| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Epidemiology and Prevention |
Department of Biochemistry, Boston University School of Medicine, Boston, Massachusetts
Requests for reprints: Michael Y. Sherman, Department of Biochemistry, Boston University Medical School, 715 Albany Street, Boston, MA 02118. Phone: 617-638-5971; Fax: 617-638-5339; E-mail: sherman{at}biochem.bumc.bu.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A body of literature indicates that normal tissues usually express a constitutive member of the Hsp70 family, Hsc73, but not Hsp72, which is induced only under stressful conditions that cause protein damage. In contrast, tumors often constitutively express both Hsc73 and Hsp72 at high levels (see refs. 6–8 for review). Furthermore, high expression of Hsp72 in human tumors correlates with high invasiveness, metastasis, resistance to chemotherapy, and poor prognosis of the disease (see ref. 9 for review). For example, in colorectal and lung cancers, levels of Hsp72 expression closely correlate with advanced clinical stages and positive lymph node involvement (10, 11). These findings suggest that Hsp72 provides a selective advantage to tumor cells during cancer progression.
A possible association of Hsp72 with cancer development was shown in several works where overproduction of recombinant Hsp72 in various systems promoted cancerous properties of cells. Accordingly, overproduction of Hsp72 in fibrosarcoma cells significantly enhanced their tumorigenic potential in athymic mice (12), whereas transgenic mice that expressed human Hsp72 at high levels developed multiple lymphomas (13). Moreover, expression of Hsp72 in Rat-1 fibroblasts led to formation of foci, anchorage-independent growth, and development of tumors in nude mice (14). It is believed that the antiapoptotic activity of Hsp72 plays an important role in cancer development. For example, Hsp72 may allow cancer cells to withstand apoptosis caused by stressful factors of the tumor microenvironment, such as hypoxia, starvation (15), TNF (16), or FAS (17). Apparently, in line with this suggestion, it was shown that depletion of Hsp72 by adenovirus-encoded Hsp72 antisense RNA causes rapid death of various types of tumor cells, whereas nontransformed cells remain resistant to such treatment (18, 19). However, most of these effects were probably caused by a subtle toxicity of adenoviruses used in these experiments, which, in combination with Hsp72 depletion, led to massive caspase-independent death of cancer cells associated with activation of cathepsin B (20). Accordingly, we have shown that although apoptosis of prostate carcinoma cells PC-3 or DU-145 can be seen after infection with adenoviruses expressing Hsp72 antisense RNA, no such apoptosis was observed when the Hsp72 depletion was achieved by delivery of this same antisense RNA or small interfering RNA (siRNA) via nontoxic retroviruses (21). Therefore, high expression levels of Hsp72 in certain cancer cell lines seem to be unrelated to protection from internal proapoptotic signals.
Besides apoptosis, cell proliferation could be limited by senescence. Cancer cells, although able to divide indefinitely, can nevertheless undergo senescence upon exposure to certain anticancer drugs and radiation therapy (22). In fact, it seems that activation of the senescence program and consequent permanent growth arrest significantly contributes to the loss of the clonogenic capacity of tumor cells and probably to tumor regression after anticancer therapy (23).
Interestingly, it was recently reported that depletion of Hsp70-2, a minor member of the Hsp70 family that is critical for testes development, led to growth arrest and senescence of tumor cells, which was associated with induction of the Cdk inhibitory protein p21 (24). Therefore, it seems that Hsp70-2 may play an important role in supporting normal growth of cancer cells. It is not clear, however, how this protein is related to cancer development because no association of Hsp70-2 expression with tumor properties has been reported thus far.
Here, to address a role of the major heat shock protein, Hsp72, in cancer cells, we used retroviral vectors to deliver siRNAs that target Hsp72 in a range of cancer cell lines. It seems that Hsp72 depletion by this method led to activation of the p53 pathway and cell senescence. We further investigated the mechanisms of activation of the p53 pathway upon Hsp72 depletion and the pathway that is involved in senescence of p53 knockout cells under similar conditions.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Irradiation of cells was done using 137Cs source (GammaCell 40, Nordion International, Inc., Ontario, Canada) at 64 rad/min. UVC irradiation was done with UV Stratalinker 1800 from Stratagene (La Jolla, CA).
Doxorubicin, MG-132, PD98059, and wortmannin were from Biomol (Plymouth Meeting, PA); emetine and caffeine were from Sigma (Atlanta, GA); and the ataxia-telangiectasia-mutated (ATM) inhibitor KU55933 was from KuDos Pharmaceuticals (Cambridge, United Kingdom).
Recombinant retroviral vectors. For knockout experiments, we used RNAi-Ready pSIREN-RetroQ vector from BD Biosciences (San Jose, CA), containing a puromycin resistance gene for selection of stable transfectants. Three sequences of human Hsp72 gene were selected as targets for RNAi:
Retroviruses were produced by transfection of 293T cells with plasmids expressing retroviral proteins Gag-Pol, G (VSVG pseudotype), or enhanced green fluorescent protein (kindly provided by Jeng-Shin Lee, Harvard Medical School, Boston, MA) or our constructs. At 48 h after transfection, supernatants containing the retrovirus were collected and frozen at –70°C. Cells were infected with twice diluted supernatant and 10 µg/mL polybrene overnight and then washed. Selection with puromycin (0.5 µg/mL) was started 48 h after infection. Luciferase reporter plasmids with p21, MDM2, and Bax promoters were a kind gift of Dr. J. Xiao (Boston University Medical School, Boston, MA).
Immunoblotting and antibodies. Cells were lysed in lysis buffer [40 mmol/L HEPES (pH 7.5), 50 mmol/L KCl, 1% Triton X-100, 1 mmol/L Na3VO4, 50 mmol/L glycerophosphate, 50 mmol/L NaF, 5 mmol/L EDTA, and 5 mmol/L EGTA, supplemented with protease inhibitor cocktail. Hsp72, Hsc73, and Hsp27 were detected with corresponding antibodies from Stressgen (San Diego, CA); antibodies for Hsp70-2 were a kind gift of Dr. M. Jaattela (Apoptosis Department and Centre for Genotoxic Stress, Institute for Cancer Biology, Danish Cancer Society, Copenhagen, Denmark). Phosphorylation of p53 was assayed using phospho-p53 antibody sampler kit from Cell Signaling (Boston, MA). Other antibodies used were as follows: ß-actin (Sigma); phospho-ATM Ser1981 (Rockland, Gilbertsville, PA); phospho-(Ser/Thr) ATM/ATM- and Rad3-related (ATR) substrate, Cdc2, phospho-Cdc2 Tyr15, phospho-Chk1 Ser345, phospho-Chk2 Thr68 (Cell Signaling); Chk1, Mdm2, p53 (DO-1; Santa Cruz Biotechnology, Santa Cruz, CA); phospho-H2AX Ser139 (Upstate Cell Signaling Solutions); p16, p21, p27, Rb (BD PharMingen, San Diego, CA); ubiquitin (Zymed, San Francisco, CA); and phospho-p38 Thr180/Tyr182 (BioLabs, Ipswich, MA). Detection of oxidatively damaged proteins was done with Oxiblot-Oxidative Protein detection kit from Chemicon International (Billerica, MA), according to manufacturer's recommendation.
Luciferase assay. Cells were transfected with reporter plasmids encoding luciferase under cytomegalovirus (CMV), p21, MDM2, or BAX promoters. For each time point, three independent transfections were done. For refolding assay, cells were subjected to heat shock at 43°C for 20 min, and then allowed to recover for the indicated time periods. Forty-eight hours later, cells were washed in ice-cold PBS and lysed with cell lysis reagent (Promega, San Luis Obispo, CA). Samples were diluted with lysis buffer to achieve equal protein concentration in all samples. Assay was done in a 96-well plate using 20 µL of lysate per plate, in triplicates. Luminescence was read by a luminometer (Bio-Rad, Hercules, CA).
Clonogenic assays. Cells were counted and plated on 60- or 100-mm Petri dishes. After 10 days, the formed colonies were stained with 0.5% crystal violet in 70% ethanol and quantified using Quantity One software (Bio-Rad).
Bromodeoxyuridine incorporation assay. Bromodeoxyuridine (BrdUrd) incorporation assay was done with BrdUrd kit from BD PharMingen according to manufacturer's recommendations.
Fluorescence-activated cell sorting analysis. Cells were grown to 60% to 70% confluence, fixed in 63% iced-cold ethanol/PBS overnight at –20°C, stained with 50 µg/mL propidium iodine in the presence of 100 µg/mL RNase A, and analyzed using Becton Dickinson FACScan Cytometer.
ß-Galactosidase assay. ß-Galactosidase assay was done as described previously (25).
Dephosphorylation assay. Cells were washed with PBS and left in PBS supplemented with 5 µmol/L rotenone and 10 mmol/L 2-deoxyglucose to prevent further phosphorylation, as described previously (26).
Comet assay. Alkaline comet assay was done according to the manufacturer's protocol (Trevigen, Helgerman, CT).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
50% of cells became flat, vacuolized, and dramatically enlarged, which is typical for senescent cells, whereas the rest of the cells became moderately enlarged and did not look flattened (Supplementary Fig. S1B). No changes in cell morphology were seen with si3-infected cells or cells infected with empty retrovirus. There was also an 2-fold reduction in the overall growth rate of si2-infected cells compared with noninfected cells (Supplementary Fig. S1A). Of note, nontransformed breast epithelial cells MCF10F do not undergo senescence or growth inhibition in response to depletion of Hsp72 (not shown).
To further characterize cellular responses to Hsp72 depletion, we investigated the proliferation potential of HeLa cells by assessing BrdUrd incorporation. As can be seen in Fig. 2A
, depletion of Hsp72 reduced the fraction of BrdUrd-positive cells by 50%, indicating cessation of DNA synthesis in a population of cells. In line with this observation, fluorescence-activated cell sorting (FACS) analysis of Hsp72-depleted HeLa cells showed that the fraction of cells in S-phase dropped by 50% (Fig. 2B), whereas the fraction of G1 cells increased correspondingly (not shown). Interestingly, in HCT116 cells, growth retardation was associated mainly with an increase in a fraction of G2 cells (Fig. 2C). As an independent measure of the proliferation defect triggered by Hsp72 depletion, we measured the colony-forming ability of HeLa and HCT116 cells upon si2 infection, and observed 40% to 70% drop in cell clonogenicity in the Hsp72-depleted cells (Fig. 2D).
|
40% of ß-gal–positive cells compared with 5% to 8% in control population (Supplementary Fig. S1C; see also Fig. 6A below). Because the ß-galactosidase activity is very low in HeLa cells, we used a different approach to test whether enlarged, flattened cells are in fact senescent. As mentioned above, although a population of cells acquired morphologic features typical of senescence upon depletion of Hsp72 (40–50% of cells), the rest of the cells were somewhat enlarged but continued to grow, although at a slower pace. We hypothesized that this fraction of cells, although not fully senescent, is in a presenescent state. This hypothesis implies that these slower-growing cells may be prone to senescence when treated with low concentrations of anticancer drugs or subjected to low-intensity stresses. Therefore, we exposed control and Hsp72-depleted HeLa cells to the DNA-damaging agent doxorubicin. At concentrations below 20 nmol/L, doxorubicin did not induce senescence-like morphologic changes in control cells; however, it dramatically increased a fraction of cells with flat morphology in culture with down-regulated Hsp72 (Fig. 2E). An even more dramatic effect was seen with H2O2, which failed to cause senescence in control HeLa cells at any concentration (high concentrations led to cell death), while causing senescence of almost the entire population of Hsp72-depleted HeLa cells (Supplementary Fig. S1D). Similar enhancement of senescence was also observed in Hsp72-depleted HCT116 cells after treatment with low doses of doxorubicin or H2O2 (not shown).
|
Hsp72 depletion activates p53. The major tumor-suppressor protein p53 is implicated in regulation of cell senescence both in primary and cancer cells (22, 27). Therefore, we hypothesized that senescence triggered by depletion of Hsp72 in cancer cells may involve activation of p53 pathway. To address this question, we assessed transcriptional activity of p53 upon depletion of Hsp72. To achieve high efficiency of transfection, these experiments were done with HeLa cells. These cells were transiently transfected with three constructs in which the luciferase reporter was placed under the control of known p53-regulated promoters, Mdm2, p21, and Bax. To control for the efficiency of transfection, Hsp72-depleted and control cells were transfected with luciferase under the control of CMV promoter (CMV-Luc). As seen in Fig. 3A
, depletion of Hsp72 from HeLa cells led to a significant activation of p53-dependent promoters, whereas CMV-Luc expression was not affected. As an independent indication of enhanced p53 activity under these conditions, we observed an 3-fold induction of endogenous p21 (Fig. 3B). Induction of p21 upon Hsp72 depletion was also seen with other cancer lines, including HCT116 and MCF-7 cells (Fig. 3B). In addition, with HCT116 cells, we have found that depletion of Hsp72 leads to induction of Hdm2 (see Fig. 3C), a distinct major target of p53. Accordingly, the level of phosphorylation of p53 at the activating site Ser15 also increased after Hsp72 depletion in HCT116 (Fig. 3D) cells. However, the increase in Ser15 phosphorylation after Hsp72 depletion was much lower than that observed after DNA damage triggered by UV irradiation (Fig. 3C and D). Interestingly, nontransformed MCF10F cells did not show p53 activation in response to depletion of Hsp72, which correlated with lack of senescence under these conditions. Therefore, activation of p53 represents an important response of cancer cell lines to reduction of Hsp72 levels. It seems that increased levels of Hsp72 in cancer cells serve to prevent p53 activation and subsequently suppress the senescence program, and down-regulation of Hsp72 in these cells activates both p53 and cell senescence.
|
We next tested whether activation and Ser15 phosphorylation of p53 observed in the Hsp72-depleted cells result from the activation of DNA damage signaling pathways. Accordingly, we suggested that Hsp72 depletion leads to DNA instability and stimulation of ATM or ATR kinases. However, under these conditions, we did not detect any significant DNA damage in either HCT116 or HeLa cells as judged by alkaline comet assay that detects various types of DNA damage. Of note, irradiation–induced damage was readily detected (Fig. 4A
). Considering that DNA damage could be below the sensitivity of the comet assay, but still sufficient to activate ATM or ATR kinases, we investigated whether various substrates of these kinases are phosphorylated upon Hsp72 depletion. Accordingly, we assayed phosphorylation of the main substrates of ATM and ATR kinases, including histone H2AX, Chk1, and Chk2, by immunoblotting HCT116 cell lysates with antibodies against the phosphorylated forms of these proteins. No phosphorylation of any of these ATM/ATR substrates was seen after depletion of Hsp72, whereas DNA damage caused by or UV irradiation led to their phosphorylation (Fig. 4B). We also did immunoblotting of cell lysates with an antibody against a phosphorylated peptide that represents a consensus site of ATM/ATR kinases, and did not observe any difference in phosphorylation of ATM/ATR substrates in lysates from Hsp72-depleted cells (not shown). Furthermore, no autophosphorylation of ATM that is associated with its activation was observed (Fig. 4B). We also did not detect significant differences in histone H3-K9 acetylation, histone H4-K12 acetylation, or histone H3-K9 demethylation (not shown), which are seen in cells with impaired genomic stability (29). These data together indicate that Hsp72 depletion does not result in DNA damage, and ATM/ATR pathways were not activated under these conditions.
|
50 min (Fig. 5B). Hsp72 depletion and irradiation (5 Gy) of HCT116 cells led to similar stabilization of p53 (Fig. 5B) and to the similar extent of p21 induction (Supplementary Fig. S3). However, Ser15 phosphorylation after irradiation was dramatically higher than that after depletion of Hsp72 (Supplementary Fig. S3), suggesting that the extent of Ser15 phosphorylation is not indicative of p53 stability. To test whether Ser15 phosphorylation plays a role in p53 stabilization upon depletion of Hsp72, we expressed p53S15A mutant in HCT116p53–/– cells using a retroviral expression system. Stability of the mutant p53 was assessed in these cells before and after depletion of Hsp72, as described above. As seen in Fig. 5C, depletion of Hsp72 strongly stabilized p53S15A mutant, indicating that the stabilization is independent of the phosphorylation.
|
2-fold increase of Hdm2 levels in Hsp72-depleted cells (Fig. 3C). Because Hdm2 is a direct target of p53, this increase probably reflects the enhanced p53 activity. The fact that p53 was stabilized and active in spite of higher levels of Hdm2 indicated that Hdm2 had reduced activity under these conditions and therefore was not proficient in inhibiting and ubiquitinating p53.
Low activity of Hdm2 could result from its self-ubiquitination and destabilization. In fact, it was recently reported that Hdm2 self-ubiquitination abolishes its interactions with p53 and leads to Hdm2 inactivation (30). It is noteworthy that Hdm2 destabilization does not always correlate with decrease in its total levels, because Hdm2 degradation may be efficiently counteracted by enhanced p53-dependent transcription. Therefore, we hypothesized that depletion of Hsp72 may lead to Hdm2 self-ubiquitination and destabilization, resulting in its inactivation. To investigate the rate of Hdm2 degradation upon Hsp72 depletion, HCT116 cells were treated with emetine, and the levels of Hdm2 were assayed at the indicated time points. We observed that in spite of higher initial levels of Hdm2 in Hsp72-depleted cells, the rate of its degradation was 3-fold higher (Fig. 3D).
Because Hdm2 and p53 represent components of a negative feedback loop, which obscures analysis of the effects of Hsp72 depletion on Hdm2 degradation, it was critical to assess such effects in cells that lack p53. Accordingly, Hsp72-depleted HCT116 p53–/– cells were treated with emetine, and the levels of Hdm2 were assayed at the indicated time points. In these cells, similar to parental cells, Hsp72 depletion led to strong acceleration of Hdm2 degradation (Fig. 5E). Therefore, effects of Hsp72 on stability of Hdm2 are independent of the presence of p53. In conclusion, it seems that Hsp72 depletion leads to Hdm2 destabilization and inhibition, which results in activation of p53.
Senescence upon depletion of Hsp72 involves p53-dependent and p53-independent pathways. As mentioned above, a Cdk inhibitor p21 that is known to regulate p53-mediated senescence in primary cells was strongly induced in HeLa, HCT116, and MCF-7 cells upon depletion of Hsp72 (see Fig. 3B). To investigate whether the observed increase in p21 was responsible for the onset of senescence under these conditions, we took advantage of the p21 knockout derivative of HCT116 cells. In contrast to control HCT116 cells, in which depletion of Hsp72 was not toxic, depletion of Hsp72 in the HCT116 p21–/– cells led to death of 50% of cells (not shown). However, senescence in the remaining population of the knockout cells was significantly lower compared with the parental cells (Fig. 6A
), suggesting that induction of p21 upon Hsp72 depletion critically contributes to activation of the senescent pathway.
To investigate the role of p53 in induction of senescence after Hsp72 depletion, we used a p53 knockout derivative of human colon tumor cells, HCT116 p53–/–. In these cells, the background levels of p21 were dramatically lower than in parental cells (Fig. 6B). Depletion of Hsp72 in HCT116 p53–/– cells led to an 2-fold increase in p21 expression. However, the p21 levels seen under these conditions were not higher than the p21 levels seen in the naïve parental HCT116 cells (Fig. 6B, top). Furthermore, the activity of p21 in HCT116 p53–/– was also dramatically reduced as judged by the relief of a p21-dependent inhibition of topoisomerase II
expression (Fig. 6B, bottom). Surprisingly, in spite of a dramatic reduction of p21 levels and activity, Hsp72 depletion in the p53 knockout cells still led to significant senescence (Fig. 6A).
Senescence of p53 knockout cells caused by Hsp72 depletion was in apparent contradiction with the lack of senescence in p21 knockout cells because p21 levels are low in p53 knockout cells under these conditions (see Fig. 6B). Such a contradiction could be explained by extensive death of p21 knockout cells upon depletion of Hsp72 (see above), associated with strong overexpression of p53 in these cells (Supplementary Fig. S4B). Therefore, the lack of senescence in p21 knockout cells upon depletion of Hsp72 could potentially result from p53-induced death of cells that otherwise would become senescent. To further address a role of p21 in senescence, we investigated the response to Hsp72 depletion of cells with combined p53 and p21 depletion. HCT116p53–/– cells were infected with two retroviruses encoding siRNAs against p21 (Supplementary Fig. S4) and Hsp72. This double infection was not toxic, allowing us to assay the extent of senescence in these cells as described above. As seen in Fig. 6A, there was a small but reproducible reduction of the number of senescent cells upon Hsp72 depletion in double p53/p21–depleted cells compared with p53 depletion alone. These data indicate that p21 contributes to cell senescence upon depletion of Hsp72, but that an alternative p53/p21–independent senescence pathway is also triggered under these conditions.
The p53-independent pathway did not involve p27 or p16 Cdk inhibitors because p27 was not induced in either p53 knockout or parental HCT116 cells after Hsp72 depletion (not shown), and p16 is not expressed in this cell line (31). Of note, no accumulation of either p16 or p27 was detected upon Hsp72 depletion in HeLa cells as well (not shown).
To gain insight into the p53-independent pathway of senescence, we analyzed the effects of Hsp72 depletion on the cell cycle distribution of HCT116 p53–/– cells using FACS analysis. We observed that Hsp72 depletion in HCT116 p53–/–cells led to an increase of G2-M population (Fig. 6C). Because G2-M checkpoint is mainly regulated by Cdc2 kinase, and because G2-M–associated senescence in certain cells is mediated by Cdc2 down-regulation (32), we investigated the status of this kinase in HCT116 p53–/– cells upon Hsp72 depletion. Activation of the G2-M checkpoint typically leads to Cdc2 inactivation through inhibitory phosphorylation of Tyr15. We observed an 2-fold increase of Tyr15-phosphorylated form of Cdc2, indicating inhibition of this kinase. Moreover, the total level of Cdc2 was significantly decreased under these conditions (Fig. 6D).
It was previously shown that DNA-damaging stresses trigger phosphorylation of Cdc2 at Tyr15 by inhibiting or destabilizing Cdc2 phosphatases, Cdc25A or Cdc25C (see ref. 33 for review). Accordingly, we measured the rates of Cdc2 dephosphorylation in control and Hsp72-depleted cells (see Materials and Methods). Surprisingly, there was no significant difference in rates of Cdc2 dephosphorylation (Fig. 6E), indicating that Cdc25 phosphatases function normally in these cells. Therefore, Hsp72 depletion inactivates Cdc2 via a novel mechanism, which probably involves stimulation of Tyr15-specific kinases, Wee1 or Myt1, rather than suppression of Cdc2-specific phosphatases. Interestingly, we observed similar accumulation of Tyr15-phosphorylated Cdc2 and similar reduction of total Cdc2 levels in parental HCT116 cells. These data suggest that elevated levels of Hsp72 in cancer cells keep in check both p53- and Cdc2-mediated senescence pathways, thus promoting cell proliferation. Based on these data, we conclude that increased levels of Hsp72 in many tumor cell lines serve to protect them from default senescence. Moreover, Hsp72 depletion promotes senescence of cancer cells via two independent pathways (i.e., either through activation of p53/p21 or stimulation of inhibitory phosphorylation and down-regulation of Cdc2).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In many cancers, suppression of the p53 pathway is achieved by mutations and deletions of p53 or Arf (34), or by amplification or activation of Mdm2 or MdmX (35). Our novel findings suggest that suppression of p53 system in cancer cells may be alternatively achieved by overexpressing Hsp72. Activation of p53 is known to either stimulate apoptosis through up-regulation of proapoptotic proteins like Bax or Puma, or it can lead to growth arrest and cell senescence via up-regulation of the Cdk inhibitor p21 (36). Interestingly, in all tested cell lines in which Hsp72 depletion led to activation of the p53 pathway, we observed a robust up-regulation of p21 and senescence but not apoptosis. It is noteworthy that in these lines, including HeLa, HCT116, and MCF-7, in contrast to lymphoid cells, activation of p53 by DNA-damaging agents at low doses also usually triggers senescence rather than apoptosis. Dominance of senescence over apoptosis in our experiments with Hsp72 depletion could also result from a prolonged activation of p53 at relatively modest levels.
Depletion of Hsp72 did not lead to any detectable reduction in the capability of the cells to handle abnormal proteins (Supplementary Fig. S2). This is not very surprising because HCT116 cells have relatively high levels of a homologous housekeeping member of the family Hsc73, as well as other chaperones (e.g., Hsp27; Fig. 1). Therefore, it seems that Hsp72 specifically control p53- and cdc2-senescent pathways. Interestingly, this activity of Hsp72 seems to be related to cancer cells because we did not observe either activation of p53 or senescence in a nontransformed breast epithelial cell line, MCF-10, which has a normal p53 pathway. These data is consistent with the idea that overexpression of Hsp72 represents a response of cancer cells to activation of p53.
p53 activation commonly follows DNA-damaging stresses. Therefore, we tested whether Hsp72 depletion activates p53 via stimulation of DNA damage signaling pathways. However, we neither observed any DNA damage (Fig. 4A) nor detected activation of any components of the DNA-damage signaling pathways (Fig. 4B) upon Hsp 72 depletion. Nevertheless, Hsp72 depletion led to accumulation of Ser15 phosphorylated form of p53 (Fig. 3C). These data suggest that background levels of ATR kinase may be important for Ser15 phosphorylation.
Hsp72 depletion led to increased levels of Hdm2, which likely resulted from increased p53 activity (Fig. 3C). Despite higher Hdm2 levels, p53 was stabilized (Fig. 5A), suggesting that Hdm2 activity was low. Importantly, despite increased expression of Hdm2, its stability was markedly reduced in Hsp72-depleted cells compared with control cells (Fig. 5D and E). Most likely, decreased stability of Hdm2 resulted from its self-ubiquitination, which, in turn, could interfere with the ability of Hdm2 to bind and ubiquitinate p53 (30). Because p53 and Hdm2 represent components of a negative feedback regulatory loop, it was difficult to identify primary events in p53 regulation. However, the effect of Hsp72 on Hdm2 degradation was seen even in the p53 knockout cells (see Fig. 5E). Therefore, the presented data strongly suggest that destabilization and subsequent inactivation of Hdm2 upon Hsp72 depletion is independent of p53 and, thus, can account for p53 activation. It is not clear, however, whether effects of Hsp72 depletion on Hdm2 stability are direct or mediated via other components, such as MdmX or c-Abl.
In addition to p53-mediated senescence, we showed that Hsp72 controls a distinct pathway that could be activated in both p53-positive and p53-knockout cells. The p53-independent pathway of senescence triggered by Hsp72 depletion leads to G2-M phase cell cycle arrest and is related to inactivation of the cell cycle kinase Cdc2 (Fig. 6). In fact, Hsp72 depletion leads to a 2-fold reduction in total Cdc2 levels and simultaneously results in accumulation of inactive, Tyr15 phosphorylated form of Cdc2. The common mechanism of controlling Cdc2 activity is through regulating the rate of its dephosphorylation by a family of Cdc25 phosphatases (33). However, upon depletion of Hsp72, we did not observe decrease in the rates of Cdc2 Tyr15 dephosphorylation (see Fig. 6E). Therefore, other mechanisms of Cdc2 regulation, probably involving Wee or Myt kinases, have to be involved.
Understanding mechanisms of senescence caused by Hsp72 depletion may have a practical outcome, because development of Hsp72 inhibitors may become a novel, interesting approach toward sensitization of cancer cells to chemotherapy and overcoming resistance.
|
Acknowledgments |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. B. Vogelstein for the HCT116 knockout cell lines, Dr. M. Jaattela for the Hsp70-2 antibody, and C. O'Callahan for helpful discussion.
|
Footnotes |
|---|
J.A. Yaglom and V.L. Gabai contributed equally to this work.
Received 10/16/06. Revised 12/ 7/06. Accepted 12/14/06.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. Korhonen, S. Sjovall, J. Viitanen, E. Ketoja, A. Makarevich, and J. Peippo Viability of bovine embryos following exposure to the green filtered or wider bandwidth light during in vitro embryo production Hum. Reprod., February 1, 2009; 24(2): 308 - 314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. L. Gabai, J. A. Yaglom, T. Waldman, and M. Y. Sherman Heat Shock Protein Hsp72 Controls Oncogene-Induced Senescence Pathways in Cancer Cells Mol. Cell. Biol., January 15, 2009; 29(2): 559 - 569. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. L. Gabai, C. O'Callaghan-Sunol, L. Meng, M. Y. Sherman, and J. Yaglom Triggering Senescence Programs Suppresses Chk1 Kinase and Sensitizes Cells To Genotoxic Stresses Cancer Res., March 15, 2008; 68(6): 1834 - 1842. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Ghosh, T. Dohi, B. H. Kang, and D. C. Altieri Hsp60 Regulation of Tumor Cell Apoptosis J. Biol. Chem., February 22, 2008; 283(8): 5188 - 5194. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. O'Callaghan-Sunol, V. L. Gabai, and M. Y. Sherman Hsp27 Modulates p53 Signaling and Suppresses Cellular Senescence Cancer Res., December 15, 2007; 67(24): 11779 - 11788. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
