| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 37, 35231-35240, September 12, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Department of Medical Chemistry, Semmelweis University, H-1088 Budapest, Hungary, the ¶Department of Biophysics, Semmelweis University, H-1088 Budapest, Hungary, and the ||Department of Microbiology and Biotechnology, St. István University, H-1118 Budapest, Hungary
Received for publication, February 7, 2003 , and in revised form, June 17, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Cell lysis is one of the most commonly used methods to test cellular integrity. Moreover, lysis rate anomalies (13, 14) together with diffusional anomalies (15, 16) were used as important arguments for the organization of the cytoplasm. Since cellular integrity is preserved after a partial cell lysis to a large extent (13, 14), partial lysis provides a highly sensitized, but still somewhat organized cellular system, where the contribution of various components to both the cytoplasmic organization and cellular stability can be studied.
The original aim of the present study was to examine whether Hsp90 inhibition induces any change in the rate of cell lysis induced by mild detergent treatment or hypotonic shock. The rationale behind these experiments was to test, whether Hsp90, a cytoprotective chaperone, binding to "thousand-and-one" substrates and other proteins is involved in the maintenance of cellular integrity (17), and whether its inhibition renders cells more "lysis-prone." The first experiments were very promising: geldanamycin, a well-established Hsp90 inhibitor (18, 19) induced a significant increase in lysis rate of various cells. However, later experiments demonstrated that the extent of geldanamycin-induced enhancement of cell lysis was dependent on the experimental conditions, namely, if cells were shaken during the experiment or not. At this time the first results of geldanamycin-induced superoxide generation appeared (20, 21). These results turned our attention to examine the contribution of superoxide-related versus Hsp-related events to diminished cellular integrity after Hsp90 inhibition. Using various Hsp90 inhibitors (18, 19, 22–26) as well as anti-Hsp90 hammerhead ribozymes we demonstrated that besides a putative increase in membrane fragility by geldanamycin-induced superoxides, inhibition or lack of Hsp90 alone also results in a compromised cellular integrity. Moreover, cell lysis after hypoxia and complement attack was also enhanced by any type of Hsp90 inhibition used, which shows that the maintenance of cellular integrity by Hsp90 is important in physiologically relevant lytic conditions of tumor cells. Our results show the first successful use of an anti-Hsp90 ribozyme in manipulating Hsp90 levels, and demonstrate a novel element of Hsp90-related cytoprotection: its role in the maintenance of cellular integrity.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Culture of Jurkat Cells and Drug Treatments—The human T-lymphocyte cell line (Jurkat, J32) was provided by M. Kamoun (Department of Pathology and Laboratory Medicine, Philadelphia, PA) and was cultured as described earlier (26). For all experiments 2 x 105 cells (at a cell density of 2 x 105 cells/ml) were used unless otherwise indicated. Drug treatments were always carried out in complete medium unless otherwise mentioned, for2hat37 °C with 5% constant CO2 supply. Cell viability was monitored by trypan blue exclusion all the time before starting the experiment. Concentrations of the N-terminal Hsp90 inhibitors, geldanamycin (18, 19) and radicicol (22), the ineffective geldanamycin analogue, geldampicin (18), the C-terminal inhibitor, cisplatin (22, 24) and novobiocin, which inhibits nucleotide binding to Hsp90 at both termini (24, 25) were optimized using the drugs at concentrations, where their Hsp90 inhibition is fully exerted, but no drug-induced cell toxicity is seen. Under "no-shaking" conditions cells were maintained at 25 °C in an Eppendorf incubator for 10 min after drug preincubation. Shaking of cells was performed for 10 min in an Eppendorf Thermomixer at a speed of 1000 rpm. Lysis conditions were standardized to have 20–25% lysis. This has been achieved using 0.005% Brij-58 for 10 min at room temperature. 100% cell lysis was obtained by sonicating cells for 30 s with a sonicator (Sonic 300, Artek Systems, Farmingdale, NY) on ice. After detergent treatment or sonication cells were centrifuged at room temperature for 10 min at 800 x g in an Eppendorf centrifuge (Model 5402). Supernatants were processed for protein determination, lactate dehydrogenase analysis, or immunoblot measurements.
Protein Content—Protein concentration of cell lysates was measured by the Bradford method (27) using bovine serum albumin as standard.
Lactate Dehydrogenase Measurements—The activity of lactate dehydrogenase (LDH) was measured using the direct spectrophotometric assay of Wroblewski and LaDue (28) in the presence of pyruvate and NADH. In 2 ml of a 50 mM Hepes buffer (pH 7.4) containing 30 µM pyruvate and 30 µM NADH 10 µl of Jurkat cell supernatant were added after the indicated treatments with drugs and detergents, and changes in optical density were measured at 340 nm for 5 min. Special care was taken to avoid the absorbtion of LDH to the cuvette wall during the activity measurement. The percentage of LDH release was calculated by dividing the activity of LDH in the supernatant by the LDH activity measured after complete cell lysis achieved by sonication. None of the detergents and drugs affected LDH activity, when added directly to the reaction mixture at the concentrations used in whole cell experiments.
Superoxide Production—Superoxide production in geldanamycin-treated Jurkat cells was assayed by the lucigenin-enhanced chemiluminescence method (29). 2 x 106 Jurkat cells (2 x 106 cells/ml) after a treatment with various drugs were washed free of the used inhibitors, resuspended in 10 mM Hepes (pH 7.4), and added to the scintillation vial containing 5 µM lucigenin in the same Hepes buffer. Lucigenin chemiluminescence with and without cells was recorded as cpm (counts per minute) for about 10 min in 0.1-min intervals in a Beckman LS7800 liquid scintillation counter using a single photon mode. Chemiluminescence values were corrected to background chemiluminescence without added cells.
Membrane Fluidity Measurements—The procedure was adapted from Revathi et al. (30). Jurkat cells after a treatment with the drug indicated were incubated with 1 µM 1,6-diphenyl-1,3,5-hexatriene (DPH) for 15 min in the dark in triplicates. Cells were washed free of DPH, resuspended in 2 ml of phosphate-buffered saline, and the fluorescence was immediately measured using a steady-state spectrofluorimeter (M300-Edinburgh Instruments) in 1-cm path length quartz cuvettes with an excitation and emission wavelength of 357 and 430 nm, respectively. Fluorescence polarization was measured after adapting cells in dark with constant stirring.
Construction of Anti-Hsp90 Hammerhead Ribozymes—The design
of anti-Hsp90 hammerhead ribozyme was adapted from the studies of Little and
Lee (31). The ribozyme was
designed to cleave in a highly conserved segment of the coding region of Hsp90
at its N or C terminus, respectively. Partially overlapping nucleotides for
both Hsp90
and Hsp90
N-terminal and C-terminal regions were
synthesized. These sense and antisense oligonucleotides were annealed at room
temperature for 15 min in a 50 mM Tris-HCl (pH 7.6) buffer. Prior
to cloning, these oligonucleotides were further extended with 5 units of
Klenow polymerase in a standard PCR buffer (50 mM Tris-HCl, pH 8.0,
10 mM MgCl2, 50 mM NaCl, and 10
µM dNTP mix) for 60 min at 37 °C. The end-filled products
were precipitated using ethanol, resuspended in sterile double distilled
water, and double digested with HindIII and NotI. For
cloning the cloning vector, pcDNA3 with an hCMV promoter and a polylinker
containing the bovine growth hormone poly(A) sequence was selected. In the
vector the upstream ATG (at nucleotide position 995) after the XbaI
and ApaI sites was disrupted by introducing a stop codon followed by
XhoI, NheI, NotI, and SacII restriction
sites. The vector was digested with HindIII and NotI
restriction enzymes, the digested vector was dephosphorylated using calf
intestinal phosphatase at 37 °C for 15 min and the enzyme was inactivated
by phenol/chloroform extraction. The anti-Hsp90 hammerhead ribozyme
oligonucleotides were then inserted as described in Sambrook et al.
(32) at the HindIII
and NotI sites within the multiple cloning site of pcDNA3. Ligation
reactions were carried out in a total volume of a 10 µl buffer containing
50 mM Tris-HCl, pH 7.6, 10 mM MgCl2, 1
mM ATP, 1 mM dithiothreitol, 5% polyethylene
glycol-8000, the double-digested insert and the vector in a molar ratio of 3:1
as well as 1 unit of T4 DNA ligase at 16 °C for 12 h. XLN-blue competent
cells were transformed with the ligation mix, and plated on Luria-Bertani (LB)
agar plates containing 60 µg/ml ampicillin. The colonies were selected from
the plate, and were subjected to plasmid miniprep. The positive clones were
selected by Southern blot analysis using the respective end-labeled, annealed
antisense oligonucleotide.
Transfection and Screening of Jurkat Cells—Jurkat cells were transfected with either pcDNA3 (as a control), or the respective anti-Hsp90 hammerhead ribozyme using the polycationic reagent, LipofectAMINE. The entire transfection protocol was performed according to the manufacturer's instructions in a 96-well plate (NUNCTM, Nalgene Nunc International, NY) in multiples. Control and ribozyme-transfected cells were screened for cell viability using trypan blue dye exclusion. Neomycin selection was not feasible, since the rate of cell death with ribozyme transfections was constantly increasing showing the importance of high Hsp90 levels for the survival of eukaryotic cells. Hsp90 content was checked by an immunoblot with anti-Hsp90 antibody. Cells on the second day of transfection, where cell death was not yet much prevalent were used for further experiments.
SDS-Polyacrylamide Gel Electrophoresis and Western Blot Analysis—Cell lysates were mixed with Laemmli buffer (33) containing 100 µM dithiothreitol, boiled for 5 min, and the samples were subjected to 10% SDS-PAGE. Proteins were transferred from the gel to nitrocellulose membrane using a semidry protein gel transfer apparatus. Transfer of proteins was confirmed by Ponceau-S staining, and the blot was processed for Western blot analysis using a primary antibody followed by a horseradish peroxidase-conjugated secondary antibody. Labeled bands were visualized using an enhanced chemiluminescence kit.
Bacterial Culture and Bacterial Protoplasts—DH5-
(34) Escherichia coli
strain (a gift from Dr. András Váradi, Institute of Enzymology,
Hungarian Academy of Sciences, Budapest, Hungary) was grown in liquid LB.
Bacteria from the mid-log phase of growth (at 37 °C) were collected and
were subjected to mild lysozyme treatment to digest the cell wall. Protoplasts
were collected by spinning cells at 2000 rpm for 2 min. Protoplasts were
stored in 0.9% NaCl isotonic solution, and the osmotic lysis conditions were
standardized as to get 20–25% of lysis. This has been achieved using
0.4% NaCl. Incubation with various concentrations of geldanamycin (from 0.001
to 5 µM) was performed at 25 °C for 60 min to see the effect
of the drug on cell lysis. Distilled water was used to achieve 100% lysis of
protoplasts.
Yeast Cultures, Yeast Spheroplasts, and Protoplasts—Wild-type strains of bakers' yeast, Saccharomyces cerevisiae (S-288) and fission yeast, Schizosaccharomyces pombe (L-972) were obtained from Department of Microbiology and Biotechnology, St. István University, Budapest, Hungary and were grown in liquid YEPD medium. Yeast cells from the mid-log phase of growth (at 30 °C) were used for making yeast spheroplasts and protoplasts. First, yeast cells were treated with 0.1 mg/ml of lyticase in case of S. cerevisiae (35) and 1 mg/ml lysing enzyme in case of S. pombe (36) at 30 °C for 30 or 45 min to make spheroplasts or protoplasts, respectively. Spheroplasts and protoplasts were collected in isotonic 1 M sorbitol, and were incubated with geldanamycin. Lysis conditions were standardized to have 20–25% lysis. This has been achieved using 0.65 M sorbitol. 100% cell lysis was achieved by treating cells with distilled water.
Hypoxia-induced Cell Lysis Measurements—The method to induce chemical hypoxia was adapted from Wang et al. (37). A stock solution of cobalt chloride was made in sterile double distilled water and various concentrations (25 µM to 2 mM) were used for a 2-h treatment to induce hypoxia in Jurkat cells. 0.2 mM CoCl2 concentration for 2 x 106 cells/ml was chosen as the optimum to have 15 to 20% cell lysis, without any cellular damage (cell integrity was measured by trypan blue exclusion). The treatment was carried out both in presence and absence of 2 mM glutathione along with various Hsp90 inhibitors.
Complement-induced Cell Lysis Measurements—2 x
106 Jurkat cells/ml were subjected to immune-mediated cell lysis
using mouse and human serum in separate experiments. Serum from mouse showed
very high cytolysis even at very low serial dilution (data not shown), hence,
human serum was chosen for the experiments shown. The human serum
concentration was optimized to have 
20% cell lysis (judged by both
protein estimation and microscopic examination). Exponentially growing Jurkat
cells were treated with various anti-Hsp90 drugs for 2 h both in presence and
absence of 2 mM glutathione followed by addition of 5 µl of 1:5
diluted human serum in 0.5 ml RPMI 1640-complete medium (serum enhanced the
complement-mediated cell lysis hence, all the lysis experiments were performed
in complete medium). Cells were then further incubated for 10 min at 30
°C. Cells were subsequently washed with phosphate-buffered saline (pH 7.6)
and centrifuged for 10 min at 2000 rpm and their supernatant was collected for
measuring the cell lysis as the percent of total cytoplasmic protein
released.
Statistical Analysis—Data are presented as means ± S.E. of minimum three independent experiments unless otherwise indicated, and analyzed with unpaired Student's t test. p < 0.05 was accepted as indicating a statistically significant difference compared with controls. In figure legends: *, p < 0.05;**, p < 0.01; ***, p < 0.001.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
20% lysis of lactate dehydrogenase (LDH) in a
10-min detergent treatment (Fig.
1A, open bars). Preincubation of Jurkat cells
with geldanamycin, a specific inhibitor of Hsp90
(18,
19) induced a significant
increase in the extent of released LDH
(Fig. 1A, filled
bars). In agreement with previous data
(18,
19,
26) geldanamycin treatment
alone, without additional Brij-58 lysis did not induce a significant lysis of
Jurkat cells (data not shown). Geldanamycin-induced lysis was not specific to
Brij-58, but could be observed if primary cell lysis was induced by other
detergents, like the plasma membrane specific digitonin or by hypotonic shock
(data not shown). The effect was concentration dependent, showing a saturation
above 2 µM geldanamycin (Fig.
1B and data not shown).
|
The geldanamycin-induced enhancement of the release of lactate dehydrogenase, a well-known cytoplasmic marker protein might be the consequence of a heretofore unknown interaction between LDH and Hsp90, where geldanamycin would dissociate LDH from Hsp90, and cause its accelerated release. However, the geldanamycin-induced enhancement of cell lysis was a general feature of all cytoplasmic proteins, including Hsp90 itself, Hsp90 substrates, like the Raf and Lck kinases and total cytoplasmic proteins (Fig. 1C and data not shown). The extent of maximal release was similar for LDH and total proteins showing that the inhibition of Hsp90 induced a fairly general destabilization of cellular structures. Geldanamycin-induced enhancement of cell lysis was not instant but required at least a 1-h preincubation with the drug to be effective (Fig. 1D). Our data are in good agreement with the concentration and time dependence of geldanamycin-induced changes in Hsp90 substrate proteins (18, 19, 26).
The geldanamycin-induced additional cell lysis was similar to the increase in cell lysis after disruption of cytoskeletal elements with colchicine or cytochalasin (data not shown). However, the effect of geldanamycin was not changed (data not shown) if lysis was performed in the presence of actin or tubulin stabilizing buffers (38, 39), which suggests that the geldanamycin-induced effect is not a direct consequence of cytoskeletal disorganization.
A highly similar pattern of geldanamycin-induced increase in Brij- or hypoosmotic shock-induced hemolysis was observed in mouse red blood cells. Time and concentration dependence of geldanamycin-induced additional hemolysis was very similar to the lysis rates obtained with Jurkat cells (see abstract published as Ref. 40, and data not shown) showing the generality of the effects observed.
Geldanamycin-induced Additional Cell Lysis Is Partially Dependent on Oxidative Stress—Interestingly, there was a marked difference in the extent of geldanamycin-induced additional lysis, if cells were shaken during the experiment or not. When geldanamycin treatment was combined with additional shaking, a larger increase in cell lysis was observed (Fig. 2A). Shaking the cells after geldanamycin treatment, but before detergent treatment might help to disrupt cellular structures otherwise preventing the faster release of LDH and cytoplasmic proteins in the presence of Brij-58. However, shaking-induced differences were larger at larger geldanamycin concentrations (Fig. 2B), which suggested that the two effects are not independent from each other. Reaching this point in our experiments the first results of geldanamycin-induced superoxide generation appeared (20, 21). What if shaking provided an additional oxygen inducing a larger amount of superoxides? The addition of reduced glutathione as an antioxidant to the incubation medium reduced the geldanamycin effect (Fig. 2C). On the contrary, the effect of other Hsp90 inhibitors, like the much more effective N-terminal inhibitor, radicicol (22), the C-terminal inhibitor, cisplatin (23, 24) and novobiocin, which inhibits nucleotide binding to Hsp90 at both termini (24, 25) was not changed by glutathione addition (Fig. 2C). Geldampicin, an ineffective geldanamycin analogue (18) did not induce a large increase in cell lysis, and the effect of all other Hsp90 inhibitors was inbetween the control and geldanamycin-induced level being roughly equal with the lysis after the simultaneous addition of geldanamycin and reduced glutathione (Fig. 2C). These data raised the possibility that the geldanamycin-induced additional cell lysis was a result of both geldanamycin-induced superoxide production and Hsp90 inhibition.
|
Geldanamycin Induces Superoxide Production in Jurkat Cells and an Increased Membrane Fluidity—Observing an oxidative stress-related component of geldanamycin-induced additional cell lysis we wanted to obtain a direct evidence for geldanamycin-induced superoxide production in Jurkat cells. Indeed, geldanamycin, but not the inactive analogue, geldampicin (18) induced a significant increase in lucigenin chemiluminescence (Fig. 3A), which is a clear indication of geldanamycin-induced superoxide production in Jurkat cells.
|
To assess, if geldanamycin affects the status of Jurkat cell membranes we opted to measure membrane fluidity by measuring the fluorescence polarization of the commonly used probe, 1,6-diphenyl-1,3,5-hexatriene (DPH, Ref. 30). Geldanamycin induced a significant decrease in DPH fluorescence polarization (Fig. 3B). The change in fluorescence polarization (0.034) was higher than several physiologically significant changes reported in the literature (41–43). On the contrary to our results with geldanamycin, another N-terminal inhibitor of Hsp90, radicicol (22) did not induce a significant change in DPH fluorescence polarization (data not shown). Since detergent treatment causes difficulties in the interpretation of fluorescence polarization data, these experiments did not give a direct analysis of the membrane status after a combined treatment of Hsp90 inhibitors and Brij-58. However, our data suggest that geldanamycin-induced superoxide production may lead to an increased membrane fluidity, membrane disorganization in Jurkat cells, which may contribute to their increased sensitivity to physiological lysis conditions.
Anti-Hsp90 Hammerhead Ribozymes Reveal a Truly Hsp90-dependent
Component of Geldanamycin-induced Additional Lysis of Jurkat
Cells—Since none of the Hsp90-inhibitors have truly specific
effects to Hsp90, (geldanamycin induces superoxide production independent of
Hsp90, Refs. 20 and
21; radicicol is an inhibitor
of citrate lyase, Ref. 44;
cisplatin and novobiocin both have a wide spectra of effects at the
concentration they inhibit Hsp90) we wanted to use a tool which really
specifically inhibits Hsp90 to assess the ratio of superoxide-dependent and
Hsp90-dependent components of geldanamycin effects. Utilizing the idea of
Little and Lee (31) to
diminish Grp94, we constructed two anti-Hsp90 hammerhead ribozymes specific to
the N and C termini of Hsp90 mRNA (Fig.
4A). Both ribozymes cleave a sequence, which is the same
in Hsp90- and Hsp90- and conserved in a wide range of species
(Fig. 4B).
Transfection of Jurkat cells with the anti-Hsp90 hammerhead ribozymes resulted
in an efficient reduction of Hsp90 levels, while keeping the pattern of total
cellular proteins intact (Fig.
4C). According to our expectations, both the N-terminal
and the C-terminal anti-Hsp90 ribozymes, either alone or in combination
induced a significant acceleration of Brij-induced Jurkat cell lysis
(Fig. 4D; the
difference between Me2SO- and GA-treated cells was statistically
significant at a level of p < 0.001). Interestingly, the
C-terminal ribozyme was somewhat less efficient than the N-terminal, which
might be related to its smaller degree of homology
(Fig. 4A). On the
contrary, LipofectAMINE-treatment or vector transfection had only marginal
effects (Fig. 4D).
Ribozyme-induced additional lysis could be further increased by geldanamycin
treatment. This increase was normalized by the addition of glutathione in all
cases (Fig. 4D).
Substitution of glutathione with the superoxide scavenger, Tiron
(45) or the NO synthase, FADH
and NADPH oxidase inhibitor, diphenyleneiodonium chloride (DPI; Refs.
21 and
46) gave smaller, but similar
effects (data not shown). The effect of DPI was the smallest of all
antioxidants studied, showing a rather small contribution of specific redox
systems to Jurkat cell lysis. The ribozyme experiments gave a strong support
to our conclusion that geldanamycin induces a destabilization of Jurkat cells
by both superoxide- and truly Hsp90-dependent mechanisms. Hsp90 seems to be
important in the maintenance of cellular integrity.
|
|
Geldanamycin-induced Lysis Acceleration in Bacterial and Yeast Cells—Observing a role for Hsp90 to maintain cellular integrity we wanted to study, whether cells with a less important contribution of Hsp90 to their viability and with a smaller degree of cellular organization than the human Jurkat cell line, namely bacterial or yeast cells are sensitive to geldanamycin-induced additional lysis or not. Lysis conditions were optimized to achieve a 20–30% lysis of total cytoplasmic proteins. On the contrary to our results with Jurkat cells, geldanamycin induced no significant additional lysis in E. coli (Fig. 5). In S. cerevisiae and S. pombe yeast cells the additional lysis after geldanamycin treatment was significant, but smaller than that of Jurkat cells (see Figs. 1B and 5A). As in the case of Jurkat cells, the lysis rate was strongly dependent on the extent of shaking in all cell types studied. No shaking induced no appreciable lysis, however, vigorous shaking induced a close-to-maximal cell lysis suggesting that bacterial and yeast protoplasts lacking a sophisticated cellular architecture are more sensitive to cell lysis than eukaryotic cells (data not shown and Fig. 5B). Since we could not be sure that geldanamycin inhibits the bacterial Hsp90 homologue, HtpG similarly than eukaryotic Hsp90 we have heat preconditioned bacterial cells at 44 °C and measured their lysis rate. Elevation of molecular chaperones in bacterial cells did not induce any change in cell lysis rate (data not shown) indicating that in bacterial cells chaperones are not playing a prominent role in the maintenance of cellular integrity.
|
Hsp90 Inhibition Enhances Hypoxia-induced Cell Lysis—To
address the physiological significance of our cell lysis experiments using
mild detergent treatment, we wanted to study if Hsp90 inhibition induces an
enhanced cell lysis in hypoxia, a usual phenomenon in tumors. Hypoxia was
induced using cobalt chloride
(37). Geldanamycin and
radicicol both caused a significant increase of hypoxia-induced cell lysis in
Jurkat cells. The effect was reduced in presence of glutathione
(Fig. 6A). Other Hsp90
inhibitors, such as cisplatin and novobiocin showed 50% less enhancement
of hypoxia-induced Jurkat cell lysis, and, on the contrary to geldanamycin and
radicicol, addition of glutathione had no significant effect on the extent of
lysis (Fig. 6A). To
see and compare the differences between Hsp90 inhibitor-induced cellular
effects and those after ribozyme-targeted Hsp90 inhibition,
ribozyme-transfected Jurkat cells were subjected to hypoxia and
hypoxia-induced cell lysis was amplified to more than 2-fold in
ribozyme-transfected cells (Fig.
6B). Vector-transfected cells did not show any
significant change. Addition of geldanamycin to ribozyme-transfected cells
showed no further change in hypoxia-induced cell lysis. Similarly, addition of
glutathione did not reduce the combined effects of anti-Hsp90 ribozyme and
geldanamycin (Fig.
6B).
|
Hsp90 Inhibition Enhances Complement-mediated Cell Lysis—To demonstrate the physiological significance of the enhanced cell lysis associated with Hsp90 inhibition further, we have chosen the complement-mediated immune lysis of the malignant Jurkat cell line, as a model of immune attack on human tumor cells. Though all effective Hsp90 inhibitors showed some increase of cytolysis when Jurkat cells were preincubated with them before the addition of human serum, again geldanamycin and radicicol showed the only significant differences (Fig. 6C). Addition of glutathione induced a slight decrease in cell lysis (Fig. 6C). The Hsp90 antisense ribozyme induced a massive increase in complement-mediated cell lysis, which was only marginally affected with the addition of geldanamycin without or with glutathione suggesting that Hsp90 loss is the major cause of increased cytolysis after complement attack (Fig. 6D).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Accelerated superoxide production is a recently observed feature of geldanamycin, the first Hsp90-independent function of this highly specific Hsp90 inhibitor (20, 21). Superoxide production has been first implicated in geldanamycin action as an alternative route of endothelial NO synthase function leading to an uncoupled superoxide production parallel with a decreased NO synthesis (47). The role of Hsp90 as a "switch" from superoxide to NO production was also demonstrated in neural NO synthase (48). However, later studies extended these findings and showed that geldanamycin is able to increase superoxide production independently of endothelial NO synthase activation both in in vivo and in vitro systems, possibly by its quinone group, which may participate in redox cycling (21). Similar data were reported on neural NO synthase (20). Importantly, the non-quinone Hsp90 inhibitor, radicicol had no direct superoxide producing effect (20). This is in agreement with our findings, that in contrast to the effects of geldanamycin, radicicol action cannot be attenuated by reduced glutathione (Fig. 2C). An earlier study proposed that radicicol might exert its antimalarial action via heme-dependent free radical generation but this assumption has not been tested directly (49). However, a recent study indicated that radicicol converts endothelial NO synthase from an NO producer to a superoxide generator independently of a direct superoxide production (50). Interestingly, the geldanamycin structural analogue, geldampicin was much less potent superoxide activator than the parent compound (Fig. 3A). Geldampicin contains a piperidine derivative directly attached to the quinoidal segment of the drug (18), which possibly hinders its participation in redox cycling by stabilization redox-independent oxygen binding to geldampicin.2
Geldanamycin-induced superoxides may induce lipid-peroxidation, which would lead to a significant membrane damage partially explaining the increased cell lysis observed in our study. However, our membrane fluidity data suggest that this might not be the case. Lipid peroxidation induces a decrease in membrane fluidity (51, 52). In contrast, we have found an increased membrane fluidity after geldanamycin action (Fig. 3B). Increased membrane fluidity may be concomitant with an increased membrane fragility (53). Thus geldanamycin seems to contribute by a partially oxidation-dependent, but heretofore unknown mechanism to the observed increase in membrane sensitivity. It cannot be excluded, however, that Hsp90 inhibition alone, independently of any geldanamycin-induced oxidative changes also induces an increase in membrane disorganization.
As a possible example of oxidation-independent membrane action of geldanamycin, Suttitanamongkol et al. (54) published an interesting study, which showed that geldanamycin treatment alone induced a lysis of platelets by inducing extensive membrane blebbing and disrupting the plasma membrane structure. These data raise the possibility that the geldanamycin-enhanced lysis of Jurkat cells is due to a detergent effect of the drug. However, several of the following pieces of evidence refute this assumption. (i) In our studies geldanamycin alone did not induce a significant lysis of Jurkat cells (Ref. 26 and data not shown). (ii) Suttitanamongkol et al. (54) used geldanamycin at a high concentration (18 µM), which most probably produced aspecific effects. A recent study (55) showed that geldanamycin is accumulated in various cells. However, in our experiments detergent treatment was performed after the excess of geldanamycin has already been washed away, and the geldanamycin-induced sensitization for lysis also occurred in experiments (i.e. when applied together with hypoxia or complement-induced lysis) when detergent treatment was not performed. This excludes the possibility that geldanamycin was especially enriched in detergent-treated cells. (iii) The effect of geldanamycin was gradually developed over a long preincubation period extending for several hours in our experiments (see Fig. 1C). This slowly developing action is a typical feature of geldanamycin effects (18, 19, 26). On the contrary, detergent effects are prompt, requiring seconds to minutes to develop. (iv) The structural analogue of geldanamycin, geldampicin did not induce the acceleration of Jurkat cell lysis (Fig. 2C). Detergent effects are not so sensitive to minor changes in detergent structure. (v) The structurally unrelated Hsp90 inhibitors, radicicol, cisplatin or novobiocin gave a smaller, but significant increase in Jurkat cell lysis compared with that observed after geldanamycin treatment (Fig. 2C). (vi) Last but not least, we have analyzed Jurkat cell structure after geldanamycin treatment with electron microscopy and saw no signs of membrane blebbing or disintegration (data not shown).
Recent studies indicated a role of Hsp90 in maintaining the membrane raft structure. Geldanamycin treatment efficiently dissociated several Hsp90-related protein complexes in lipid rafts (56, 57). Since lipid rafts are usually isolated after a nonionic detergent extraction, the similar experimental conditions argue for a role of Hsp90 inhibition induced raft disorganization in the decrease of cellular integrity observed in our studies.
Our experiments with anti-Hsp90 hammerhead ribozymes demonstrated that a diminished Hsp90 function alone compromises cellular integrity (Fig. 4D). The role of Hsp90 can be a direct disorganization of cellular structures after the disruption of various Hsp90-related complexes or an indirect action of Hsp90 inhibition via the incorrect folding of an Hsp90 client protein important in the maintenance of cellular integrity. At present we cannot exclude any of these possibilities. Our data showing a prolonged action after Hsp90 inhibition (Fig. 1C) suggest the contribution of an Hsp90 client protein to the effects observed.
The observation, that Hsp90 inhibition does not affect the lysis of bacterial protoplasts, and has a smaller enhancement of yeast protoplast lysis than that of higher eukaryotic cells (Fig. 5) is in good agreement with the increasing role of this chaperone in the maintenance of cellular life from prokaryotes to yeast and higher eukaryotes (1–5) as well as with the development of more and more sophisticated cytoskeleton in these organisms. Hsp90 is well known to interact with filamentous actin and tubulin (6–8) and the involvement of the cytoskeleton in the traffic of Hsp90 substrates has been repeatedly demonstrated (5, 9). These findings raise the possibility that Hsp90 might contribute to an increased cellular integrity by the maintenance of cytoskeleton-related structural elements in eukaryotic cells.
Hsp90 inhibitors (geldanamycin analogues, radicicol, and purine scaffold inhibitors) were recently introduced to the clinical practice as anticancer agents (58, 59). Our findings may help to establish a novel element of the mechanism of action of these drugs by showing the role of Hsp90 inhibition to sensitize cells for various lytic events. To assess the decreased cell stability after the inhibition of Hsp90 function in experiments, which are more relevant to physiological conditions than mild detergent treatment, or hypotonic shock, we have examined the effect of Hsp90 inhibitors and the disruption of Hsp90 by anti-Hsp90 ribozyme on hypoxia-induced and complement-mediated cytolysis of Jurkat cells. Both conditions mimic quite well the lytic conditions usual for tumor cells. Our results demonstrated a clear enhancement of cell lysis under both conditions after any type of Hsp90 inhibition used. Our findings suggest the probability of a general contribution of Hsp90 to maintain cellular integrity. The role of heat shock proteins in natural cell reactivity is well demonstrated (60). Similarly, immune cell-mediated lysis is also associated with the production of superoxides (61). Our results show that both effects take place when Hsp90 inhibitors enhance cell lysis, but our anti-Hsp90 ribozyme experiments show that the inhibition of Hsp90 itself, seems to be predominant. Sodium arsenite was shown to sensitize Jurkat cells for immune-mediated cytolysis (62). However, in this case the complexity of the stress response both in tumor and bystander cells as well as the relative toxicity makes the treatment non-suitable for selective tumor therapy. On the contrary, selective depletion of Hsp90 seems to be an effective mode of cell sensitization to both hypoxia- and immune-mediated cell lysis, which adds a novel element to the mechanism of action of Hsp90 inhibitor drug candidates. This phenomenon may help the immune system to attack tumor cells. Similarly, a lysis sensitization may cause a shift from tumor cell apoptosis to necrosis, which gives a further help for the activation of the immune system (63).
In summary, our experiments demonstrated that a diminished Hsp90 function alone compromises cellular integrity. This effect is characteristic to eukaryotic organisms, and is extended by a putative increase in membrane fragility partially due to the increased superoxide production by geldanamycin, an Hsp90 inhibitor. Our results show the first successful use of an anti-Hsp90 ribozyme in manipulating Hsp90 levels, and demonstrate a novel element of Hsp90-related cytoprotection: its role in the maintenance of cellular integrity.
|
FOOTNOTES |
|---|
Recipient of a National Overseas Fellowship of the State of India. On leave
from the Center for Cellular and Molecular Biology, Hyderabad, 500 007
India.
High school student of the St. István High School of the Cistercian
Order in Székesfehérvár, Hungary obtaining a 2nd prize of
the 13th European Union Contest for Young Scientists for his contribution to
this work as well as a chance to participate in the 2001 Nobel Ceremonies.
** To whom correspondence should be addressed: Semmelweis University, Dept. of Medical Chemistry, P. O. Box 260, H-1444 Budapest 8, Hungary. Tel.: 36-1-266-2755 (ext. 4102); Fax: 36-1-266-7480; E-mail: csermely{at}puskin.sote.hu.
1 The abbreviations used are: Hsp90, 90 kDa heat shock protein; Brij,
Brij-58: C16E20, polyethylene glycol hexadecyl ether;
CDDP, cisplatin; DPH, 1,6-diphenyl-1,3,5-hexatriene; DPI, diphenyleneiodonium
chloride; GA, geldanamycin; GP, the inactive geldanamycin analogue,
geldampicin; GSH, reduced glutathione; LDH, lactate dehydrogenase; NB,
novobiocin; RA, radicicol; Tiron, 4,5-dihydroxy-1,3-benzene disulfonic acid
disodium salt.
2 Z. Riedl and J. Jakus, personal communication.
|
ACKNOWLEDGMENTS |
|---|
E. coli strain. |
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
