ZNF300 promotes chemoresistance and aggressive behaviour in non‐small‐cell lung cancer

Abstract Objectives Chemoresistance induced by cisplatin has become the major impediment to lung cancer chemotherapy. This study explored the potential chemoresistant genes and underlying mechanisms of chemoresistance in NSCLC. Materials and methods Gene expression profile was integrated with DNA methylation profile to screen the candidate chemoresistant genes. Bioinformatic analysis and immunohistochemistry were used to analyse the association of a candidate gene with the characteristics of NSCLC patients. Recombinant lentivirus vectors were utilized to overexpress or silence candidate gene. Microarrays and immunoblotting were applied to explore the downstream targets of candidate gene. Xenograft models were established to validate the findings in vitro. Results An increased ZNF300 expression was detected in three chemoresistant cell lines of NSCLC, and the higher expression of ZNF300 was associated with poor OS of NSCLC patients. Cells with upregulated ZNF300 presented chemoresistance and enhanced aggressive growth compared to cells with downregulated ZNF300. ZNF300 inhibited MAPK/ERK pathways and activated CDK1 through inhibiting WEE1 and MYT1 and modulating MYC/AURKA/BORA/PLK1 axis. ICA and ATRA improved the anti‐tumour effect of cisplatin on chemoresistant cells by inducing differentiation. Conclusions ZNF300 promotes chemoresistance and aggressive behaviour of NSCLC through regulation of proliferation and differentiation by downregulating MAPK/ERK pathways and regulation of slow‐cycling phenotype via activating CDK1 by inhibiting WEE1/MYT1 and modulating MYC/AURKA/BORA/PLK1 axis. Cisplatin, combined with ATRA and ICA, might be beneficial in chemoresistant cases of NSCLC.


| INTRODUC TI ON
Global Cancer Statistics 2018 demonstrates that lung cancer remains the most commonly diagnosed cancer and is the leading cause of cancer death in both sexes combined worldwide. 1 Non-small-cell lung cancer (NSCLC) comprises approximately 80%-85% of all lung cancers, with adenocarcinoma and squamous cell carcinoma being the predominant histological subtypes of NSCLC. Clinically, except for a small portion of NSCLC patients diagnosed at an early stage (stage I or II), over 60% of lung cancer patients present with locally advanced or metastatic disease (stage III or IV) at the time of diagnosis.
For NSCLC patients with advanced stages, the multiple therapeutic choices including molecular targeted therapy and immunotherapy have emerged and reached the clinic, leading to improved clinical outcomes in recent years. 2,3 However, the approved molecular targeted therapy or immunotherapy is limited to a small portion of NSCLC patients with advanced stages who possess the genetic alteration. Therefore, the platinum-based chemotherapeutic agents continue to be the most widely prescribed chemotherapy for the vast majority of NSCLC patients, including those who failed in the genetic alteration-guided targeted therapies and patients of stage I to IIIa who have undergone the complete surgical resection. Thus, the use of platinum agents, including cisplatin (DDP), remains unwavering despite chemoresistance associated with treatment failures.
Over the past three decades, intense research has been conducted and several mechanisms that account for the cisplatin-resistant phenotype of tumour cells were explored. 4 Though the known mechanisms explain the cisplatin resistance at the molecular level to a certain extent, cisplatin resistance often exhibits a multifactorial nature. One of the hard facts is that the micromolecules targeting the currently known mechanisms of chemoresistance have not achieved the expected effect on the tumour suppression in the related clinical researches having little or no impact on progression-free survival (PFS) or overall survival (OS) of NSCLC. 5 Moreover, clinically, once the tumour cells present the cisplatin-resistant phenotype, the disease progresses malignantly and develops multidrug resistance to different chemotherapeutic agents with different pharmacological mechanisms, distant metastasis, and relapse, eventually leading to the treatment failure. Thus, there is an urgent need to continue exploring the mechanisms of chemoresistance in NSCLC.
In the study, the increased ZNF300 expression induced by cisplatin was found to be related to chemoresistance and malignant pro- DDP, H1650/DDP, H1915/DDP, H520/DDP and H446/DDP were established as described in our previous work. 6,7 All of the cells were tested for mycoplasma routinely. The identity and purity of cells were validated by short tandem repeat (STR) analysis profiling.

| Lentivirus vectors and transfection
Four sets of ZNF300-shRNA oligonucleotide were cloned into the LV3 lentiviral vectors (shZNF300) to silence ZNF300 in A549/DDP cells, whereas the full-length sequence coding for human ZNF300 transcript variant 1 was cloned into the LV5 lentiviral vectors to overexpress ZNF300 in A549 cells. The sequences of ZNF300-shRNA oligonucleotides are shown in Supplemental Materials and Experimental Procedures (Document S1).

| Gene expression profiles
For processing on microarrays, total RNA (1-2 µg) extracted using TRIzol and prepared using the 3'IVT Express Kit was utilized to synthesize double-stranded cDNA, which was hybridized on the microarrays with probes for >54 000 genome-wide transcripts. Microarray quality was verified by a signal intensity ratio of GAPDH 3′ to 5′ probe sets ≤ 3.0 and multi-chip normalization scaling factor ≤ 10.0.
Gene expression was normalized using the single-channel array normalization (SCAN)/Universal exPression Codes (UPC) method.
Genes with a UPC value of <0.2 in all samples were excluded.

| Immunohistochemistry of clinic specimens and tissue microarray
Immunohistochemistry (IHC) was performed, and percentage of the positively staining cells was scored as described previously. 8 Since by inhibiting WEE1/MYT1 and modulating MYC/AURKA/BORA/PLK1 axis. Cisplatin, combined with ATRA and ICA, might be beneficial in chemoresistant cases of NSCLC. the OS data of patients collected from Xinqiao Hospital were incomplete, a commercial tissue microarray containing 80 tumorous cases and 80 para-carcinoma tissues (controls) with the OS data of lung adenocarcinoma patients were stained with ZNF300 antibody (Shanghai Zhuohao Medical Science and Technology Co. Ltd). The appraisal and statistics were completed as described above.

| Co-immunoprecipitation assay (Co-IP)
The coding regions of ZNF300 were cloned into the ptt5 vector. A FLAG-tag was fused at the N-terminus of ZNF300. The recombinant vectors were then transformed into A549 cells to express ZNF300.
The whole-cell lysates were pre-cleared and incubated with rabbit anti-Flag antibody. The IP targets were disassociated from the immobilized antibodies and subjected to immunoblotting.

| Video and imaging of co-cultured cells
After seeded in 96-well plate and cultured overnight, cells were subjected to the real-time monitoring by High Content Analysis System (PerkinElmer Operetta CLS™) in fresh medium containing cisplatin (2 µg/mL), ATRA (5 µg/mL), ICA (20 µg/mL), separately or combined, for 7 days successively. Random vision fields were selected and transformed into videos. Pictures were captured at the end of each video.

| Xenograft tumorigenic experiment
Healthy Balb/c-nu mice (3-to 4-week-old female, n = 40) were purchased from Animal Laboratory of Beijing Vital River Laboratory Animal Technology Co. Ltd and housed in a climate-control specific pathogen-free (SPF) facility. Mice were divided into four groups that were inoculated subcutaneously with A549/DDP-shZNF300-NC, A549/DDP-shZNF300, A549-ZNF300-NC and A549-ZNF300 cells into the right flank (100 µL, 1 × 10 6 cells/animal), respectively. Each group was then randomly divided into two subgroups and raised by professional breeders. Tumour volume was measured every 3 days from day 0. When the tumour volume reached 100 mm 3 , one subgroup in each group received an intraperitoneal injection with cisplatin (0.02 mg/10 g: cisplatin/animal weight) every 3 days while the other subgroup with normal saline as controls (NS). After 42 days, all animals were sacrificed by cervical dislocation, and xenograft tumours were excised. Tumour size and weight were recorded, and tu-

| Statistical analyses
The numerical data were expressed as means ± standard deviation. One-way ANOVA was used to analyse the difference between means, and Pearson's chi-square test or Fisher's exact test was used for categorical variables. The independent effect of ZNF300 was assessed by performing the multivariate logistic regression analysis with adjustments for the possible confounding factors of age, gender and smoking habit to calculate the adjusted P-value. The Statistical Package for Social Science 15 for Windows was used for all statistical analyses (SPSS Inc). A P-value <.05 was considered to be statistically significant. All experiments were performed a minimum of three times.

| ZNF300 mediates chemoresistance of NSCLC
As shown in Figure S1A, IC 50 of cisplatin, gemcitabine, paclitaxel, docetaxel and pemetrexed on A549/DDP cells was significantly higher than that of its progenitor A549 cells. Because mitochondrial apoptosis induction accounts for one of the primary mechanisms of cisplatin anti-tumour activity, we compared the function changes of mitochondria between A549/DDP cells and A549 cell after treated with cisplatin. As shown in Figure S1B Table S1. Among the 15 upregulated genes, ZNF300 was selected as the candidate chemoresistant gene to be investigated comprehensively in the study because ZNF300 expression was negligible in A549 cells. However, higher expression of ZNF300 was detected in A549/DDP cells. Western blotting validated that ZNF300 expression was significantly higher in A549/DDP, H1650/ DDP and H520/DDP cells compared to the corresponding progenitor cells. ZNF300 was endogenously overexpressed in H1915 and H446 cells. No significant difference in ZNF300 expression was observed between H1915/DDP and H1915 cells, or between H446/DDP and H446 cells ( Figure 1A). The findings suggested that ZNF300 might mediate the chemoresistance of NSCLC rather than SCLC. BSP results ( Figure 1B) and changes of ZNF300 expression F I G U R E 1 The upregulated ZNF300 induced by cisplatin promoted the chemoresistance in NSCLC. A, Expression of ZNF300 in the chemoresistant cells and their progenitor cells of human lung cancer (A549/DDP, A549, H1650/DDP, H1650, H1915/DDP and H1915: cell lines of human lung adenocarcinoma; H520/DDP and H520: cell lines of human lung squamous carcinoma; H446/DDP and H446: cell lines of human SCLC). B, Methylation status of the ZNF300 promoter in the cells mentioned above. C, Expression of ZNF300 in A549, A549/ DDP, H1650, H1650/DDP, H520 and H520/DDP cells after treatment with azacitidine and belinostat, separately or combinedly. D, IC 50 of cisplatin on A549-ZNF300-NC, A549-ZNF300, A549/DDP-shZNF300-NC and A549/DDP-shZNF300 cells, respectively. E, Apoptosis of A549-ZNF300-NC, A549-ZNF300, A549/DDP-shZNF300-NC and A549/DDP-shZNF300 cells after treatment with cisplatin. F, Observation of xenograft tumours in nude mice treated with cisplatin or normal saline. The nude mice were inoculated subcutaneously with A549-ZNF300-NC, A549-ZNF300, A549/DDP-shZNF300-NC and A549/DDP-shZNF300 cells into the right flank, respectively. G, Growth curves of xenograft tumours in each subgroup treated with cisplatin or normal saline. H, Measurement of the final tumour weight in each subgroup treated with cisplatin or normal saline. Data are expressed as the mean ± SD (significant difference between two groups: *P < .05, **P < .01, ***P < .001) in A549, H1650 and H520 cells after incubation of 5-azacitidine or belinostat or combined usage of both ( Figure 1C) demonstrated that ZNF300 expression was regulated by promoter methylation. In the following experiments, the recombinant lentivirus vectors were utilized to overexpress ZNF300 in A549 cells or silence ZNF300 in A549/DDP cells to explore the role of ZNF300 in the chemoresistance of NSCLC ( Figure 1A).
To evaluate the potential role of ZNF300 in lung cancer, we retrieved ZNF300 in normal human tissues (http://www.biogps.   Figure S4A). Interestingly, two ZNF300 datasets of lung cancer appeared in the outlier analysis (Broet-Lung and Larsen-Lung, Figure S4B,C). Because only age, sex and stage were listed in Broet-Lung, 9 we used the data from Larsen-Lung to analyse the association of ZNF300 with the clinical characteristics of lung cancer patients. 10 The analysis revealed that ZNF300 was associated with poor OS of NSCLC patients when ZNF300 expression was stratified into three levels of <0.00, 0-0.25 and >0.25 ( Figure S4D). Additionally, the expression of ZNF300 in anaplastic oligodendroglioma ( Figure S4E,F), glioblastoma ( Figure S4G), skin basal cell carcinoma ( Figure S4H), invasive ductal breast carcinoma ( Figure S4I) and invasive lobular breast carcinoma ( Figure S4J) was significantly higher than that of the corresponding normal tissues.
To clarify the relationship between ZNF300 and MAPK/ERK pathways, AZD6244, an inhibitor, and hesperetin, an agonist of MAPK/ERK pathways, were used to treat the relevant cells. As displayed in Figure 4Aa, p-ERK1/2, p-p38 MAPK, and CD61 decreased, while p15 and Nanog increased in A549-ZNF300-NC and A549/DDP-shZNF300 cells after treatment with AZD6244.
ZNF300 was predicted to interact with E2F3 and SPI1 (https:// www.gcbi.com.cn/). The immunoblotting showed that ZNF300flag, used as bait, could bind to E2F3 (Figure 4Ga), indicating that ZNF300 might interact with E2F3 to function in the chemoresistance and aggressive behaviour of tumour cells. As shown in Figure 4Gb, there was no significant difference of E2F3 expression between the cells with upregulated ZNF300 and the cells with downregulated ZNF300, implying that E2F3 might not be regulated by ZNF300.

| ICA and ATRA improve the anti-tumour effect of cisplatin on chemoresistant cells by inducing differentiation
All-trans retinoic acid (ATRA) and icariin (ICA) were reported to induce the cell differentiation. 11,12 Thus, we used the two chemical compounds to test whether they could enhance the anti-tumour effect of cisplatin on the chemoresistant cells by inducing differentiation. As shown in Figure 5A

| D ISCUSS I ON
Chemoresistance induced by cisplatin has become the major im- ZNF300 gene was originally isolated from a human embryo and overexpressed in the heart, brain, skeletal muscle and testis. 15,16 Being a typical member of the zinc finger protein family, ZNF300 was reported to exert its transcriptional impact in the nucleus via the KRAB domain. 15 It has been shown to play a pivotal role in cell differentiation, 17  Since ZNF300 was one of the three methylated genes identified in the plasma cell-free DNA (cfDNA) of hepatocellular carcinoma as compared to liver tissue DNA, it can be used as a biomarker for early detection and high-risk screening of hepatocellular carcinoma. 22 ZNF300 expression was higher in the desmoid tumours than in the normal tissues, but its function is still unknown. 23 Since the cells with upregulated ZNF300 presented slower proliferation and slower cell cycling compared to the cells with downregulated ZNF300, we focused on the differentially expressed genes related with proliferation, growth, differentiation and cell cycle between A549-ZNF300-NC and A549-ZNF300 cells. The analysis displayed that most of the genes associated with growth, differen-  29,30 Recently, the wild-type p53 expression has been revealed to be a driver of therapy resistance by initiating a slow-cycling phenotype in melanoma. 31 In the study, we also observed a slow-cycling phenotype

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

E TH I C S A PPROVA L A N D CO N S E NT
Informed consent was obtained from all the patients participating in this study. All animal experiments were approved by the Experimental