Loss of CDKN2A at chromosome 9 has a poor clinical prognosis and promotes lung cancer progression

Abstract Objective This study aimed to identify critical genes involved in the tumor biology of lung cancer via datamining of The Cancer Genome Atlas (TCGA) with special focus on gene copy number variation. Methods Genomic deletion and amplification were analyzed with cBioportal online tools. Relative expression of Cyclin Dependent Kinase Inhibitor 2A (CDKN2A) was analyzed by both real‐time polymerase chain reaction (PCR) and Western blot. The abundance of methylthioadenosine phosphorylase (MTAP) and epithelial‐mesenchymal transition markers were analyzed by real‐time PCR. Cell proliferation was determined by cell counting kit‐8 method and cell viability was measured with 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide assay. The cell migration and invasion were measured with transwell chamber assay, and migrative capacity was further evaluated by wound healing assay. Results We found the frequent loss of CDKN2A was associated with its downregulation in lung cancer, and siRNA‐mediated CDNKN2A knockdown significantly stimulated cell proliferation, invasion, and migration. Mechanistically, we unraveled that MTAP, which was positively correlated with CDKN2A, predominantly mediated the antitumoral function of CDKN2A in lung cancer. Conclusion Our study consolidated the involvement of CDKN2A‐MTAP signaling in the context of lung cancer.

of air pollution. Therefore, avoidance of smoke and air pollutant constitutes the first line of prevention (Simon, (2016)). Clinical treatments for lung cancer are heavily dependent on molecular subtype, progression stage and the overall personal health status, and include surgery, chemotherapy, radiotherapy, targeted therapy, and emerging immunotherapy (Hirsch et al., 2017;Liu et al., 2020). Despite of the advances in therapeutic interventions, the overall survival of lung cancer is still relatively unsatisfactory with the 5-year survival rate of around 20% in the United States.
Activation of oncogenes and/or inactivation of tumor suppressor genes have been long acknowledged as the initiating events in the tumorigenesis of lung cancer. The environmental carcinogens are frequently identified to be involved in the generation of mutations in these genes. K-ras proto-oncogene mutations are reported to account for 10%-30% cases of lung adenocarcinomas (Ferrer et al., 2018), and 40% of NSCLCs are characterized with echinoderm microtubule associated protein-like 4/anaplastic lymphoma kinase tyrosine kinase fusion gene (Sasaki et al., 2010). The epidermal growth factor receptor (EGFR) which signals cell proliferation and angiogenesis is commonly and aberrantly over-activated in NSCLC, which in general shows favorable response to EGFR-inhibitor treatments (Tu et al., 2017). In addition, epigenetic mechanisms have also critically contributed to lung cancer via DNA methylation, histone modification, and microRNA network modulation. In this case, Cyclin Dependent Kinase Inhibitor 2A (CDKN2A, OMIM association number 600160), the essential cell cycle regulating factor, is increasingly recognized to be involved in the pathological process of lung cancer. The first two studies identified deletion of CDKN2A in multiple human cancers in 1994 (Kamb et al., 1994;Nobori et al., 1994), and subsequent investigation uncovered that CDKN2A loss was restricted to a subpopulation of lung cancers with intact retinoblastoma (RB) (Otterson et al., 1994), which was further disclosed as reciprocal RB inactivation and CDKN2A expression in primary lung cancers and cell lines (Shapiro et al., 1995). In the following study, Merlo et al. identified that 5′CpG island methylation was associated with suppressed transcription of CDKN2A in human cancers , which was then found to be frequently associated with aberrant inactivation of CDKN2A . The study performed by Belinsky et al. suggested that aberrant methylation of CDKN2A was an early event in lung cancer patients and served as a potential biomarker for diagnostic purpose (Belinsky et al., 1998). Here, we retrieved the publicly available database and unraveled the causal relationship between genomic deletion and downregulation of CDKN2A in lung cancer patients, which clinically associated with unfavorable prognosis. The antitumoral properties of CDKN2A was uncovered in cell culture. Most importantly, here, we further identified a positive correlation between CDKN2A and Methylthioadenosine Phosphorylase (MTAP, OMIM association number, 156540) in lung cancer. The suppressed expression of MTAP predominantly contributed to the oncogenic signaling in CDKN2A-deficient lung cancer.

| The Cancer Genome Atlas (TCGA) data analysis
Genomic deletion/amplification, copy number variation of CDKN2A, and survival curve in terms of CDKN2A deletion status in lung cancer patients were analyzed against TCGA datasets containing multiple subtypes of lung cancer, such as NSCLC, SCLC, and mesothelioma, using the cBioportal algorithm (http://www.cbiop ortal.org).

| Cell culture
The human lung cancer cell lines A549 and H322 were purchased from the American Type Culture Collection (NY, USA) and maintained in RPMI-1640 medium containing 10% of fetal bovine serum (FBS, Gibco, MA, USA) and 1% of penicillin/streptomycin (Hyclone, MA, USA). All cells were cultured in a humidified CO 2 incubator (5%). Cell lines were verified using the short tandem repeat analysis. Mycoplasma contamination was regularly monitored by PCR method.

| 2,5-diphenyl tetrazolium bromide (MTT) assay
MTT Assay Kit (ab211091, Abcam) was employed to determine the viability in exponentially growing cells. The indicated cells were plated in 6-well plates and cultured overnight for attachment. The 1:1 mixture of serum-free media and MTT reagent (100 μl) was then replaced and followed by 3 h of incubation at 37℃. 150 μl of MTT solvent solution was then replaced and followed by 15 min of incubation at room temperature on a shaker. Absorption at 590 nm was recorded on a microplate reader (Berthold Technologies, Bad Wildbad, Germany).

| Cell counting assay
Cell counting kit-8 (CCK-8, Dojindo, Dalian, China) was used to measure cell proliferation following the manufacturer's instructions. Cells were first seeded in 96-well plates (1500 cells/well) and cultured for 24 h. About 10 μl/well of reagent solution was added and followed by incubation for 2 h at 37℃. Absorption was recorded with a microplate reader (Berthold Technologies).

| Cell invasion, migration, and wound healing assays
Cell invasion and migration were evaluated with Boyden chambers (BD Biosciences, NJ, USA) with or without Matrigel precoating, respectively. Wound scratch was created with sterile tips on 6-well plates. Gap closure was continuously monitored for 24 h. All experiments were performed in triplicate and result was normalized to cell numbers.

| Statistical analysis
Results are presented as means ±standard deviation (SD). All the experiments were repeated as biological replicates for at least three times. Differences between groups were estimated with Student's t test, one and two-way ANOVA analysis with a post hoc test where appropriate. p values <0.05 were considered as statistically significant.

| Loss of CDKN2A in lung cancer
We first analyzed the gain and loss of genomic fragments in lung cancer at genome-wide level in the TCGA database with the cBioportal algorithm (Figure 1a), and uncovered a characteristic depletion of genomic region in chromosome 9 across the CDKN2A coding sequence (Figure 1b). We further demonstrated the highest frequency of CDKN2A depletion among four datasets in lung cancer patients including PanCancer Atlas (Figure 1c), Nat Genet 2016 (Figure 1d), Provisional (Figure 1e), and Cancer Discov 2017 (Figure 1f), which indicated the universal genomic depletion of CDKN2A and its essential roles in the tumorigenesis of lung cancer. Therefore, our following investigation focused on the importance of CDKN21 in this disease via analyzing both clinical and in vitro data.

| Loss of CDKN2A correlated with poor survival outcomes in lung cancer
We next investigated the potential linkage between CDKN2A depletion and clinical outcomes in lung cancer patients via analyzing the survival curve. As shown in Figure 2a, depletion of CDKN2A significantly associated with poorer survival. And consistent observation was noticed in the "TCGA, Provisional" dataset as well (Figure 2b). Likewise, CDKN2A deficiency indicated a poorer disease-free survival in the "TCGA, Provisional" dataset ( Figure 2c). These results suggested the potent tumor suppressor role of CDKN2A in lung cancer, especially in tumorigenesis and tumor progression. Further analysis uncovered the relatively low expression of CDKN2A transcript in lung cancer patients with CDKN2A depletion (Figure 2d), which highlighted the important contribution of genomic loss to the downregulation of CDKN2A.

| Loss of CDKN2A promoted lung cancer migration and invasion
We next sought to clarify the potential effects of CDKN2A on cell migrative and invasive behaviors in lung cancer cells. As shown in Figure 4a, siRNA-mediated knockdown of CDKN2A greatly stimulated both cell migration and invasion in A549 cells, with statistical result presented in Figure 4b. The similar observation was noticed in H322 cells as well (Figure 4c,d). The stimulatory action of CDKN2A deficiency on cell migration was further interrogated with wound healing assay. As presented in Figure 4e, the gap closure was tremendously accelerated in CDKN2A-deleted A549 cells. In line with the suppressive roles of CDKN2A on cell migration and invasion, the molecular profiling showed decreased epithelial markers including CDH1, TJP1, and OCLN, and increased mesenchymal markers including ZEB1, FN1, and EZH2 ( Figure 4f). Therefore, our data suggested the suppressive roles of CDKN2A on both cell migration and invasion in addition to cell proliferation. shown as mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant

| CDKN2A-regulated MTAP expression
Next, we sought to understand the molecular mechanism underlying the CDKN2A-stimulated cell proliferative and metastatic phenotype. In view of the nature of CDKN2A as a critical cell cycle-related protein, alteration of which fundamentally influenced cell cycle progression and expression of an array of genes. Therefore, we analyzed the transcriptome in terms of CDKN2A expression status, and the results are presented in Figure 5a as a volcano plot. We identified MTAP with high correlation to CDKN2A in TCGA dataset (r = 0.5691, p < 0.0001, Figure 5b). We further demonstrated significant downregulation of MTAP transcript in CDKN2A-silenced A549 cells (Figure 5c). To clarify the role of MTAP in mediating the antitumoral activities of CDKN2A, we then ectopically over-expressed MTAP in CDKN2A-deficient A549 cells (Figure 5d). Cell proliferation stimulated by CDKN2A knockdown was completely suppressed by simultaneous overexpression of MTAP (Figure 5e). Likewise, both cell migration and invasion that were greatly induced in CDKN2A-deficient A549 were compromised by supplementation with ectopic MTAP (Figure 5f). Therefore, our data clearly suggested that MTAP predominantly mediated the tumor suppressor roles of CDKN2A in lung cancer.

| DISCUSSION
In this study, we systematically analyzed genomic alterations with a specific focus on copy number variation in lung cancer patients. Retrieval of TCGA dataset showed significant deletion in chromosome 9 spanning the CDKN2A coding region, which was further identified as one of the most altered genes in multiple datasets. Importantly, lung cancer patients with CDKN2A loss manifested relatively poor overall survival and disease-free survival. The CDKN2A transcripts were notably downregulated in patients with genomic depletion. We further demonstrated that ectopic CDKN2A expression remarkably inhibited cell proliferation and viability in lung cancer cells, while siRNA-mediated CDKN2A knockdown greatly promoted cell proliferation and viability. In addition, knockdown of CDKN2Astimulated cell migrative and invasive behaviors, while ectopic introduction of CDKN2A significantly compromised the migrative and invasive capacities. Profiling the epithelial-mesenchymal transition-related molecular markers demonstrated decreased CDH1, TJP1, and OCLN and increased ZEB1, FN1, and EZH2 in response to CDKN2A silencing in lung cancer cells. Mechanistically, we identified MTAP as a positively correlated gene with CDKN2A in lung cancer, which was significantly downregulated in CDKN2A-deficient cells. Complementation with MTAP completely inhibited cell proliferation, migration, and invasion stimulated by CDKN2A knockdown. Therefore, our study highlighted the critical contributions of loss of CDKN2A in the tumorigenesis and progression of lung cancer, which predominantly depended on downstream MTAP. Our study highlighted the antitumor properties of both CDKN2A and MTAP. The downregulation of CDKN2A was long recognized as a major player in the tumorigenesis of lung cancer via epigenetic suppression. Bradly et al. (2012) reported that CDKN2A promoter hypermethylation impacted the outcome in young lung cancer patients. Xiao et al. (2014) suggested the diagnostic values of CDKN2A methylation in exhaled breath condensate for early detection of NSCLC. Tuo et al. (2018) proposed CDKN2A promoter methylation as a valuable biomarker for NSCLC as well via meta-analysis. In addition, genomic deletion spanning CDKN2A in lung cancer patients has been increasingly acknowledged. Chen et al. reported the deletion of both FHIT and CDKN2A mRNA in biopsy specimens acquired from lung cancer patients via bronchoscopy for diagnostic purposes (Chen et al., 2013). Jiang et al. (2016) showed that coexistence of CDKN2A deletions with overactivation of EGFR signified a poorer response to EGFR-targeting inhibitor in lung adenocarcinoma patients. Panani et al. (2009) demonstrated that numerous abnormalities in chromosome 9 and CDKN2A deletion were detected by FISH in NSCLC patients. Andjelkovic et al. (2011) proposed the concurrent alterations of both CDKN2A and PTEN as potential biomarkers for particular subgroups of NSCLC patients. Our data were in support of the antitumor properties of CDKN2A, deletion of which was notably detected in many lung cancer patients, and therefore, suggested a fundamental role in the tumorigenesis of this disease. Along with epigenetic mechanisms, the biallelic inactivation of CDKN2A might heavily contribute to lung cancer incidence.
Our results also highlighted the predominant roles of MTAP in mediating the antitumoral activities of CDKN2A in lung cancer cells. Complementation with MTAP completely inhibited cell proliferation, migration, and invasion which was greatly induced in CDKN2A-deficient cells. We provided the direct evidence in support of the regulation of MTAP by CDKN2A, and MTAP was significantly downregulated in