Inhibition of Heat Shock Protein 90 Function by 17-Allylamino-17-Demethoxy-Geldanamycin in Hodgkin's Lymphoma Cells Down-Regulates Akt Kinase, Dephosphorylates Extracellular Signal–Regulated Kinase, and Induces Cell Cycle Arrest and Cell Death

Materials and Methods

Cell lines and cell culture. The human Hodgkin's lymphoma–derived cell lines HD-MyZ, HD-LM-2, L-428, and KM-H2 were obtained from the German Collection of Microorganisms and Cell Cultures, Department of Human and Animal Cell Cultures (Brunswick, Germany). The phenotypes and genotypes of these cell lines have been previously published (22). The human T-cell lymphoblastic leukemia cell line Jurkat was obtained from the American Type Culture Collection (Manassas, VA). Unless otherwise specified, all cell lines were cultured in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies, Inc., Gaithersburg, MD), 1% l-glutamine, and penicillin/streptomycin in a humid environment of 5% CO₂ at 37°C.

Tumor samples. The study group included 43 patients with Hodgkin's lymphoma (37 classic nodular sclerosis or mixed cellularity and 6 lymphocyte predominant Hodgkin's lymphoma). The diagnosis of Hodgkin's lymphoma was based on currently used histologic and immunohistochemical criteria as defined in the WHO classification (23). The tumor tissue specimens used in this study were obtained from previously untreated patients.

Reagents, antibodies, and recombinant proteins. 17-AAG and the pan-caspase inhibitor Z-VAD-FMK were obtained from Alexis Corporation (San Diego, CA). Both reagents were dissolved in DMSO to reach a final concentration of 0.1% in cell culture. Antibodies specific to HSP90 were obtained from Alexis; antibodies to HSP70, Akt, extracellular signal-regulated kinase (ERK) 1/2, cyclin-dependent kinases 4 and 6 (cdk4 and cdk6), cyclin B1, procaspase 8, caspase 9, active caspase 3, cleaved poly(ADP-ribose) polymerase, and p53 were obtained from Cell Signaling Technology (Beverly, MA); antibodies specific to murine double minute-2 (MDM2), p21, pERK, cyclin D1, Mcl-1, cellular inhibitor of apoptosis protein 2 (CIAP2), cellular FLICE-like inhibitory protein (cFLIP), X-linked inhibitor of apoptosis protein (XIAP), and apoptosis-inducing factor (AIF) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA); antibodies to HSP60 were from Stressgen (Victoria, British Columbia, Canada); antibodies to Bcl-2 were from DAKO (Carpinteria, CA); and antibodies to β-actin were from Sigma-Aldrich Co. (St. Louis, MO). Doxorubicin was a gift from The University of Texas M.D. Anderson Cancer Center pharmacy. The two fully human agonistic antibodies to the tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) death receptor R1 (HGS-ETR1) and TRAIL-R2 (HGS-ETR2) were provided by Human Genome Sciences, Inc. (Rockville, MD).

In vitro proliferation assay. Cells were cultured in 6-, 12-, and 24-well plates at a concentration of 0.5 × 10⁶/mL. Cell viability was assessed with the nonradioactive cell proliferation 3-(4,5-dimethyl-thiazol-2yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assay using CellTiter96 AQ_ueousOne Solution Reagent (Promega, Madison, WI) as previously published (24). Briefly, 80 μL of cell suspension and 20 μL of CellTiter96 AQ_ueousOne Solution Reagent were incubated in 96-well plates for 1 hour at 37°C, 5% CO₂, and formazan absorbance was measured at 490 nm on a μQuant plate reader equipped with KC4 software (Biotek Instruments, Winooski, VT). Each measurement was done in triplicate and the mean value was calculated.

Combination index calculation. The effectiveness of the agents used in this study and their combination were analyzed by the Calcusyn Software (Biososoft, Ferguson, MO). The combination index and the isobologram plot were calculated according to the Chou-Talalay method (25–27). A combination index value of 1 indicates an additive effect between two drugs. Combination index values <1 indicate synergy, and the lower the value, the stronger the synergy. In contrast, combination index values >1 indicate antagonism.

Flow cytometry. Apoptosis was determined using Annexin V-FLUOS staining (Roche Molecular Biochemicals, Indianapolis, IN) according to the instructions of the manufacturer and as previously published (24). Cell cycle fractions were determined by propidium iodide nuclear staining (20). Briefly, cells were harvested, washed in PBS, fixed with 70% ethanol, and incubated with propidium iodide for 30 minutes at 37°C. Data were collected and analyzed with fluorescence-activated cell sorting analysis (FACS) on a Becton Dickinson FACSCalibur flow cytometer using CellQuestPro software (BD Biosciences, San Jose, CA). Cell cycle analysis was done by applying diploid cell cycle model using the ModFit LT (Verity Software House, Inc., Topsham, ME).

Western blot analysis. Whole cellular protein was extracted by incubating cells in lysis buffer (Cell Signaling Technology) for 30 minutes on ice. Subcellular fractionation was done by incubating cells in a buffer containing 100 mmol/L KCl, 2.5 mmol/L MgCl₂, 250 mmol/L sucrose, 20 mmol/L HEPES/KOH, 20 μmol/L cytochalasin B, and 0.025% digitonin (all from Sigma-Aldrich), and centrifugation at 800 × g to pellet nuclei and at 8,000 × g to pellet mitochondria, and subsequent incubation in lysis buffer (Cell Signaling Technology) for 30 minutes on ice. Lysates were then centrifuged to remove cellular debris. The protein in the resulting supernatant was quantified by the bicinchonic acid method (Pierce Chemical Co., Rockford, IL) according to the instructions of the manufacturer, diluted 1:2 in protein SDS loading buffer (Cell Signaling Technology), and heated to 95°C for 5 minutes. A total of 30 μg of protein were loaded onto 12% Tris-HCl SDS-polyacrylamide electrophoresis Ready Gels (Bio-Rad, Hercules, CA), transferred to a nitrocellulose transfer membrane (Pierce Chemical), and detected by using SuperSignal WestDura Extended Duration Substrate (Pierce Chemical) as previously published (19, 20).

Tissue microarray and immunohistochemical methods. A manual tissue microarrayer (Beecher Instruments, Silver Springs, MD) was used to construct a tissue microarray as previously described (28). The tissue microarray used included quadruplet tumor cores from 43 lymph nodes involved by Hodgkin's lymphoma (37 classic and 6 lymphocyte predominant) all from previously untreated patients. Briefly, tissue microarray sections (3 or 4 mm thick) were deparaffinized in xylene and rehydrated in a graded series of ethanols. Heat-induced epitope retrieval was done. The slides were incubated with a monoclonal antibody specific for HSP90 (NCL-HSP90, Novocastra, Newcastle upon Tyne, United Kingdom) at a dilution of 1:50 overnight at 4°C. The immunoreaction was detected by using the LSAB+ kit (DakoCytomation, Carpinteria, CA), which contains a secondary biotinylated antibody and streptavidin/horseradish peroxidase complex. We used 3,3′ diaminobenzidine/H₂O₂ (DakoCytomation) as the chromogen and hematoxylin as the counterstain. Any cytoplasmic staining of tumor cells for HSP90 was considered positive irrespective of intensity. Rarely, reactive lymphocytes were also positive for HSP90. A 10% cutoff for the percentage of HSP90-positive tumor cells was chosen to define HSP90 expression.

Results

Expression of HSP90 in primary and cultured Hodgkin's lymphoma cells. We first determined the levels of HSP90 in Hodgkin's lymphoma cell lines. For this experiment, whole-cell lysates from the HD-MyZ, HD-LM2, L-428, and KM-H2 cell lines were used. The T-cell lymphoblastic leukemia cell line Jurkat, as well as two normal peripheral blood lymphocyte samples, were used as controls. HSP90 was abundantly expressed in all cell lines tested (Fig. 1A). These cell lines also expressed abundant levels of HSP70. In contrast, normal peripheral blood lymphocytes only weakly expressed HSP90 and HSP70.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Expression of HSP90 and HSP70 in Hodgkin's lymphoma cell lines shown by Western blot analysis. A, expression of HSP90 and HSP70 in 4 Hodgkin's lymphoma cell lines (HD-MyZ, HD-LM2, L-428, and KM-H2), in Jurkat cells, as well as in two primary samples of peripheral lymphocytes from normal donors (NPL), determined by Western blot. B, three representative cases of classic Hodgkin's lymphoma sections stained for HSP90. HSP90 is strongly expressed by HRS cells with relatively low levels by the reactive cells. HSP90 localization is predominantly cytoplasmic.

To explore the expression of HSP90 in primary HRS cells in tumor specimens, we evaluated 43 lymph node biopsy specimens obtained from previously untreated patients with Hodgkin's lymphoma. HRS cells in 35 of 37 (95%) cases of classic Hodgkin's lymphoma and all 6 cases of lymphocyte predominant Hodgkin's lymphoma (100%) overexpressed HSP90. Virtually all HRS cells in positive cases overexpressed HSP90 and the staining pattern was predominantly cytoplasmic with some weak nuclear staining (Fig. 1B).

17-AAG induces antiproliferative effect in Hodgkin's lymphoma cell lines. 17-AAG is a small-molecule inhibitor of HSP90 that has antiproliferative activity in a variety of cancer cells and recently entered clinical trials in patients with solid tumors and lymphoma (29). To investigate the biological effect of HSP90 inhibition in Hodgkin's lymphoma, we incubated Hodgkin's lymphoma cell lines with increasing doses of 17-AAG (0-5 μmol/L) or DMSO (0.1% final concentration) and cell viability was determined by the MTS proliferation assay. 17-AAG inhibited cell viability in a dose-dependent manner in all four Hodgkin's lymphoma cell lines. When Hodgkin's lymphoma cells were incubated with 5 μmol/L of 17-AAG, 30.38% of HD-LM2 and 51.44% of L-428 remained viable compared with DMSO control cells (Fig. 2A). The IC₅₀ of 17-AAG was determined at 2.32 μmol/L for HD-LM2 and at 5.75 μmol/L for L-428 at 48 hours (data not shown). In the remaining two cell lines, 5 μmol/L of 17-AAG left 55% of the HD-MyZ cells and 45% of the KM-H2 cells alive after 48 hours (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Activity of HSP90 inhibitor 17-AAG in Hodgkin's lymphoma cell lines and molecular effects. A, dose-effect curve of 17-AAG on HD-LM2 and L-428 cell lines at 48 hours. The number of viable cells was determined with the MTS assay. Points, mean of three independent experiments done in triplicate; bars, SE. B, effect of 17-AAG (5 μmol/L) on the expression of HSP90, HSP70, AKT, ERK1/2, and pERK1/2. C, effect of 17-AAG (5 μmol/L) on cellular FLICE-like inhibitory protein (cFLIP), Bcl-2, and inhibitor of apoptosis protein (cIAP2 and XIAP). Cells were incubated for 24, 48, and 72 hours with DMSO or 17-AAG before they were subjected to analysis by Western blot.

Decreased cell viability was associated with up-regulated expression of HSP90 and HSP70 as early as 24 hours and lasted for at least 72 hours, indicative of stress response (Fig. 2B). Within the same time frame, the prosurvival protein kinase Akt was depleted in both cell lines and ERK phosphorylation was completely inhibited (Fig. 2B). Furthermore, 17-AAG reduced the cellular contents of the short form of cellular FLICE-like inhibitory protein but had no significant effect on Bcl-2, cellular inhibitor of apoptosis protein, or X-linked inhibitor of apoptosis protein levels (Fig. 2C).

17-AAG induces apoptosis and cell cycle arrest in Hodgkin's lymphoma cell lines. HSP90 is known to chaperone a variety of proteins that regulate cell cycle and apoptosis (5, 30). Thus, we determined whether the antiproliferative activity of 17-AAG was due to apoptosis, cell cycle arrest, or both. For these experiments, HD-LM2 and L-428 cells were incubated with 5 μmol/L 17-AAG or DMSO for as long as 72 hours. Apoptosis was determined by the Annexin V binding assay with a propidium iodide counterstain and fluorescence-activated cell sorting analysis (Fig. 3A). In HD-LM2 cells, 17-AAG induced 30% cell death (cells positive for Annexin V, propidium iodide, or both) after 24 hours compared with 7% in control DMSO. After 72 hours, 65% of the cells treated with 17-AAG died compared with 10% in control DMSO. After 72 hours, 17-AAG rendered 39% cell death in L-428 cells compared with 14% in control DMSO (Fig. 3A).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Inhibition of HSP90 induces apoptosis in Hodgkin's lymphoma cell lines. A, HD-LM2 and L-428 cell lines were incubated with and without 5 μmol/L 17-AAG for up to 72 hours. Apoptosis was determined with Annexin V-FLUOS/propidium iodide staining and fluorescence-activated cell sorting analysis. B, treatment of HD-LM2 and L-428 cell lines with 5 μmol/L 17-AAG for up to 72 hours showed prominent caspase 9, caspase 8, caspase 3, and poly(ADP-ribose) polymerase (PARP) cleavage, indicating the triggering of the mitochondrial/intrinsic caspase pathway. C, HD-LM2 and L-428 cells were incubated with DMSO, Z-VAD-FMK (20 μmol/L), 17-AAG, and the combination of Z-VAD-FMK and 17-AAG. The number of viable cells was determined with the MTS assay ± SE. D, L-428 cells were incubated with DMSO (0.1%) or 17-AAG (5 μmol/L) for 48 hours and subcellular fractions were isolated. Apoptosis-inducing factor (AIF) shows clear translocation from mitochondria to the nucleus. HSP60 was used as mitochondrial marker and Rb for nucleus.

Cell death was associated with delayed caspase-8 and caspase-3 activation, caspase-9 depletion, and poly(ADP-ribose) polymerase cleavage, which were more prominent in the HD-LM2 cell line (Fig. 3B). These events were observed after 48 hours of incubation (Fig. 3B). To confirm the involvement of the caspase pathway in 17-AAG-induced cell death, we treated Hodgkin's lymphoma cells with 17-AAG (5 μmol/L) with or without the pan-caspase inhibitor Z-VAD-FMK (20 μmol/L). Z-VAD-FMK partially reversed 17-AAG-induced cell death in the HD-LM2 cell line and to a lesser extent in the L-428 cell line (Fig. 3C). These data suggest that caspase activation is not the only mechanism for 17-AAG-induced cell death in Hodgkin's lymphoma cell lines.

Next, we examined whether 17-AAG can induce cell death by caspase-independent mechanisms (31). For this experiment, L-428 cells were incubated with DMSO (0.1%) or 17-AAG (5 μmol/L) for 48 hours and subcellular fractionation was done to discriminate mitochondria from nuclei. The effect of 17-AAG on subcellualr localization of apoptosis-inducing factor was determined by Western blot. As shown in Fig. 3D, 17-AAG induced translocation of apoptosis-inducing factor from the mitochondria to the nucleus, consistent with induction of caspase-independent apoptosis (31–33).

Induction of apoptosis was preceded by cell cycle arrest. Treatment of HD-LM2 cells with 17-AAG predominantly induced G₀-G₁ cell cycle arrest (Fig. 4A), which was associated with down-regulation of several cell cycle regulatory proteins, including cdk4, cdk6, polo-like kinase 1 (Plk1), cyclin B1, and, to a lesser extent, cyclin D1 (Fig. 4B). In contrast, treatment of L-428 cells with 17-AAG was associated with cell cycle arrest in the G₂-M phase with no significant effect on of cyclin D1, Plk1, or cyclin B1 levels. However, 17-AAG down-regulated cdk4 and cdk6 proteins (Fig. 4B). In both cell lines, 17-AAG reduced MDM2 levels but had a different effect on p53 levels; it reduced p53 levels in HD-LM2 cells but stabilized p53 levels in L-428 cells for up to 72 hours (Fig. 4B). Finally, 17-AAG had no significant effect on p27 level in both cell lines (Fig. 4B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

17-AAG induces cell cycle arrest in Hodgkin's lymphoma cell lines. HD-LM2 and L-428 cell lines were incubated with and without 5 μmol/L 17-AAG for up to 72 hours. A, cell cycle arrest at G₀-G₁ phase for HD-LM2 and at G₂-M phase for L-428 cells was determined for 24, 48, and 72 hours with propidium iodide staining and fluorescence-activated cell sorting analysis. B, effect of 17-AAG (5 μmol/L) on cell cycle regulatory proteins.

17-AAG enhances the effect of chemotherapy and anti-TRAIL death receptor monoclonal antibodies. Several reports showed the role of Akt activation in mediating resistance to chemotherapy and to TRAIL (34–37). Because 17-AAG depleted Akt in Hodgkin's lymphoma cells, we investigated whether 17-AAG enhances the effect of doxorubicin chemotherapy and agonistic antibodies to the TRAIL death receptors. HD-LM2 cells were incubated for 48 hours with increasing doses of doxorubicin (0.02-0.1 μg/mL), 17-AAG (1-5 μmol/L), or both at a constant ratio. Cell viability was determined by the MTS assay. In all cases, the combination of doxorubicin and 17-AAG was more effective than either agent alone (Fig. 5A). The combination index is shown in Table 1 and was consistently <1, indicating synergy (Table 1).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Inhibition of HSP90 enhances the effect of doxorubicin and anti-TRAIL monoclonal antibodies HGS-ETR1 and HGS-ETR2. HD-LM2 cells were incubated with 2.5 μmol/L 17-AAG and doxorubicin (0.05 μmol/L; A), HGS-ETR1 (2 μg/mL; B), and HGS-ETR2 (2 μg/mL; C), alone or in combination, for up to 72 hours. The isobologram for each combination was plotted for 48 hours at ED₅₀. The number of viable cells was determined with the MTS assay. Columns, mean of three independent experiments done in triplicate; bars, SE.

Similarly, 17-AAG showed synergy with agonistic monoclonal antibodies to TRAIL-R1 (HGS-ETR1 antibody) and TRAIL-R2 (HGS-ETR2 antibody) death receptors (24, 38). Treatment of HD-LM2 cells for 48 hours with either antibody (0.8-4 μg/mL), with 17-AAG (1-5 μmol/L), or with the combinations showed synergistic effect when both drugs were combined (Fig. 5B and C). The combination index was indicative of synergy between anti-TRAIL monoclonal antibodies and 17-AAG (Table 1).

Discussion

In the present study, we identified HSP90 as a potential target for novel therapeutic approaches in patients with Hodgkin's lymphoma. We have previously reported that activation of CD30, CD40, and RANK receptors in Hodgkin's lymphoma cell lines activates shared survival pathways, including ERK, Akt, and nuclear factor κB (18–20). In this study, we showed that several components of these signaling pathways are disrupted by 17-AAG, suggesting that targeting HSP90 may be as effective as targeting several signaling pathways. HSP90 overexpression has been described in a variety of tumor types and clinical trials using small-molecule inhibitors of HSP90 are currently being conducted for the treatment of cancer (29). In this study, we showed that HSP90 is abundantly expressed in Hodgkin's lymphoma–derived cell lines and in >95% of primary Hodgkin's lymphoma sections. Importantly, HSP90 expression was significantly higher in HRS cells compared with reactive benign cells in Hodgkin's lymphoma tissue specimens, suggesting that inhibitors of HSP90 may preferentially target tumor cells and spare the reactive benign cells.

17-AAG showed antiproliferative activity in Hodgkin's lymphoma cell lines. This antiproliferative activity is due to deregulation of several survival proteins, including Akt and ERK (19, 39, 40). 17-AAG rapidly and completely depleted Akt from the HD-LM2 and L-428 cell lines, which was recently reported to promote Hodgkin's lymphoma cell survival (39, 40). The ability of 17-AAG to dephosphorylate ERK is likely to be induced by upstream inhibition of HSP90-client proteins, such as mitogen-activated protein kinase/ERK kinase and raf (41–43). The inhibition of HSP90 function by 17-AAG in Hodgkin's lymphoma was associated with an increase in HSP90 and HSP70 levels, possibly as cellular response to stress, but these elevated levels did not prevent the antiproliferative effect of 17-AAG.

In addition to the effect of 17-AAG on prosurvival proteins Akt and ERK, 17-AAG also induced cell death through caspase-dependent and caspase-independent mechanisms. Caspase-dependent cell death does not seem to be the predominant mechanism because caspase activation did not occur until 48 hours of incubation and because cell death was marginally reversed by pan-caspase inhibition (Fig. 3B and C). Thus, caspase-independent apoptosis played an important role in 17-AAG-induced cell death and at least apoptosis-inducing factor was involved in this process (Fig. 3D; ref. 44).

HSP90 is known to chaperone a variety of proteins that regulate the cell cycle progression at multiple levels and inhibition of HSP90 has been reported to induce cell cycle arrest in the G₁ or G₂-M phases, depending on the cell type (5). Although cyclin D1 is not a client protein for HSP90, it was modestly down-regulated by 17-AAG in HD-LM2 cells, which was associated with a transient cell cycle arrest in the G₁ phase (Fig. 4). However, because Akt may regulate cyclin D1 translation, it is likely that reduction of cyclin D1 is a consequence of Akt depletion in these cells. However, although Akt was also depleted from L-428 cells, cyclin D1 remained unchanged and the cells were arrested in the G₂-M phase rather than at G₁ phase. This suggests that cyclin D1 expression is regulated by mechanisms other than Akt in L-428 cells.

The ability of HSP90 inhibitors to induce cell cycle arrest in the G₂-M phase has been recently linked to HSP90 association with Plk1 (45). Both HD-LM2 and L-428 expressed Plk1 but 17-AAG treatment decreased Plk1 level more effectively in HD-LM2 cells. Plk1 in turn activates cyclin B1, which regulates the transition of G₂-M phase to mitosis. Both Plk1 and cyclin B1 were depleted from HD-LM2 cells after 48 hours of incubation, which was associated with a modest increase in the G₂-M cell cycle fraction but with significant cell death, perhaps due to mitotic catastrophe. However, within the same time frame, p53 was also depleted in addition to MDM2, cdk4, and cdk6. Thus, because HSP90 chaperones several cell cycle regulatory proteins, including those involved in G₀-G₁ and G₂-M entry, the final predominant effect on cell cycle cannot be predicted by single molecular changes but rather by the net effect on all these proteins.

The ability of 17-AAG to disrupt several survival and resistance signaling proteins makes it a potentially appealing drug for combination therapy. In this study, we explored the potential synergy between 17-AAG and two other treatment modalities: chemotherapy and monoclonal antibody. We showed that 17-AAG synergized with doxorubicin chemotherapy. We have previously reported that the agonistic anti-TRAIL death receptor monoclonal antibodies HGS-ETR1 and HGS-ETR2 can induce apoptosis in a panel of lymphoid malignancies, including Hodgkin's lymphoma (24). Here we show that 17-AAG synergized with agonistic monoclonal antibodies to TRAIL death receptors R1 and R2, perhaps by disrupting several mechanisms that are known to mediate resistance to the TRAIL death pathway, including cFLIP and Akt (46–49). Because these antibodies are currently being evaluated in clinical trials, a combination trial of 17-AAG with one of these antibodies may be warranted.

Our study showed that HSP70 cochaperone was also abundantly expressed by Hodgkin's lymphoma–derived cell lines. We did not evaluate HSP70 expression in primary HRS cells because currently there is no reliable antibody that can be used for immunohistochemical analysis using fixed tissue specimens. Furthermore, there are no specific small-molecule inhibitors of HSP70 that are currently available. Whether inhibiting HSP70, or the dual inhibition of HSP90 and HSP70, may be of therapeutic value in Hodgkin's lymphoma is currently unknown but would of interest to explore in the future.

Collectively, our data show that 17-AAG may be of therapeutic value for the treatment of patients with Hodgkin's lymphoma. Our group is currently investigating this possibility in a prospective phase II study using 17-AAG in patients with relapsed classic Hodgkin's lymphoma.