to comparable levels of thermotolerance and also induced resistance to arsenite-, ethanol-, and diamide-induced toxicity (clonogenic ability). Stress-induced expression of the major heat shock proteins (hsp27, hsc70(p73), hsp70(p72), and hsp90) was generally highest in H-TT cells and lowest in A-TT cells. Interestingly, the four types of TT cells showed distinct differences in certain aspects of resistance against thermal protein damage. Thermal protein denaturation and aggregation determined in isolated cellular membrane fractions was found to be attenuated when they were isolated from H-TT and A-TT cells but not when isolated from E-TT and D-TT cells. The heat resistance in the proteins of the membrane fraction corresponded with elevated levels of hsp70(p72) associated with the isolated membrane fractions. In the nuclear fraction, only marginal (not significant) attenuation of the formation of protein aggregates (as determined by TX-100 (in)solubility) was observed. However, the postheat recovery from heat-induced protein aggregation in the nucleus was faster in H-TT, E-TT, and D-TT cells, but not in A-TT cells. Despite the fact that elevated levels of hsp27, hsp70(p73), and hsp70(p72) were found in the TX-100 insoluble nuclear fraction derived from all TT cells, no correlation was found with the degree of resistance in terms of the accelerated recovery from nuclear protein aggregation. The only correlation between accelerated recovery from nuclear protein aggregates was that with total cellular levels of hsp27. The data indicate that heat-induced loss of clonogenic ability may be a multitarget rather than a single target event. A threshold of damage may exist in cells after exposure to heat; multiple sets of proteins in (different compartments of) the cell need to be damaged before this threshold is exceeded and the cell dies. As a consequence, stabilization of only one of these sets of proteins is already sufficient to render cells thermotolerant at the clonogenic level.

Selective inhibition of repair of active genes by hyperthermia is due to inhibition of global and transcription coupled repair pathways.

Hyperthermia specifically inhibits the repair of UV-induced DNA photolesions in transcriptionally active genes. To define more precisely which mechanisms underlie the heat-induced inhibition of repair of active genes, removal of cyclobutane pyrimidine dimers (CPDs) was studied in human fibroblasts with different repair capacities and different transcriptional status of the adenosine deaminase gene, i.e. normal human cells, human cells carrying an inactive copy of the adenosine deaminase gene and xeroderma pigmentosum complementation group C fibroblasts. The results indicate that repair of active genes is impaired by inhibition of two repair pathways: (i) a global repair system involved in the repair of CPDs in potentially active genes; and (ii) the transcription-coupled repair pathway responsible for the accelerated repair of the transcribed strand. Since X-ray-induced DNA damage is also preferentially removed from the transcribed strand of active genes, selective inhibition of repair of radiation-induced DNA damage in active genes may play a key role in radiosensitization due to hyperthermia.

Cells overexpressing Hsp27 show accelerated recovery from heat-induced nuclear protein aggregation.

Protein denaturation/aggregation upon cell exposure to heat shock is a likely cause of cell death. In the nucleus, protein aggregation has often been correlated to inhibition of nuclear located processes and heat-induced cell killing. In Chinese hamster O23 cells made thermotolerant by a prior heating (20'44 degrees C + 10h 37 degrees C) which induces the whole spectrum of heat shock proteins (hsps), the extent of nuclear protein aggregation during heat shock is reduced and the rate of recovery from aggregation after heat shock is enhanced. In contrast, a heat resistant Chinese hamster cell line overexpressing only hsp27 shows an unaltered sensitivity to formation of nuclear protein aggregates by heat, but shows the same enhanced rate of recovery from nuclear protein aggregation as thermotolerant cells. This suggests that accelerated recovery of protein aggregation could be partly responsible for hsp27-mediated thermoprotection.

Heat-shock treatment selectively affects induction and repair of cyclobutane pyrimidine dimers in transcriptionally active genes in ultraviolet-irradiated human fibroblasts.

The effect of hyperthermia on induction and repair of UV-radiation-induced cyclobutane pyrimidine dimers was investigated in the genome overall and in transcriptionally active and inactive genes in confluent human fibroblasts. Hyperthermia treatment (30 min, 45 degrees C) of human fibroblasts resulted in an increase in the protein content of isolated nuclei (protein aggregation) similar to that observed for HeLa S3 cells. The faster rate of disaggregation of nuclear proteins and the higher survival rate of heated fibroblasts in comparison with those for HeLa cells provide further evidence for a possible role of protein aggregation in heat-induced cell killing. Determination of the frequencies of cyclobutane pyrimidine dimers in the genome overall and in restriction fragments of the active adenosine deaminase (ADA) gene and inactive 754 locus revealed that hyperthermia selectively inhibits the induction of cyclobutane pyrimidine dimers in transcriptionally active DNA. Removal of cyclobutane pyrimidine dimers from the ADA gene was strongly delayed during the first 8 h in 10 J/m2 UV-irradiated fibroblasts. Such inhibition of repair of cyclobutane pyrimidine dimers was not observed for the 754 gene, indicating that inhibition of repair by hyperthermia is generally not mediated by inactivation of repair enzymes. It is proposed that the inhibition of induction and repair of cyclobutane pyrimidine dimers in active genes by hyperthermia is related to the heat-induced aggregation of proteins with the nuclear matrix, proximal to which active genes are located. Our results are consistent with a functional compartmentalization of DNA repair at the nuclear matrix.

Association of HSP72 with the nuclear (TX-100-insoluble) fraction upon heating tolerant and non-tolerant HeLa S3 cells.

HSP72 levels in the cellular and the nuclear (TX-insoluble) fraction before and after heating of heat- and sodium arsenite-induced thermotolerant and non-tolerant HeLa S3 cells have been investigated by 1D- and 2D-electrophoresis, followed by Western blotting and immunostaining, using a newly developed monoclonal antibody that specifically detects HSP72 (Heine et al. 1991). HSP72 was constitutively expressed in HeLa S3 cells and elevated upon heat or arsenite stress. Immediate association of HSP72 with the nuclear fraction was induced by heat but not arsenite. However, at the time of maximal thermotolerance, elevated levels of HSP72 were found associated with nuclei isolated from both heat- and arsenite-induced thermotolerant cells. After (test) heat treatments (0-60 min at 45 degrees C) translocation of HSP72 to the nuclear fraction in all cells was observed, albeit with different kinetics and to different plateau values. When tolerant and non-tolerant cells were allowed to recover from a heat stress (at 37 degrees C) before isolation of the nuclei, no dissociation of HSP72 from the nuclear fraction was observed within a 5 h time period. Our data indicate that association/dissociation of HSP72 with/from the nuclear fraction is not related to the recovery from heat-induced intranuclear protein aggregation (Kampinga et al. 1992), nor to the extent of thermotolerance in the human HeLa S3 cell line.

Acquisition of thermotolerance induced by heat and arsenite in HeLa S3 cells: multiple pathways to induce tolerance?

Recent data indicate that cells may acquire thermotolerance via more than one route. In this study, we observed differences in thermotolerance development in HeLa S3 cells induced by prior heating (15 minutes at 44 degrees C) or pretreatment with sodium-arsenite (1 hour at 37 degrees C, 100 microM). Inhibition of overall protein and heat shock protein (HSP) synthesis (greater than 95%) by cycloheximide (25 micrograms/ml) during tolerance development nearly completely abolished thermotolerance induced by arsenite, while significant levels of heat-induced thermotolerance were still apparent. The same dependence of protein synthesis was found for resistance against sodium-arsenite toxicity. Toxic heat, but not toxic arsenite treatments caused heat damage in the cell nucleus, measured as an increase in the protein mass of nuclei isolated from treated cells (intranuclear protein aggregation). Recovery from this intranuclear protein aggregation was observed during post-heat incubations of the cells at 37 degrees C. The rate of recovery was faster in heat-induced tolerant cells than in nontolerant cells. Arsenite-induced tolerant cells did not show an enhanced rate of recovery from the heat-induced intranuclear protein aggregation. In parallel, hyperthermic inhibition of RNA synthesis was the same in tolerant and nontolerant cells, whereas post-heat recovery was enhanced in heat-induced, but not arsenite-induced thermotolerant cells. The more rapid recovery from heat damage in the nucleus (protein aggregation and RNA synthesis) in cells made tolerant by a prior heat treatment seemed related to the ability of heat (but not arsenite) to induce HSP translocations to the nucleus.