Targeting survivin as a potential new treatment for chondrosarcoma of bone

Abstract

Chondrosarcomas are malignant cartilage-forming bone tumors, which are intrinsically resistant to chemo- and radiotherapy, leaving surgical removal as the only curative treatment option. Therefore, our aim was to identify genes involved in chondrosarcoma cell survival that could serve as a target for therapy. siRNA screening for 51 apoptosis-related genes in JJ012 chondrosarcoma cells identified BIRC5, encoding survivin, as essential for chondrosarcoma survival. Using immunohistochemistry, nuclear as well as cytoplasmic survivin expression was analyzed in 207 chondrosarcomas of different subtypes. Nuclear survivin has been implicated in cell-cycle regulation while cytoplasmic localization is important for its anti-apoptotic function. RT-PCR was performed to determine expression of the most common survivin isoforms. Sensitivity to YM155, a survivin inhibitor currently in phase I/II clinical trial for other tumors, was examined in 10 chondrosarcoma cell lines using viability assay, apoptosis assay and cell-cycle analysis. Survivin expression was found in all chondrosarcoma patient samples. Higher expression of nuclear and cytoplasmic survivin was observed with increasing histological grade in central chondrosarcomas. Inhibition of survivin using YM155 showed that especially TP53 mutant cell lines were sensitive, but no caspase 3/7 or PARP cleavage was observed. Rather, YM155 treatment resulted in a block in S phase in two out of three chondrosarcoma cell lines, indicating that survivin is more involved in cell-cycle regulation than in apoptosis. Thus, survivin is important for chondrosarcoma survival and chondrosarcoma patients might benefit from survivin inhibition using YM155, for which TP53 mutational status can serve as a predictive biomarker.

Figure 1

Screening for apoptotic regulators identifies BIRC5 as an important survival gene in chondrosarcoma cell line JJ012. (a) Occupied area of cells present in the well as a percentage to control siRNA of 51 apoptosis-related genes. Mean values and range of duplicates are shown. (b) Raw data measured as Hoechst area of controls used in the siRNA screen. Dots indicate individual measurements and mean values are shown for each control. siCTR indicates control siRNA of either GFP or GAPDH. (c) Area of cells as a percentage to control siRNA of selected hits. BIRC5 shows 4/4 siRNAs that mimic the smart pool. Data represent means of duplicate values with range. Black bars represent the smart pool and gray bars represent individual siRNAs.

Figure 2

Survivin is highly expressed in high-grade chondrosarcoma. (a) Immunohistochemical analysis of survivin in conventional chondrosarcoma showing higher cytoplasmic expression in grade II and III central chondrosarcoma compared with ACTs. Also higher expression of nuclear survivin is observed in grade III chondrosarcomas compared with ACTs. Each dot represents one patient, and mean value with standard deviation is shown for each group. (b) High cytoplasmic survivin expression in a grade III chondrosarcoma. (c) Immunohistochemical analysis of nuclear and cytoplasmic survivin in dedifferentiated, clear cell and mesenchymal chondrosarcoma. Nuclear survivin is highly expressed in nuclei of the well-differentiated part of dedifferentiated chondrosarcoma and clear cell chondrosarcoma compared with the cytoplasmic part. Each dot represents one patient, and mean value with standard deviation is shown for each group. (d) High nuclear survivin expression in dedifferentiated chondrosarcoma (dedifferentiated part).

Figure 3

P53 is highly expressed in high-grade conventional chondrosarcoma and correlated with nuclear survivin in grade III chondrosarcoma. (a) Immunohistochemical analysis of p53 expression in conventional chondrosarcoma. Grade II and III chondrosarcomas show significantly higher expression compared with grade I chondrosarcoma (P<0.0001). Each dot represents one patient, and mean value with standard deviation is shown for each group. (b) High nuclear p53 expression in a high-grade chondrosarcoma. (c) Low nuclear p53 expression in a high-grade chondrosarcoma. (d) Correlation between nuclear survivin and p53 overexpression in conventional chondrosarcoma. P53 is considered as overexpressed when it reached a sum score of 4 (indicated by +). Grade I chondrosarcomas did not show p53 overexpression. In grade II chondrosarcoma, higher nuclear survivin was seen in p53 overexpressing chondrosarcomas, as well as in grade III chondrosarcoma (P<0.01) compared with low p53 expressing tumors. Each dot represents one patient, and mean value with standard deviation is shown for each group.

Figure 4

Survivin wt, 2b and Δex3 isoforms are highly correlated and show higher expression in high-grade chondrosarcoma as determined by Q-PCR analysis. (a–c) Survivin isoforms wt (a), 2b (b) and Δex3 (c) are highly expressed in high-grade (grade II and III) chondrosarcoma compared with low-grade (ACT) chondrosarcoma. No expression is found in cartilage while variable expression is seen in growth plate tissue. Each dot represents one measurement, and mean values are shown for each group.

Figure 5

Chondrosarcoma cell lines are sensitive to YM155, which is p53 dependent. (a) JJ012 cell line showing strong survivin expression. (b) Normalized RNA expression of three survivin isoforms in eight chondrosarcoma cell lines. JJ012, SW1353 and L835 show a clear reduction in survivin expression after YM155 treatment. Growth plate and cartilage are taken as a control. Bars represent mean with standard deviation. (c, d) Dose response curves for YM155 (72 h) in conventional subtypes (c) and rare chondrosarcoma subtypes (d). Error bars are shown for three experiments performed in triplicate. (e) IC₅₀s determined for TP53 wt and mutant cell lines showing a significantly increased sensitivity of TP53 mutant cell lines toward YM155. Bars represent mean values of three experiments performed in triplicate.

Figure 6

YM155 does not cause caspase 3/7 and PARP-dependent apoptosis, but cell-cycle deregulation in chondrosarcoma cell lines. (a, b) Caspase 3/7 activity (a) and viability (b) after 24 h as percentage to untreated control measured in nine chondrosarcoma cell lines by caspase-glo assay and presto blue viability assay. No caspase 3/7 activity is seen after YM155 treatment in all cell lines. Pan Caspase inhibitor z-vad was not able to rescue the YM155-dependent reduction in viability. MCS170 cells treated with doxorubicin and ABT-737 were used as a positive control. Error bars are shown for two independent experiments performed in duplicate. (c) Western blot analysis for PARP and cleaved PARP expression in four chondrosarcoma cell lines. No differences are seen between treated (+) and untreated (–) samples. Alpha tubulin was used as a loading control. Jurkat cell lysates treated with 25 μm etoposide obtained from Cell Signaling were used as a positive control. (d) FACS analysis of three chondrosarcoma cell lines treated with YM155 for 48 h. JJ012 and CH2879 show a reduction in G1 and an increase in S phase after treatment with YM155. NDCS1 is not showing a difference in cell-cycle distribution.