Hsp90 inhibition sensitizes DLBCL cells to cisplatin

Abstract

Purpose: Platinum-containing therapy is standard treatment for relapsed Diffuse Large B-Cell Lymphoma (DLBCL). However, the efficacy of treatment is limited by drug resistance leading to relapse. Cisplatin resistance has been linked to impairments of the DNA damage response, and several DNA repair proteins have been identified as clients of the molecular chaperone Hsp90. Here, we investigated the combinatory treatment of cisplatin and the Hsp90 inhibitor, 17AAG, in DLBCL cells to evaluate if inhibition of Hsp90 could sensitize DLBCL cells to cisplatin treatment.

Methods: Cell viability was assessed for cisplatin and 17AAG as monotherapies and for 25 different combinations in 7 DLBCL cell lines, where the Bliss Independence Model and the Combination Index were applied to assess their interaction. Induction of apoptosis and DNA damage response were evaluated by measuring Annexin V and γH2AX levels after 48 h of exposure.

Results: 17AAG synergized with cisplatin in DLBCL cells as detected in both interaction assessment models, resulting in a lower viability after 48 h for the combination-treated cells compared to both vehicle and single drug-treated cells. The combination also induced a stronger apoptotic response and an increase in DNA damage in 17AAG, cisplatin- and combination-treated cells compared to vehicle-treated cells, with the effect of the combination generally being higher than compared to both single drugs.

Conclusion: This study demonstrates that 17AAG sensitizes DLBCL cells to cisplatin treatment. This effect is correlated with increased apoptotic and DNA damage response, potentially mediated by downregulation of Hsp90 clients in DNA repair pathways. Thus, cisplatin resistance could plausibly be overcome by combining the treatment with an Hsp90 inhibiting drug.

Keywords: 17AAG; Cisplatin; DNA repair; Diffuse large B-cell lymphoma; Drug combination; Hsp90.

Fig. 1

Single drug screening in 7 DLBCL cell lines. A Dose–response curves for cisplatin and 17AAG in the 7 DLBCL cell lines DB, HBL-1, NU-DHL-1, SU-DHL-5, RIVA, OCI-Ly7 and SU-DHL-4. Response was calculated as a viability ratio to vehicle-treated controls, where controls are set to 1. B Overview of Area Under Dose–Response Curve (AUC) across the 7 cell lines. Cell lines sorted top to bottom from highest AUC, corresponding to lowest response, to lowest AUC, corresponding to highest response

Fig. 2

Drug combination screening in DLBCL cell lines. The drug combination cisplatin + 17AAG was tested in a combination matrix consisting of 5 doses of each drug, resulting in a total of 25 different combinations. A Left panel, dose–response curve for cisplatin at different set doses of 17AAG in RIVA. Right panel, combination response compared to both single drugs for the concentration resulting in the highest Bliss score (also marked with a gray arrow in the left panel). B Left panel, dose–response curve for cisplatin at different set doses of 17AAG in OCI-Ly7. Right panel, combination response compared to both single drugs for the concentration resulting in the highest Bliss score (also marked with a gray arrow in the left panel). C Left panel, dose–response curve for cisplatin at different set doses of 17AAG in DB. Right panel, combination response compared to both single drugs for the concentration resulting in the highest Bliss score (also marked with a gray arrow in the left panel). For all graphs, response was calculated as a viability ratio to vehicle-treated controls, where controls are set to 1

Fig. 3

Comparisons of shift in AUC between cisplatin and combination. A Table with all 7 DLBCL cell lines showing doses for the combination resulting in the highest Bliss score, AUC for cisplatin and combination and the corresponding fold change, i.e., AUC_combination/AUC_cisplatin. A low fold change equals a stronger response. B Visualization of AUC shift using results for RIVA

Fig. 4

Combination index for combination in DLBCL cell lines. A combination index, an alternative estimation of combination effect, was calculated from the drug combination screen. On the top, a table with Combination index (CI) for RIVA, OCI-Ly7 and DB can be seen. The three columns on the left show the single doses in five different combinations, and the three columns to the right show the CI for each cell line. Visualization of the CI in the three DLBCL cell lines is shown on the bottom. A CI > 1 is an antagonistic interaction, a CI = 0 equals no, or additive, drug interaction, and a CI < 1 equals a synergistic drug interaction

Fig. 5

Single drug and combination effect on apoptosis in DLBCL cell lines. Annexin V expression was used to assess the apoptotic response in RIVA, OCI-Ly7 and DB after treatment with single drugs and drug combination for 48 h. A Annexin V expression after flow analysis in RIVA. Doses used were 1.7 µg/mL for cisplatin and 0.68 µg/mL for 17AAG. B Annexin expression after flow analysis in OCI-Ly7. Doses used were 1.7 µg/mL for cisplatin and 0.68 µg/mL for 17AAG. C Annexin expression after flow analysis in DB. Doses used were 1.7 µg/mL for cisplatin and 0.17 µg/mL for 17AAG. All treated cells had values significantly higher than the control cells, except for the 17AAG-treated OCI-Ly7 cells. Response was calculated as a Annexin V expression ratio to vehicle-treated controls, where controls are set to 1. The percentage Annexin V-positive cells were identified using gates from positive controls for each cell line. Student’s Independent t-test was used for statistical analysis, *p value < 0.05, **p value < 0.01, ***p value < 0.001, ****p value < 0.0001

Fig. 6

DNA damage response after treatment measured by anti-γ-H2AX staining. Three DLBCL cell lines were treated with vehicle, single drugs or combination for 48 h. A + B Results for RIVA. Figure 6A shows histograms for γH2AX for all treatments, where the gate for γH2AX positive (γH2AX +) cells were defined using unstained samples. Percentages for γH2AX + cells are calculated from triplicates and written within the figure. Figure 6B shows γH2AX mean intensity for cisplatin- and combination-treated cells calculated and shown as a ratio to vehicle-treated cells, where controls are set to 1. Same panel set-up is shown for OCI-Ly7 (C + D) and DB (E + F). Histograms show one representative sample, and mean intensities (± SD) are calculated from triplicates from one representative experiment from each cell line (n = 2). Student’s Independent t-test was used for statistical analysis, *p value < 0.05, **p value < 0.01, ***p value < 0.001, ****p value < 0.0001