The expression of SIRT6 is associated with poor prognosis of osteosarcoma patients
Immunohistochemical expression of SIRT6 in human osteosarcoma is presented in Fig. 1a. The cut-off point of the immunohistochemical staining score for SIRT6 expression was determined at the point with the highest area under the curve, which was nine (Fig. 1b). The cases with immunohistochemical staining scores equal to, or greater than, nine were considered SIRT6-positive. With this cut-off point, SIRT6 positivity was significantly associated with latent distant metastasis of osteosarcoma patients (P = 0.005) (Table 1). However, there was no significant association between SIRT6 expression and tumor stage or histologic grade (Table 1).
Table 1
The clinical characteristics of 27 osteosarcomas and their correlation with the immunohistochemical expression of SIRT6
Characteristics
|
|
No.
|
SIRT6
|
|
|
|
|
Positive
|
P
|
Age, years
|
< 30
|
24
|
11 (46%)
|
0.642
|
|
≥ 30
|
13
|
7 (54%)
|
|
Sex
|
Male
|
25
|
13 (52%)
|
0.556
|
|
Female
|
12
|
5 (42%)
|
|
Tumor size
|
≤ 8 cm
|
19
|
9 (47%)
|
0.873
|
|
> 8 cm
|
18
|
9 (50%)
|
|
Stage
|
I
|
11
|
4 (36%)
|
0.331
|
|
II, III, & IV
|
26
|
14 (54%)
|
|
Histologic grade
|
1 & 2
|
10
|
3 (30%)
|
0.167
|
|
3 & 4
|
27
|
15 (56%)
|
|
Lymph node metastasis
|
Absence
|
34
|
17 (50%)
|
0.580
|
|
Presence
|
3
|
1 (33%)
|
|
Distant metastasis at diagnosis
|
Absence
|
30
|
14 (47%)
|
0.618
|
|
Presence
|
7
|
4 (57%)
|
|
Latent distant metastasis
|
Absence
|
28
|
10 (36%)
|
0.005
|
|
Presence
|
9
|
8 (89%)
|
|
In univariable analyses, age, tumor size, stage, histologic grade, lymph node metastasis, distant metastasis, and SIRT6 expression were significantly associated with OS or RFS (Table 2). Especially, SIRT6-positivity indicated a 4.012 fold greater risk of death of osteosarcoma patients (95% CI, 1.419–11.344; P = 0.009) and had a 4.151 fold greater risk of relapse of tumor or death of patients (95% CI, 1.570-10.971; P = 0.004) (Table 2). Kaplan-Meier survival analysis also indicated that SIRT6 expression is significantly associated with OS (Log-rank, P = 0.005) and RFS (Log-rank, P = 0.002) (Fig. 1c). Multivariable analyses were performed with standard variables significantly associated with OS or RFS in univariable analyses. Age, tumor size, stage, lymph node metastasis, distant metastasis, histologic grade, and SIRT6 expression were included in multivariable analyses. Multivariable analyses showed tumor size (OS; P = 0.023, RFS; P = 0.018), distant metastasis (OS; P = 0.008, RFS; P = 0.050), and SIRT6 expression (OS; P = 0.001, RFS; P < 0.001) as independent prognostic indicators of osteosarcoma patients (Table 3). SIRT6-positivity had a 6.797 fold greater risk of death of osteosarcoma patients (95% CI, 2.129–21.702) and had a 6.516 fold greater risk of relapse of tumor or death of patients (95% CI, 2.234–19.001) (Table 3).
Table 2
Univariate Cox proportional hazards regression analysis for the overall survival and relapse-free survival of osteosarcoma patients
Characteristics
|
No.
|
OS
|
|
RFS
|
|
|
|
HR (95% CI)
|
P
|
HR (95% CI)
|
P
|
Age, years, ≥ 30 (vs. <30)
|
13/37
|
2.398 (0.938–6.134)
|
0.068
|
2.925 (1.210–7.070)
|
0.017
|
Sex, male (vs. female)
|
25/37
|
1.140 (0.402–3.229)
|
0.806
|
1.513 (0.541–4.228)
|
0.430
|
Tumor size, > 8 cm (vs. ≤ 8 cm)
|
18/37
|
3.975 (1.401–11.280)
|
0.010
|
3.067 (1.204–7.813)
|
0.019
|
Stage, ≥ II (vs. I)
|
26/37
|
4.829 (1.107–21.058)
|
0.036
|
3.267 (0.953–11.206)
|
0.060
|
Histologic grade, 3 & 4 (vs. 1 & 2)
|
27/37
|
4.502 (1.031–19.666)
|
0.045
|
5.130 (1.184–22.222)
|
0.029
|
Lymph node metastasis, presence (vs. absence)
|
3/37
|
5.373 (1.026–28.150)
|
0.047
|
2.981 (0.628–14.142)
|
0.169
|
Distant metastasis, presence (vs. absence)
|
7/37
|
5.347 (1.793–15.950)
|
0.003
|
3.802 (1.295–11.163)
|
0.015
|
SIRT6, positive (vs. negative)
|
18/37
|
4.012 (1.419–11.344)
|
0.009
|
4.151 (1.570-10.971)
|
0.004
|
OS, overall survival; RFS, relapse-free survival; HR, hazard ratio; 95% CI, 95% confidence interval. |
Table 3
Multivariate survival analysis in overall osteosarcoma patients.
Characteristics
|
OS
|
|
RFS
|
|
|
HR (95% CI)
|
P
|
HR (95% CI)
|
P
|
Tumor size, > 8 cm (vs. ≤ 8 cm)
|
3.795 (1.205–11.953)
|
0.023
|
3.680 (1.249–10.842)
|
0.018
|
Distant metastasis, presence (vs. absence)
|
6.200 (1.603–23.984)
|
0.008
|
3.450 (0.997–11.934)
|
0.050
|
SIRT6, positive (vs. negative)
|
6.797 (2.129–21.702)
|
0.001
|
6.516 (2.234–19.001)
|
< 0.001
|
OS, overall survival; RFS, relapse-free survival; HR, hazard ratio; 95% CI, 95% confidence interval. The variables included in multivariate analysis were age, tumor size, stage, lymph node metastasis, distant metastasis, histologic grade, and SIRT6 expression. |
Furthermore, we performed additional survival analysis with 26 osteosarcoma patients who were treated with postoperative adjuvant chemotherapy. In this sub-group of osteosarcoma patients, SIRT6 expression was an independent indicator of poor prognosis of osteosarcoma patients with univariable and multivariable analyses. SIRT6 expression predicted a 6.988 fold greater risk of death of osteosarcoma patients (95% CI, 1.439–33.930; P = 0.016) and had a 6.930 fold greater risk of relapse of tumor or death of patients (95% CI, 1.727–27.814; P = 0.006) (Table 4). Kaplan-Meier survival curves for the expression of SIRT6 on OS (Log-rank, P = 0.009) and RFS (Log-rank, P = 0.012) are presented in Fig. 2.
Table 4
Univariate and multivariate survival analysis in 26 osteosarcoma patients who treated with adjuvant chemotherapy
Characteristics
|
No.
|
OS
|
|
RFS
|
|
|
|
|
HR (95% CI)
|
P
|
HR (95% CI)
|
P
|
Univariate analysis
|
|
|
|
|
|
|
SIRT6, positive (vs. negative)
|
15/26
|
5.855 (1.290-26.563)
|
0.022
|
4.523 (1.245–16.434)
|
0.022
|
Multivariate analysis
|
|
|
|
|
|
|
Tumor size, > 8 cm (vs. ≤ 8 cm)
|
|
|
|
3.560 (1.055–12.009)
|
0.041
|
|
Distant metastasis, presence (vs. absence)
|
|
48.868 (2.759-865.448)
|
0.008
|
10.649 (1.186–95.614)
|
0.035
|
|
SIRT6, positive (vs. negative)
|
|
6.988 (1.439–33.930)
|
0.016
|
6.930 (1.727–27.814)
|
0.006
|
OS, overall survival; RFS, relapse-free survival; HR, hazard ratio; 95% CI, 95% confidence interval. The variables included in multivariate analysis were age, tumor size, stage, lymph node metastasis, distant metastasis, histologic grade, and SIRT6 expression. |
Suppression of SIRT6 potentiates the effect of doxorubicin on the proliferation of osteosarcoma cells.
Based on the prognostic significance of SIRT6 expression in osteosarcoma patients, especially in the patients who received adjuvant chemotherapy, we evaluated the effects of SIRT6 expression on the anti-tumor activity of doxorubicin in osteosarcoma cells. As shown in Fig. 3a and 3b, the knock-down or over-expression of SIRT6 did not influence the proliferation of U2OS and KHOS/NP osteosarcoma cells is they were untreated or treated with 0.05 µM doxorubicin. However, overexpression of SIRT6 attenuated the anti-proliferative effect of doxorubicin and knock-down of SIRT6 potentiated the effect of doxorubicin at concentrations of doxorubicin at or above 0.1 µM (Fig. 3).
Suppression of SIRT6 potentiates the cytotoxic effect of doxorubicin by increasing the apoptosis of osteosarcoma cells.
Based on the cytotoxic activity of doxorubicin, we further evaluated the effects of SIRT6 expression on apoptosis of osteosarcoma cells when co-treated with 0.1 µM doxorubicin. The effect of SIRT6 expression on apoptotic signaling was minimal without the treatment of doxorubicin (Fig. 4a). However, in the cells treated with doxorubicin, the knock-down of SIRT6 increased expressions of cleaved PARP1, BAX, and cleaved caspase 3, but decreased expression of BCL2; conversely, overexpression of SIRT6 decreased the expression of cleaved PARP1, BAX and cleaved caspase 3, but increased expression of BCL2 in both U2OS and KHOS/NP cells (Fig. 4a). Flow cytometric apoptotic analysis also showed more apoptosis of KHOS/NP cells with knock-down of SIRT6 and decreased apoptosis of KHOS/NP cells with overexpression of SIRT6 in a condition of co-treatment of 0.1 µM doxorubicin (Fig. 4b).
SIRT6 influence on in vivo growth of osteosarcoma cells with treatment of doxorubicin
We further evaluated the expression of SIRT6 on in vivo tumor growth in mice treated with doxorubicin. The treatment of doxorubicin itself significantly decreased tumor growth. Furthermore, knock-down of SIRT6 significantly potentiated the inhibitory effect of doxorubicin on in vivo tumor growth (Fig. 5). Compared to the group receiving only doxorubicin treatment, tumor growth was significantly decreased with knock-down of SIRT6 (Fig. 5). However, the effect of the anti-tumor effect of doxorubicin was attenuated by overexpression of SIRT6. There was no significant difference between the EVs group and SIRT6-overexpression with the doxorubicin treated group (Fig. 5). There was no toxic effect or distant metastasis on internal organs in four experimental groups.
Blocking of the DNA damage repair pathway with an ATM inhibitor and the PARP inhibitor, olaparib, potentiates the cytotoxic effects of doxorubicin in osteosarcoma cells overexpressing SIRT6
Recently, a role of SIRT6 in the DNA damage repair pathway has been reported [1, 8, 25]. Moreover, the effects of SIRT6 on the proliferation and apoptosis of osteosarcoma cells under the treatment of doxorubicin raise the possibility that the SIRT6-mediated DNA damage repair pathway might affect the cytotoxic effect of doxorubicin. Consistently, under treatment of doxorubicin, overexpression of SIRT6 increased expression of phosphorylated ATM and phosphorylated Chk2, while knock-down of SIRT6, especially when combined with 0.1 µM doxorubicin treatment, decreased the expression of phosphorylated ATM and phosphorylated Chk2 (Fig. 6a). In addition, the expression of γH2AX increased with the knock-down of SIRT6 and decreased with overexpression of SIRT6 under treatment of 0.1 µM doxorubicin (Fig. 6a). These findings suggest that overexpression of SIRT6 induces resistance to the genotoxic agent doxorubicin by repairing damage to escape apoptosis. Therefore, we evaluated the effect of co-treatment of doxorubicin and an ATM inhibitor/PARP inhibitor in osteosarcoma cells in which the overexpression of SIRT6 had been induced. Although the treatment of the ATM inhibitor KU-55933 was not effective in preventing the proliferation of osteosarcoma cells, the treatment of olaparib or doxorubicin significantly inhibited the proliferation of U2OS and KHOS/NP osteosarcoma cells (Fig. 6b and 6c). The cytotoxic effect of doxorubicin was attenuated with the overexpression of SIRT6 in U2OS and KHOS/NP osteosarcoma cells (Fig. 6b and 6c). However, co-treatment of KU-55933 or the PARP inhibitor olaparib with doxorubicin synergistically inhibited the proliferation of SIRT6-overexpressing U2OS and KHOS/NP osteosarcoma cells (Fig. 6b and 6c).