ISG15 is downregulated by KLF12 and implicated in maintenance of cancer stem cell-like features in cisplatin-resistant ovarian cancer

Abstract

Drug resistance is often developed during clinical chemotherapy of ovarian cancers. The ubiquitin-like protein interferon-stimulated gene 15 (ISG15) is possibly dependent on tumour context to promote or suppress progression of various tumours. The ubiquitin-like protein interferon-stimulated gene 15 (ISG15) was decreased in cisplatin-resistant ovarian cancer cells. The current study identified that both ectopic wild type and nonISGylatable mutant ISG15 expression inhibited CSC-like phenotypes of cisplatin-resistant ovarian cancer cells. Moreover, ectopic ISG15 expression suppressed tumour formation in nude mice. In addition, ISG15 downregulation promoted CSC-like features of cisplatin-sensitive ovarian cancer cells. Furthermore, low ISG15 expression was associated with poor prognosis in patients with ovarian cancer. Transcriptional repressor Krüppel-like factor 12 (KLF12) downregulated ISG15 in cisplatin-resistant cells. Our data indicated that downregulating ISG15 expression, via weakening effect of KLF12, might be considered as new therapeutic strategy to inhibit CSC phenotypes in the treatment of cisplatin-resistant ovarian cancer.

Keywords: ISG15; KLF12; cancer stem cell; cisplatin resistance; ovarian cancer.

FIGURE 1

Downregulated ISG15 expression promotes cancer stem cell like features of cisplatin‐resistant ovarian cancer cells. A, Total proteins were isolated from the indicated cells, ISG15 expression was evaluated using Western blot. B, Cells were cultured under serum‐free media and supernatant were collected after 3 d' culture, extracellular release of ISG15 was analysed using dot blot. C, Cells were infected with lentivirus containing wild type (WT) or nonconjugatable mutant (G156/157A) ISG15, ISG15 expression were confirmed by Western blot. D, The indicated cell were treated with 10 μg/mL of cisplatin for 24 h, and cell viability was assessed using CCK8 assay. E, The indicated cell were treated with 10 μg/mL of cisplatin for 24 h, cell apoptosis was analysed. F‐K, capacities of colony formation (F‐G), migration (H), invasion (I) and spheroid formation (J‐K) were evaluated in the indicated cells. Representative images of colony formation (F) and spheroid formation (J) were presented. An asterisk (*) represents significant difference with P < .05. Error bars are indicative of means ± SD. n.s., not significant

FIGURE 2

Loss of ISG15 rescues cancer stem cell like phenotypes of cisplatin‐sensitive ovarian cancer cells. A, SKOV3 and A2780 cells were infected with lentivirus containing scramble or shISG15, downregulation of ISG15 was confirmed by Western blot. B, The indicated cell viability was assessed using CCK8 assay, after exposure to 10 μg/mL of cisplatin for 24 h. C, The indicated cell were treated with 10 μg/mL of cisplatin for 24 h, cell apoptosis was analysed. D‐I, capacities of colony formation (D‐E), migration (F), invasion (G) and spheroid formation (H‐I) were analysed in the indicated cells followed by the infection with lentivirus containing scramble or shISG15. Representative images of colony formation (D) and spheroid formation (I) were presented. An asterisk (*) represents significant difference with P < .05. Error bars are indicative of means ± SD. n.s., not significant

FIGURE 3

Ectopic ISG15 expression suppresses tumour formation in nude mice and low ISG15 expression is associated with poor prognosis. A, Serially diluted cells were inoculated intracutaneously into nude mice. Experimental mice were killed and tumours were excised on day 28 (n = 3 mice/group). B, ELDA was used to predict the frequency of cancer stem cell, and the limiting dilution model was plotted as log‐fraction. The log‐active cell fraction was presented as the dotted lines and the 95% confidence interval was also given. *P < .05. NS not significant. C, Immunohistochemistry staining with ISG15 using epithelial ovarian cancer tissue microarray. D, Representative immunohistochemistry staining indicated in (C). E, ISG15 expression was categorized as high and low expression, and Kaplan‐Meier plot was used to analyse the overall survival of patients with ovarian cancer. Log‐rank test was used to determine P‐value

FIGURE 4

Transcriptional regulation of ISG15 and stability of ISG15 mRNA. A, Total RNA was isolated from the indicated cells, and ISG15 mRNA expression was analysed by real‐time RT‐PCR. B, Newly synthesized ISG15 mRNA was captured using a ClickiT in the indicated cells and then measured using real‐time RT‐PCR. C, The luciferase reporter vector bearing SV40, null (Empty) or ISG15 promoter (pISG15) was constructed. SKOV3 and SKOV3/DDP cells were cotransfected with the indicated luciferase reporter vector and Renilla reporter vector. The activity of luciferase and Renilla was analysed 2 d later. Renilla activity was used to normalize the luciferase activity. D‐E, Cisplatin‐sensitive and cisplatin‐resistant SKOV3 (D) or A2780 cells (E) were exposed to actinomycin D for the indicated time, and ISG15 mRNA was measured using real‐time RT‐PCR ISG15 mRNA levels were normalized to 18S rRNA and plotted as a percentage from three experiments repeated independently. An asterisk (*) represents significant difference with P < .05. Error bars are indicative of means ± SD. n.s., not significant

FIGURE 5

ISG15 transcription is regulated by KLF12 via the CACCC element located in −1187/−1013 and −672/−503 fragment. A, The luciferase reporter vector bearing p‐1767/+53, p‐1237/+53, p‐896/+53 or p‐411/+53 segment of ISG15 promoter was constructed. The indicated luciferase reporter vector and Renilla reporter vector was used to transfect SKOV3 or SKOV3/DDP cells. Luciferase activity was evaluated 2 d later. Renilla activity was used to normalize luciferase activity. B, Two potential binding motifs in ISG15 promoter for KLF12 (CACCC elements) were located at −1130/−1126 and −598/−594 sequence. Luciferase reporter bearing ISG15 promoter with −1130/−1126 deletion, −598/−594 deletion or both of them deletion was constructed. SKOV3 or SKOV3/DDP cells were cotransfected with the indicated luciferase reporter vector and Renilla reporter vector. Luciferase activity was evaluated 2 d later. Renilla activity was used to normalize luciferase activity. C, Total proteins were derived from paired cisplatin‐sensitive and cisplatin‐resistant SKOV3 or A2780 cells, and KLF12 and ISG15 expression was assessed using Western blot. D‐E, ChIP assay of KLF12 recruitment to the indicated DNA fragment of ISG15 promoter in cisplatin‐sensitive and cisplatin‐resistant SKOV3 (D) and A2780 (E) cells. F, Cisplatin‐sensitive and cisplatin‐resistant SKOV3 or A2780 cells were infected with lentivirus containing scramble or shKLF12, ISG15 and KLF12 expression was analysed using Western blot. G‐H, ChIP assay of KLF12 recruitment to the −1187/−1013 fragment of ISG15 promoter in cisplatin‐sensitive and cisplatin‐resistant SKOV3 or A2780 cells transfected with scramble or shKLF12. I‐J, ChIP assay of KLF12 recruitment to the −672/−503 fragment of ISG15 promoter in cisplatin‐sensitive and cisplatin‐resistant SKOV3 or A2780 cells transfected with scramble or shKLF12. An asterisk (*) represents significant difference with P < .05. Error bars are indicative of means ± SD. n.s., not significant