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.

of Hsp70-interacting protein 13 and inhibiting cancer cell growth and promoting apoptosis. 14 Furthermore, free form ISG15 can also be released extracellularly and alters the tumour microenvironment via functioning as an immunomodulatory cytokine. 7,15,16 Recently, our group have also revealed ISG15 suppresses translation of ABCC2 via ISGylation of hnRNPA2B1 and enhances drug sensitivity in cisplatinresistant ovarian cancer cells. 17 Krüppel-like factors (KLFs) are homologues of the Drosophila melanogaster Krüppel protein, which manages body segmentation during Drosophila embryo developing. 18 KLFs contain evolutionarily conserved zinc (Zn)-finger domains in their C-terminal regions. KLFs and specificity proteins (SPs) share high similarity with regards to their structure and DNA-binding capacity, thereby they are commonly known as SP/KLF transcription factors. 19 To date, 18 KLFs are reported in research work. These transcription factors participate in a variety of key cellular processes in normal tissues, including proliferation, differentiation, pluripotency and homeostasis. [20][21][22] In recent studies, it is uncovered that KLFs expression and activity are altered in individual cancers, [23][24][25][26] even one certain KLF can function as tumour promoter and suppressor based on tumour types or stage. 26,27 KLF4 is extremely important for preserving CSC-like properties in breast cancer, 23 colorectal 28 and pancreatic cancer cells. 29 KLF4, especially, can induce cells with cancer stem cell properties through somatic reprogramming. 30,31 In both normal stem cells and CSCs, KLF4 can maintain telomerase activity, which reveals a role for KLF4 in long-term proliferative potential of stem cells. 32 Several other KLF family members have been involved in CSCs, but their regulatory functions and mechanisms are not clarified. For example, KLF5 encourages CSC viability in ovarian cancer, 24 and KLF9 inhibits glioblastoma-initiating stem cells. 25 The current study displayed that ISG15 evidently decreased in cisplatin-resistant cell lines, when compared with their cisplatin-sensitive partners. Ectopic expression of wild-type ISG15 elevated cellular responses to cisplatin as well as ectopically expressed both wild-type ISG15 and nonISGylatable mutant ISG15 leaded to decreasing CSC population in resistant cell lines.
Although ISG15 knock-down in parental cell lines maintained the sensitivity in cisplatin, attenuated ISG15 strengthened CSC features of the cells. In addition, it was shown in the present study that ISG15 expression was negatively regulated by KLF12.
Therefore, these studies suggested that KLF12 might function as a potential therapeutic target via regulating ISG15 expression for inhibition of CSC phenotype in the treatment of cisplatinresistant ovarian cancer.

| Cell line cultivation
SKOV3 and A2780 cells were acquired from ATCC and GenChem, respectively. The initial cells were identified using STR profiling. Initial dose-response studies of cisplatin over 72 hours showed that IC50 of SKOV3 and A2780 cells were 0.72 ± 0.13 and 1.21 ± 0.23 μg/mL, respectively. Cisplatin-resistant variants of each cell line (SKOV3/ DDP and A2780/DDP) were generated by continuous exposure to cisplatin with IC50 concentrations for 72 hours and then allowed to recover for a further 72 hours. This procedure was carried out for approximately 6 months, and IC50 of SKOV3/DDP and A2780/ DDP cells was reassessed to be 5.72 ± 0.38 and 11.23 ± 0.46 μg/ mL, respectively. Cells were then maintained continuously in the presence of cisplatin with these new IC50 concentrations for a further 6 months. Two pairs of cisplatin-sensitive (SKOV3 and A2780) and cisplatin-resistant (SKOV3/DDP and A2780/DDP) human ovarian cancer cell lines were cultured in RPMI1640 (Life Technologies) containing 10% foetal bovine serum (FBS, Sigma), 100 IU/mL of penicillin and 100 µg/mL of streptomycin (Sigma).

| Apoptosis assay
Flow cytometry was performed following PI-and FITC-labelled annexin V staining according to the manufacturer's protocol (KeyGen Biotech, Nanjing, China). Briefly, after 48-hour incubation, cells were washed, resuspended in 200 μL binding buffer at 1 × 10 6 cells/mL and incubated with 5 μL FITC-annexin V. After 15-minute incubation at room temperature in the dark, 300 μL binding buffer together with 5 μL PI was added to each tube. The samples were incubated for 30 minutes at room temperature in the dark. Flow cytometry was performed within 1 hour.

| Transwell migration and invasion assays
In vitro transwell migration assays were performed in modified Boyden chambers with 8-mm-pore filter inserts in 24-well plates (BD Biosciences, San Jose, CA, USA). Briefly, the lower chamber was filled with DMEM containing 10% foetal bovine serum. A2780 and SKOV3 cells were collected after trypsinization, resuspended in 200 mL of conditional medium collected and transferred to the upper chamber. After 24 hours of incubation, the filter was gently removed from the chamber, and the cells on the upper surface were removed using a cotton swab. Cells fixed with 4% paraformaldehyde for 15 minutes and stained with 0.1% crystal violet.
In transwell invasion assay, the cells went through the Matrigel matrix membrane of the upper chamber. After 24 hours of incubation, the filter was gently removed from the chamber, and the cells on the upper surface were removed. Then cells were fixed and stained. 10 4 cells/well were seeded in ultra-low-attachment six-well plates (Corning, Acton, MA, USA) and cultured in serum-free DMEM/F12 containing B27 (1:50, Invitrogen, Carlsbad, CA, USA), 20 ng/mL human recombinant EGF (epidermal growth factor, Sigma-Aldrich, Saint Louis, MO, USA), 20 ng/mL bFGF (basic fibroblast growth factor, Sigma-Aldrich), 4 μg/mL heparin (Sigma-Aldrich) and 5 μg/ mL insulin (Sigma-Aldrich). Spheroids were imaged and its numbers were counted under phase-contrast microscopy after 7-10 days of cell seeding. Only the spheroid exceeding 50 μm in diameter was counted in the result.

| Lentiviral vector construction and recombinant lentivirus production
Gene encoding ISG15 and ISG15 (G156/157A) was cloned into the lentiviral vector (GeneChem Co., Ltd., Shanghai, China). DNA sequencing was performed by GeneChem to verify the sequence of the insert, and the identities were 100%. Following construction, A2780 and SKOV3 cells were cotransfected. Recombinant lentiviruses were harvested at 48 and 72 hours post-transfection, centrifuged to get rid of cell debris and then filtered. Ultimately, a concentrated lentivirus solution was obtained.
Quantification of extracted proteins was performed using the BCA protein assay kit. 20 μg of total protein was subjected to 12% SDS-PAGE and then transferred to PVDF membrane (Millipore Corporation). Antibody of ISG15, ISG15-Flag and KLF12 was diluted at 1:1000. And GAPGH antibody was diluted at 1:5000.

| Dot blot
3 × 10 5 cells/well were seeded into six-well plates with RPMI1640 containing 10% FBS, and the culture medium was replaced by the serum-free RPMI1640 after cell attachment. The cells were incubated for additional 3 days, and the supernatant was collected and centrifuged. Dot blot was performed by loading 100 μL of the supernatant on the NC membrane. The blots on the NC membrane were blocked with 5% skim milk for 1 hour and reacted with specific primary ISG15 antibody. ISG15 antibody was diluted at 1:1000. The acquired signals were detected using the ECL Western blotting system.

| Real-time reverse transcription polymerase chain reaction (RT-PCR)
Total RNAs were isolated using TRIzol reagent (Invitrogen) and reversely transcribed using SuperScript™ II RNase H-reverse transcriptase (Invitrogen). Amplification of cDNA was performed using SYBR Green PCR Master Mix (Applied Biosystems) on LightCycler480 II System (Roche). Each data was normalized against GAPDH and presented as ratio vs vehicle-treated control. The experiments were repeated for three times in triplicate.

| Label and capture nascent RNA
Labelling and isolation of newly synthesized RNA were performed using the ClickiT Nascent RNA Capture kit (Thermo Fisher Scientific) as previously reported. 33 Briefly, after pulsing with 0.2 mmol/L 5-ethynyl uridine (EU) for 4 hours, total RNA was isolated and subjected to nascent RNA capture, followed by analysis using real-time PCR.

| Analysis of mRNA stability
5 μg/mL of actinomycin D is selected as the optimal concentration to inhibit ISG15 transcription. Cells were exposed to 5 μg/mL of actinomycin D for the indicated time, and total RNA was isolated and analysed by quantitative RT-PCR. ISG15 mRNA expression was normalized to 18S rRNA. The value at time zero was set at 100%, and data were presented as a percentage of the value at time zero from three experiments repeated independently.

| Chromatin immunoprecipitation (ChIP)
Chromatin immunoprecipitation assay was performed using the Upstate Biotechnology Inc kit, and the detailed protocol was as previously reported. 34 The output DNA produced using this protocol was analysed using qPCR.

| Nude mice xenograft experiments
BALB/c-nu/nu mice (4-5 weeks old, female) (Beijing Vital River Laboratory Animal Technology) were subcutaneously inoculated with the serially diluted viable SKOV3/DDP cells. Euthanasia of the experiment mice were performed using overdose of sodium pentobarbital on day 28, and primary tumours were excised and weighed.
All animal procedures were approved and compiled with the guidelines of the Institutional Animal Care Committee of China Medical University.

| Statistics
ANOVA and post hoc Dunnett's test were used to analyse the statistical significance of the difference, which was defined as P < .05.
All experiments were repeated three times independently, and data from a representative experiment were presented as the mean ± SD (standard deviation).

| Ectopic ISG15 expression suppresses tumour formation in nude mice and low ISG15 expression is associated with poor prognosis in patients with ovarian cancer
Tumour formation in vivo was then investigated in nude mice. Results  (Table 1). According to chi-squared tests, low ISG15 expression was highly associated with histological grade, TNM classification and death rate (with P < .01, Table 1). Survival

| Nascent ISG15 mRNA decreased and degraded ISG15 mRNA increased in cisplatin-resistant ovarian cancer cells
The obvious involvement of ISG15 in the cancer stem cell-like features urged us to elucidate the mechanisms underlying regulation of ISG15 expression. Compared with the cisplatin-sensitive SKOV3 and A2780 cells, ISG15 mRNA decreased by more than 80% in SKOV3/DDP and A2780/DDP cells ( Figure 4A). On the other hand, only about 40% reduction of nascent ISG15 mRNA was detected ( Figure 4B). These

| ISG15 is downregulated by KLF12 at the transcriptional activation level via the CACCC elements located in −1187/−1013 and −672/−503 fragment
To identify the potential cis-acting regulatory elements located on the ISG15 promoter, luciferase reporter constructs containing truncated ISG15 promoter were constructed. Luciferase activity demonstrated that reporter construct containing −411/+53 segment did not generate any difference in SKOV3 and SKOV3/DDP cells  (Table S1). Pan-cancer analysis using starBase v3.0 project observed negative correlation between ISG15 and KLF12 in multiple cancers (Table S2). Western blot also demonstrated that KLF12 increased in cisplatin-resistant ovarian cancer cells ( Figure 5C). To further demonstrate the molecular mechanisms regulated by KLF12, ChIP-PCR assay was then performed. Compared with cisplatin-sensitive parental cells, recruitment of KLF12 by both −1187/−1013 and −672/−503 fragments was consistently increased in SKOV3/DDP ( Figure 5D) and A2780/DDP cells ( Figure 5E). KLF12 was then knocked down using lentivirus containing shKLF12 in cisplatin-sensitive and cisplatin-resistant SKOV3 or A2780 cells. ISG15 increased in SKOV3/DDP and A2780/DDP cells with KLF12 knock-down ( Figure 5F). ChIP-PCR data showed F I G U R E 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 | 4403 that KLF12 knock-down had no effect on the recruitment of KLF12 to −1187/−1013 fragment of ISG15 promoter in SKOV3 and A2780 cells ( Figure 5G), while KLF12 recruitment was significantly decreased due to KLF12 knock-down in SKOV3/DDP and A2780/DDP cells ( Figure 5H). The identical effects have arisen on −672/−503 fragment from KLF12 knock-down ( Figure 5I,J).

| D ISCUSS I ON
Cisplatin is utilized as the first-line medicine for patients with various cancers including ovarian cancer. Although the survival length of patients was largely improved by the combination of cisplatin chemotherapy and cytoreduction, its therapeutic application in ovarian cancer was compromised by the drug resistance and adverse sideeffect. Therefore, understanding the molecular mechanisms underlying cisplatin resistance might lead to potential therapeutic strategy for ovarian cancer treatment.
ISG15 plays an apparently contradictory role in cancers. On one side, ISG15 is highly expressed and functions as a tumour-promoting molecule in some cancers, 7,10,36,37 and its high expression contributes to cancer progression, including oesophageal, 38 oral, 39 nasopharyngeal 11 and pancreatic cancer. 7 It has been reported that free ISG15 promotes cancer stem cell-like phenotypes of PDAC via autocrine 40 and paracrine-mediated pattern. 7 On the other side, ISG15 and ISG15 conjugated targets are also been reported to suppress progression of some cancers, such as lung cancer, 13,41 glioblastoma 42 and cervical cancer. 14 ISG15 upregulation has been reported to promote cancer stem cell phenotype and increases cell resistance to cisplatin (DDP) treatment in nasopharyngeal carcinoma, 11 while ISG15 downregulation increases cisplatin resistance in lung cancer. 43 Consistent with the phenomena observed in lung cancer, we currently reported that significant decrease in ISG15 was observed in cisplatin-resistant ovarian cancer cell lines. Moreover, ectopic overexpression of wild-type ISG15 increased the sensitivity of cisplatinresistant cell lines. Interestingly, paradoxical role of ISG15 has been assigned to breast cancer, as free ISG15 plays a antitumour role by activating immune system in vivo in breast cancer, 15 while the conjugated ISG15 triggers a malignant transformation of breast cells. 12,44 Thereby, the difference between free and conjugated form of ISG15 might be an alternative explanation for its paradoxical function in distinct cancers. Our work showed that only conjugatable ISG15 (wild type) increased the sensitivity of cisplatin in ovarian cancer cells, while both conjugatable and nonISGylatable mutant ISG15 were involved in CSC-like characters. In addition, knock-down of ISG15 did not alter the responsiveness to cisplatin, but promoted CSC-like features of sensitive ovarian cancer cells. Importantly, our data showed that ISG15 positive expression was correlated with good prognosis in the patients with ovarian cancer. These data indicated that only ISG15 downregulation may not be enough to induce cisplatin resistance, but its downregulation might be implicated in maintenance of CSC-like features and make a significant contribution to cisplatin resistance of ovarian cancer cells. while KLF 12 serves as transcriptional repressor by interacting with F I G U R E 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 Therefore, these studies might suggest that KLF12 might function as a potential target to increase ISG15 expression for inhibition of CSC phenotype in the treatment of cisplatin-resistant ovarian cancer.

F I G U R E 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 F I G U R E 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

ACK N OWLED G EM ENTS
This work was partly supported by the National Natural Science Foundation of China (81602510, 81872257 and 81602439) and the Natural Science Foundation of Liaoning Province of China (2020-MS-185).

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.