Anticancer Drugs Up-regulate HspBP1 and Thereby Antagonize the Prosurvival Function of Hsp70 in Tumor Cells*

The 70-kDa heat shock protein (Hsp70) is up-regulated in a wide variety of tumor cell types and contributes to the resistance of these cells to the induction of cell death by anticancer drugs. Hsp70 binding protein 1 (HspBP1) modulates the activity of Hsp70 but its biological significance has remained unclear. We have now examined whether HspBP1 might interfere with the prosurvival function of Hsp70, which is mediated, at least in part, by inhibition of the death-associated permeabilization of lysosomal membranes. HspBP1 was found to be expressed at a higher level than Hsp70 in all normal and tumor cell types examined. Tumor cells with a high HspBP1/Hsp70 molar ratio were more susceptible to anticancer drugs than were those with a low ratio. Ectopic expression of HspBP1 enhanced this effect of anticancer drugs in a manner that was both dependent on the ability of HspBP1 to bind to Hsp70 and sensitive to the induction of Hsp70 by mild heat shock. Furthermore, anticancer drugs up-regulated HspBP1 expression, whereas prevention of such up-regulation by RNA interference reduced the susceptibility of tumor cells to anticancer drugs. Overexpression of HspBP1 promoted the permeabilization of lysosomal membranes, the release of cathepsins from lysosomes into the cytosol, and the activation of caspase-3 induced by anticancer drugs. These results suggest that HspBP1, by antagonizing the prosurvival activity of Hsp70, sensitizes tumor cells to cathepsin-mediated cell death.

The 70-kDa heat shock protein (Hsp70) is up-regulated in a wide variety of tumor cell types and contributes to the resistance of these cells to the induction of cell death by anticancer drugs. Hsp70 binding protein 1 (HspBP1) modulates the activity of Hsp70 but its biological significance has remained unclear. We have now examined whether HspBP1 might interfere with the prosurvival function of Hsp70, which is mediated, at least in part, by inhibition of the death-associated permeabilization of lysosomal membranes. HspBP1 was found to be expressed at a higher level than Hsp70 in all normal and tumor cell types examined. Tumor cells with a high HspBP1/Hsp70 molar ratio were more susceptible to anticancer drugs than were those with a low ratio. Ectopic expression of HspBP1 enhanced this effect of anticancer drugs in a manner that was both dependent on the ability of HspBP1 to bind to Hsp70 and sensitive to the induction of Hsp70 by mild heat shock. Furthermore, anticancer drugs upregulated HspBP1 expression, whereas prevention of such upregulation by RNA interference reduced the susceptibility of tumor cells to anticancer drugs. Overexpression of HspBP1 promoted the permeabilization of lysosomal membranes, the release of cathepsins from lysosomes into the cytosol, and the activation of caspase-3 induced by anticancer drugs. These results suggest that HspBP1, by antagonizing the prosurvival activity of Hsp70, sensitizes tumor cells to cathepsin-mediated cell death.
Members of the 70-kDa heat shock protein (Hsp70) 2 family play an essential role in quality control of cellular proteins (1)(2)(3) and include stress-inducible Hsp70, constitutively expressed Hsc70, mitochondrial Hsp75, and endoplasmic reticulum GRP78 (4,5). Under normal conditions, Hsp70 proteins function as ATP-dependent molecular chaperones by facilitating the folding of newly synthesized polypeptides, the assembly of multiprotein complexes, and the transport of proteins across cellular membranes. Under stressful conditions, the synthesis of inducible Hsp70 enhances the ability of cells to cope with increased concentrations of unfolded or denatured proteins (1)(2)(3)(4)(5)(6).
The Hsp70 proteins undergo a cycle of substrate binding and release that is accelerated by ATP hydrolysis (1)(2)(3). The substrate binding domain of Hsp70 is localized to a 25-kDa COOH-terminal region, with substrate access to this domain being controlled by a COOH-terminal "lid" that exposes the domain in the ATP-bound form and allows substrate binding to occur when Hsp70 is in the ADP-bound form. Opening and closing of the lid are governed by conformational changes associated with ATP binding and hydrolysis, which occur within the cleft of the 45-kDa NH 2 -terminal ATPase domain. The exchange of bound ADP for ATP results in substrate release, thus allowing Hsp70 to enter a new round of substrate binding and release.
The chaperone activity of Hsp70 proteins is regulated by various accessory proteins, known as cochaperones (1)(2)(3). For example, Hsp40 binds to the COOH terminus of Hsp70, stimulates its ATPase activity, and thereby stabilizes the substratebound form. The cochaperones Hip and Bag-1 bind to the ATPase domain of Hsp70 and function as nucleotide exchange factors: Hip prevents the dissociation of ADP from Hsp70 and thereby stabilizes the substrate-bound form, whereas Bag-1 promotes the release of ADP and rebinding of ATP, thereby triggering the premature unloading of bound substrate proteins from Hsp70.
In addition to their essential role in protein quality control, Hsp70 proteins have been shown to contribute to tumorigenesis. A high level of expression of these proteins thus enhances the tumorigenic potential of rodent cells in syngeneic animals (7) and is associated with a poor therapeutic outcome in several human cancers (8,9), whereas depletion of Hsp70 promotes tumor regression (10). The tumorigenic potential of Hsp70 is thought to be attributable to an ability to confer a survival advantage on tumor cells through direct interference with several key components of the apoptotic signaling pathway, including JNK (11)(12)(13), AIF (14,15), and apoptotic proteaseactivating factor-1 (16 -18). The prosurvival function of Hsp70 proteins has been suggested to be distinct from their chaperone activity (14). Hsp70 has been localized to membranes of the endosomal-lysosomal compartment and shown to inhibit the cell death-associated release of cathepsins from these vesicles in tumor cells, thus presumably contributing to its prosurvival function (19,20).
Hsp70-binding protein 1 (HspBP1) was originally identified as a protein that interacts with Hsp70 and inhibits its chaperone activity (21). It was subsequently shown to bind to the ATPase domain of Hsp70 and to inhibit its ATPase activity (22). However, HspBP1 has also been shown to function as a nucleotide exchange factor for Hsp70 and Hsc70 (23) and to stimulate their chaperone activities (24). Although HspBP1 appears to represent a new member of the class of nucleotide exchange factors for Hsp70/Hsc70, its biological significance has remained poorly understood.
We have now investigated the physiological function of HspBP1, focusing on its possible interference with the prosurvival activity of Hsp70. Our results suggest that anticancer drugs up-regulate the expression of HspBP1, which then specifically binds to Hsp70 and antagonizes its ability to inhibit the cell death-associated permeabilization of lysosomes. HspBP1 thus contributes to the induction of cathepsin-mediated programmed cell death by anticancer drugs in tumor cells.

EXPERIMENTAL PROCEDURES
Materials-Vincristine and paclitaxel were obtained from Sigma, and etoposide was from Wako Pure Chemical (Osaka, Japan). CA-074 Me was obtained from Biomol (Plymouth Meeting, PA). Rabbit polyclonal antibodies to HspBP1 were generated in response to a peptide corresponding to residues 342 to 359 (EKLLQTCFSSPADDSMDR) of the human protein.
Plasmids and Cell Transfection-An HspBP1 cDNA was isolated from a human heart cDNA library by yeast two-hybrid screening with the ATPase domain of human Hsp70 as the bait (DDBJ number AB020592). To generate expression plasmids encoding EGFP-HspBP1 or the deletion mutant EGFP-HspBP1(⌬C), we amplified cDNA fragments from human TIG-3 diploid fibroblast cDNA by polymerase chain reaction with the primers 5Ј-CGGAATTCCATGTCAGACGAAGG-CTC-3Ј (forward), with the underlined sequence corresponding to an EcoRI site, and either 5Ј-CGGCTCGAGTCACCGATC-CATGCTGTC-3Ј (reverse) for HspBP1 or 5Ј-CGGCTCGAG-TCACGGCTCCCGACACTC-3Ј (reverse) for HspBP1(⌬C), with the underlined sequences corresponding to a XhoI site. The PCR products were digested with EcoRI and XhoI, and the resulting fragments were cloned into the EcoRI-and SalI-digested pEGFP-C1 expression vector (Clontech, Palo Alto, CA) and verified by sequencing. To generate expression plasmids for EGFP-HspBP1(⌬M) or EGFP-HspBP1(⌬MC), we amplified DNA fragments corresponding to COOH-terminal portions of HspBP1 (amino acids 196 to 359 for HspBP1(⌬M) and 196 to 313 for HspBP1(⌬MC)) by PCR with the primers 5Ј-GGA-AGGTACCTGGACCGCGACGCCTGCGAC-3Ј (forward), with the underlined sequence corresponding to a KpnI site, and either 5Ј-GCAGGATCCTCACCGATCCATGCTGTC-3Ј (reverse) for HspBP1(⌬M) or 5Ј-GCAGGATCCTCACGGCT-CCCGACACTC-3Ј (reverse) for HspBP1(⌬MC), with the underlined sequences corresponding to a BamHI site. The PCR products were digested with KpnI and BamHI, and the resulting fragments were cloned into the EGFP-HspBP1 expression vector that had been digested with KpnI and BamHI (to delete the DNA sequence encoding amino acids 154 to 359) and were verified by sequencing. Transfection of HeLa S3 cells with expression plasmids was performed with the use of the Lipofectamine 2000 reagent (Invitrogen). For establishment of HeLa S3 cells stably expressing EGFP or EGFP-HspBP1, the cells were subjected to selection in culture medium supplemented with Geneticin (400 g/ml) after transfection, and individual resistant colonies were isolated to obtain cell clones.
Recombinant Proteins-The HspBP1 coding sequence was inserted into the pET24b vector (Novagen, Madison, WI), and the resulting construct was introduced by transformation into Escherichia coli strain BL21(DE3). Expression of His 6 -tagged HspBP1 was induced by the addition of isopropyl ␤-D-thiogalactopyranoside to the bacterial culture. Bacterial cells were lysed by ultrasonic disruption in 20 ml of a solution containing 20 mM Tris-HCl (pH 8.0) and 100 mM NaCl. Cell lysates were centrifuged at 12,000 ϫ g for 30 min at 4°C, and the resulting supernatant was applied to a Talon metal affinity column (Clontech). Recombinant His 6 -HspBP1 was eluted from the column with the cell lysis solution supplemented with 100 mM imidazole and subjected to dialysis against phosphate-buffered saline. Recombinant human Hsp70 (NSP-555) and recombinant bovine Hsc70 (SPP-751) were obtained from Stressgen Bioreagents.
Flow Cytometry-Cells were harvested by exposure to trypsin, fixed with ice-cold 70% ethanol, treated with RNase A (100 g/ml) (DNase-free, Sigma), and stained with propidium iodide (20 g/ml) as described previously (27). At least 1 ϫ 10 4 cells were analyzed for DNA content (excitation at 488 nm, emission at 620 nm) or the expression of EGFP-tagged proteins (excitation at 488 nm, emission at 530 nm) with the use of a FACSCalibur flow cytometer and Cell Quest Pro software (BD Biosciences, San Jose, CA).
Clonogenic Cell Survival Assay-Cells seeded in 12-well plates (250 cells per well) were allowed to grow for 48 h before treatment with various drugs for 48 h. They were then allowed to form colonies by incubation in drug-free medium for 5 days. The resulting colonies were fixed with 70% ethanol and stained with 0.5% crystal violet. Colonies consisting of more than 50 cells were counted (29).
Immunofluorescence Microscopy-Cells grown on glass coverslips were exposed to various reagents, fixed with methanol at Ϫ20°C, and transferred to blocking buffer (phosphate-buffered saline containing 2.5% bovine serum albumin). Cells were stained with monoclonal antibodies to cathepsin L or cathepsin B (1/200 dilution in blocking buffer), and immune complexes were then detected with Alexa Fluor 546-conjugated goat antibodies to mouse immunoglobulin G (1/200 dilution in blocking buffer) (Molecular Probes, Eugene, OR). Nuclei were stained with Hoechst 33342 (Sigma). Confocal images were obtained with an Axiovert 200M microscope equipped with an LSM 5 Pascal system (Carl Zeiss, Jena, Germany).
Statistical Analysis-Data are presented as mean Ϯ S.D. and analyzed where indicated by the two-tailed Student's t test. A p value of Ͻ0.05 was considered statistically significant.  of HeLa S3 cells were subjected to immunoprecipitation (IP) with antibodies to HspBP1, Hsp70, or Hsc70 or with nonimmune immunoglobulin G. The resulting precipitates, as well as the original cell lysates (40 g of protein), were then subjected to immunoblot analysis (IB) with the same antibodies, as indicated. Arrowheads and asterisks indicate specific and nonspecific bands, respectively. B, HeLa S3 cells were transfected for 24 h with expression plasmids for EGFP-tagged HspBP1 or its deletion mutants (shown in the upper panel). Cell lysates (200 g of protein) were subjected to immunoprecipitation with anti-Hsp70, and the resulting precipitates were subjected to immunoblot analysis with antibodies to Hsp70 or GFP (lower left panel). Cell lysates (40 g of protein) were also subjected directly to immunoblot analysis with the same antibodies (lower right panel). All data are representative of three separate experiments.

Preferential Association of HspBP1 with Hsp70 in Cells-
Hsp70 or Hsc70, we examined the association of these proteins in HeLa S3 cells with a reciprocal co-immunoprecipitation assay. Immunoprecipitates prepared with antibodies to HspBP1 were found to contain not only HspBP1 but also a substantial amount of Hsp70 and a small amount of Hsc70 (Fig.  1A). Conversely, HspBP1 was detected in substantial amounts in immunoprecipitates prepared with anti-Hsp70 but was virtually undetectable in those prepared with anti-Hsc70. Essentially identical results were obtained with MKN1 cells (data not shown). Hsc70 is an abundant protein that is constitutively expressed in normal and tumor cells at similar levels; its concentration in most of the cell lines examined was 20 to 30 g/mg of cellular protein, a value more than 10 times that for Hsp70 (Fig. 2,  A and B). These results thus indicated that HspBP1 binds to Hsp70 in cells with much higher affinity than it does to Hsc70. Given that only a small proportion of Hsc70 molecules are associated with HspBP1 in cells, it is likely that HspBP1 exerts little effect on the physiological function of Hsc70. We therefore focused on the interaction between HspBP1 and Hsp70 in subsequent experiments.
Increased Sensitivity of Tumor Cells with a High HspBP1/ Hsp70 Molar Ratio to Anticancer Drugs-The molar ratio of HspBP1 to Hsp70 in cells might be expected to be an important determinant of the interaction between these two proteins as well as of the function of the resulting complex. In this regard, an HspBP1/Hsp70 ratio of ϳ4 has been suggested to be required for inhibition of Hsp70 activity by 50% (30). We therefore next determined the amounts of Hsp70 and HspBP1 in a variety of tumor cell types by immunoblot analysis, with recombinant Hsp70 and HspBP1 (His 6 -tagged) as respective standards. The concentration of Hsp70 in all the tumor cell lines examined was severalfold greater than that in diploid fibroblasts (Fig. 2, A and B). Although the concentration of HspBP1 in many tumor cell types was greater than that in normal fibroblastic cells, the difference in the amount of HspBP1 between tumor cells and normal cells was not as marked as that for Hsp70. Furthermore, no correlation between the expression level of Hsp70 and of HspBP1 was apparent in the tumor cells. The molar ratio of HspBP1 to Hsp70 was thus highly variable among the tumor cell types, ranging from a value similar to that for diploid fibroblasts (more than 10) for HLF, MKN1, and DLD-1 cells to less than 2 for A431, NUGC-3, and MKN28 cells (Fig. 2C).
HspBP1 has been shown to bind to Hsp70 and modulate its activity either negatively (21,22) or positively (24). In addition, the increased expression of Hsp70 in tumor cells is thought to contribute to the resistance of these cells to the induction of cell death by anticancer drugs (8,9). To obtain insight into the biological effect of HspBP1 on Hsp70 activity, we compared the drug sensitivities of tumor cell types with a high HspBP1/Hsp70 ratio (MKN1, DLD-1, and HLF cells) and those with a low HspBP1/Hsp70 ratio (HeLa S3, WiDr, and MKN28 cells). The cells were treated with vincristine (100 nM), paclitaxel (100 nM), or etoposide (34 M) for up to 48 h, and the proportion of dead cells with a fractional DNA content (cells in sub-G 1 phase) was then determined by flow cytometry. Tumor cells with a high HspBP1/ Hsp70 molar ratio were more susceptible to the anticancer drugs than were those with a low ratio (Fig. 3, A and B). The differences in drug sensitivity between these two groups of tumor cell lines were confirmed by clonogenic cell survival assays performed with much lower concentrations of the anticancer drugs compared with those used for assay of the induction of cell death (Fig. 3C, data not shown). These results suggested that HspBP1 antagonizes the ability of Hsp70 to protect tumor cells from anticancer drug-induced cell death.
Ectopic Expression of HspBP1 Increases the Susceptibility of Tumor Cells to Anticancer Drugs-To elucidate further the biological significance of HspBP1, we established HeLa S3 cells that stably express EGFP-HspBP1. Immunoblot analysis revealed that the expression level of EGFP-HspBP1 was about four times that of endogenous HspBP1 in several of the HeLa-EGFP-HspBP1 cell clones (Fig. 4A). These HeLa-EGFP-HspBP1 cells exhibited an increased sensitivity to the induction of cell death by anticancer drugs compared with control HeLa-EGFP cells (Fig. 4B). Furthermore, induction of an ϳ2-fold increase in Hsp70 expression by mild heat shock (Fig. 4B, inset)

HspBP1 Antagonizes the Prosurvival Function of Hsp70
etoposide or paclitaxel on cell death in both HeLa-EGFP and HeLa-EGFP-HspBP1 cells (Fig. 4B). Clonogenic cell survival assays confirmed these results. Colony formation efficiency (relative to that of corresponding control cells not exposed to drugs) was thus 84.64 Ϯ 1.14 and 46.15 Ϯ 5.81% for HeLa-EGFP cells (clone 3) and HeLa-EGFP-HspBP1 cells (clone 1), respectively, exposed to 1 nM paclitaxel, and these values were increased to 96.33 Ϯ 2.51 and 82.30 Ϯ 3.87%, respectively, for cells first subjected to mild heat shock (data are mean Ϯ S.D. from three independent experiments).   Transient transfection of HeLa S3 cells with the expression vector for EGFP-HspBP1 also potentiated the induction of cell death by anticancer drugs (Fig. 4C). Transient expression of the EGFP-HspBP1(⌬C) mutant, which exhibits a reduced ability to bind to Hsp70 (Fig. 1B), increased the susceptibility of HeLa S3 cells to etoposide or paclitaxel only slightly (Fig. 4D), whereas transient expression of EGFP-HspBP1(⌬MC), which lacks Hsp70 binding activity (Fig. 1B), did not affect the susceptibility of HeLa S3 cells to these anticancer drugs (Fig. 4D). These results indicated that the biological activity of HspBP1 is completely dependent on its ability to bind to Hsp70.
Anticancer Drugs Up-regulate HspBP1 in Tumor Cells-We next examined the possible effects of anticancer drugs on the expression of Hsp70 and HspBP1 in tumor cells. All three anticancer drugs examined (vincristine, paclitaxel, and etoposide) induced a ϳ2.0 -2.5-fold increase in the amount of HspBP1 in HeLa S3 cells (representative of tumor cells with a low HspBP1/ Hsp70 molar ratio) as well as in MKN1 cells (representative of tumor cells with a high HspBP1/Hsp70 molar ratio), with the highest expression levels being apparent between 6 and 24 h after the onset of drug treatment (Fig. 5). In contrast, the expression level of Hsp70 was not substantially affected by exposure of the tumor cells to these anticancer drugs for up to 24 h.
Depletion of HspBP1 Reduces the Susceptibility of Tumor Cells to Anticancer Drugs-We examined the effect of RNAimediated depletion of endogenous HspBP1 on the susceptibility of tumor cell to anticancer drugs. For these experiments, we studied MKN1 cells, which express HspBP1 at a relatively high level. Immunoblot analysis revealed that transfection of MKN1 cells with siRNAs (A or B) specific for HspBP1 mRNA, but not that with a control RNA duplex, resulted in a pronounced reduction in the abundance of HspBP1 as well as in marked suppression of the up-regulation of HspBP1 induced by anticancer drugs (Fig. 6A). Depletion of HspBP1 by RNAi reduced the susceptibility of MKN1 cells to anticancer drugs. Clonogenic cell survival assays thus revealed that RNAi-mediated depletion of HspBP1 resulted in a marked increase in the colony formation efficiency of MKN1 cells exposed to vincristine, paclitaxel, or etoposide compared with that observed for drug-treated cells transfected with the control siRNA (Fig. 6B). We confirmed these results by showing that depletion of HspBP1 resulted in marked inhibition of the induction of cell death (cells with a fractional DNA content) by vincristine (100 nM), paclitaxel (100 nM), or etoposide (34 M) (data not shown).

Promotion of Anticancer Drug-induced Permeabilization of Lysosomal Membranes and Cathepsinmediated Cell Death by Ectopic
Expression of HspBP1-Finally, we examined the mechanism by which HspBP1 interferes with the prosurvival function of Hsp70. Hsp70 has been shown to promote cell survival through inhibition of death-associated permeabilization of lysosomal membranes (19,20). Treatment with anticancer drugs induces both the translocation of lysosomal cathepsins from the lysosomal lumen to the cytosol as well as permeabilization of the mitochondrial outer membrane, events that are followed by caspase-or AIF-mediated programmed cell death (31,32). We therefore examined the effect of ectopic expression of HspBP1 on anticancer drug-induced permeabilization of lysosomal membranes.
Immunostaining of HeLa S3 cells transiently expressing EGFP, EGFP-HspBP1, or EGFP-HspBP1(⌬C) revealed a predominantly perinuclear and punctate distribution of cathepsin L (Fig. 7A), consistent with the lysosomal localization of this enzyme. Immunostaining with antibodies to cathepsin B yielded essentially identical results (data not shown). Treatment with etoposide for 24 h induced the translocation of these cathepsins from cytoplasmic vesicles to the cytosol in cells expressing EGFP-HspBP1 but not in those expressing EGFP or EGFP-HspBP1(⌬C). Furthermore, prior exposure of the cells expressing EGFP-HspBP1 to mild heat shock to induce up-regulation of Hsp70 (see Fig. 4B, inset) resulted in inhibition of the etoposide-induced translocation of cathepsins. Exposure to etoposide for more than 60 h induced the translocation of cathepsins B and L to the cytosol also in the cells expressing EGFP or EGFP-HspBP1(⌬C) (data not shown). These results suggested that ectopic expression of HspBP1 promotes the anticancer drug-induced permeabilization of lysosomal membranes and the consequent release of cathepsins from lysosomes, and that this action of HspBP1 is sensitive to the up-regulation of Hsp70 induced by mild heat shock.
Cathepsins B and L have been shown to cleave Bid, a proapoptotic BH3-only member of the Bcl-2 family, resulting in the generation of its active ϳ15-kDa fragment (tBid), which in turn induces permeabilization of the mitochondrial outer membrane, the release of cytochrome c into the cytosol, and caspase-3-dependent programmed cell death (31,33). We found that the cleavage of Bid as well as the activation of caspase-3 (as revealed by the appearance of its active 19-and 17-kDa fragments) induced by paclitaxel or etoposide were enhanced in HeLa-EGFP-HspBP1 cells compared with those in HeLa-EGFP cells (Fig. 7B). Furthermore, CA-074 Me, a specific inhibitor of cathepsin B, suppressed the cleavage of Bid and the activation of caspase-3 induced by anticancer drugs in both cell types (Fig.  7B), and these effects were accompanied by inhibition of paclitaxel-induced cell death (Fig. 7C).

DISCUSSION
HspBP1 is an Hsp70-binding protein that appears to be ubiquitously expressed in mammalian cells and which we have now shown binds to Hsp70 with a much higher affinity than it does to Hsc70 in cells. The abundance of Hsp70 is increased in many tumor cell types, with this up-regulation contributing to the reduced susceptibility of these cells to the induction of cell death by anticancer drugs (6,8,9). We have also now shown that the amount of Hsp70 in all the tumor cell types examined was greater than that in diploid fibroblasts. Furthermore, although our results now indicate that the biological activity of HspBP1 is dependent on its ability to bind to Hsp70, the expression levels of these two proteins in tumor cells do not appear to be correlated. Similar results suggesting that the expression of HspBP1 is not coordinately regulated with that of Hsp70 were recently described (34).
Our present results show that HspBP1 is a relatively abundant protein and is expressed at higher levels than is Hsp70 in both normal and tumor cell types. Moreover, tumor cells with a high HspBP1/ Hsp70 molar ratio were found to be more susceptible to anticancer drugs than were those with a low ratio. Ectopic expression of HspBP1 also enhanced the induction of tumor cell death by anticancer drugs; this effect appeared to be dependent on the ability of HspBP1 to bind to Hsp70 and was inhibited by heat shock-induced up-regulation of Hsp70. Conversely, RNAimediated depletion of endogenous HspBP1 markedly reduced the susceptibility of tumor cells to anticancer drugs. Together, these results indicate that, under physiological conditions, HspBP1 functions as an inhibitor of the prosurvival function of Hsp70.
Hsp70 protects tumor cells from a wide range of lethal stimuli by various mechanisms, which include prevention of JNK activation, release of AIF from mitochondria, and nuclear import of the released AIF (4 -6). However, overexpression of HspBP1 did not substantially affect either the activation of JNK (as determined by immunoblot analysis of the phosphorylation of c-Jun) or the nuclear import of AIF (as examined by immunostaining with anti-AIF) induced by anticancer drugs in HeLa S3 cells (data not shown). Hsp70 localizes to membranes of the endosomal-lysosomal compartment of tumor cells and promotes survival of tumor cells through inhibition of the deathassociated permeabilization of lysosomal membranes (19,20). Furthermore, the expression, secretion, or activity of lysosomal proteases such as cathepsins B, D, and L has been shown to be increased in most human tumor types examined (35,36). Consistent with these observations, we found that ectopic expression of HspBP1 in HeLa S3 cells promoted anticancer drug- induced permeabilization of lysosomal membranes, as reflected by the release of cathepsins B and L from lysosomes into the cytosol, resulting in enhancement of cathepsin-mediated cell death. Furthermore, heat shock suppressed the anticancer drug-induced permeabilization of lysosomal membranes in cells overexpressing HspBP1, likely as a result of up-regulation of Hsp70. These results suggest that the antagonistic effect of HspBP1 on the prosurvival function of Hsp70 is mediated, at least in part, through suppression of the ability of Hsp70 to inhibit the death-associated permeabilization of lysosomal membranes. It remains to be determined whether binding of HspBP1 to Hsp70 interferes with the localization of the latter protein to lysosomal membranes or indeed suppresses its ability to inhibit the death-associated permeabilization of these membranes.
Anticancer drugs induced the up-regulation of HspBP1 in tumor cells, without affecting the expression level of Hsp70.
Furthermore, the observed RNAimediated abrogation of such up-regulation of HspBP1 was associated with suppression of drug-induced cell death. These observations thus indicate that the up-regulation of HspBP1 by anticancer drugs increases the antagonism of Hsp70 by HspBP1 and thereby sensitizes tumor cells to cathepsin-mediated cell death. The suppression of anticancer drug-induced cell death either by RNAimediated depletion of HspBP1 or by cathepsin inhibitors was partial, indicating that the death response is mediated by multiple pathways, including those dependent on and independent of cathepsins (36 -38).
The abundance of HspBP1 in cells is regulated at the transcriptional, post-transcriptional, and post-translational levels. Reverse transcription and PCR analysis did not reveal a marked increase in the amount of HspBP1 mRNA in HeLa S3 cells in response to treatment with vincristine or etoposide for up to 24 h (data not shown). The precise mechanism for the up-regulation of HspBP1 induced by anticancer drugs in tumor cells thus remains to be determined.
The amounts of Hsp70 and HspBP1 in normal and tumor cells were previously estimated (30). Although the amounts of Hsp70 in these cells were similar to those determined in the present study, those of HspBP1 were estimated at ϳ0.1 ng/g of protein in normal cells and 0.4 to 0.6 ng/g of protein in tumor cells (30), values that are less than one-tenth of those determined in the present study. Given that an HspBP1/Hsp70 molar ratio of ϳ4 is thought to be required for inhibition of Hsp70 activity by 50%, the authors of this previous study concluded that global inhibition of Hsp70 activity by HspBP1 in cells is unlikely (30).
We found that tumor cells with a high HspBP1/Hsp70 molar ratio are more sensitive to anticancer drugs than are those with a low ratio. Furthermore, depletion of HspBP1 by RNAi reduced susceptibility to anticancer drugs not only in MKN1 cells (Fig. 6) but also in DLD-1 cells (data not shown). Our present results are thus inconsistent with the previous observations suggesting that HspBP1 lacks a biological effect in cells. Rather, our results indicate that HspBP1 interferes with the prosurvival function of Hsp70 in tumor cells under physiological conditions. Finally, our finding that anticancer drugs induce the up-regulation of HspBP1 in tumor cells suggests that this protein, by antagonizing the prosurvival activity of Hsp70, functions to sensitize tumor cells to the cathepsin-mediated cell death induced by anticancer drugs.