ABSTRACT
Background Frequency of somatic copy number deletion of CDKN2A gene is upto 60% in human esophageal squamous cell cancer. However, it is unknown whether CDKN2A deletion could be a biomarker for esophageal squamous cell dysplasia (ESCdys) due to absence of a feasible detection method.
Methods Information on base-resolution common deletion region (CDR) for CDKN2A were extracted from published articles and confirmed with whole genome sequencing (WGS). A quantitative PCR targeted to the CDR (P16-Light) was established and used to detect CDKN2A copy number in ESCdys biopsies from patients (n=205) enrolled in a multicentre follow-up study.
Results A 5.1-kb CDR from the CDKN2A/P16INK4A promoter to intron-2 was firstly characterized in 90% (83/92) of cancer cell lines and confirmed with WGS. The CDR covers CDKN2A exon-2 which is the essential coding exon for both P16INK4a and P14ARF. And CDKN2A exon-2 deletion markedly promoted the proliferation and invasion and inhibited the apoptosis of HEK293T cells. In the follow-up study, both somatic CDKN2A deletion and amplification are prevalent in mild/moderate (m/M) ESCdys. CDKN2A deletion was less common among 70 patients whose ESCdys regressed than among 135 patients whose ESCdys progressed or remained stable, and CDKN2A amplification was more common in the patients who regressed than in the patients whose m/M ESCdys persisted or progressed over a median of 37 months of follow-up (p<0.0001).
Conclusion There is A 1.5-kb CDR within CDKN2A gene in many cancers. CDR deletion could inactivate both P16INK4a and P14ARF and associate wtih prognosis of ESCdys.
INTRODUCTION
Somatic CDKN2A copy number deletion is a landmark of human cancer (Beroukhim et al. 2010). The frequency of CDKN2A deletion detected by single nucleotide polymorphism (SNP) microarray or whole genome sequencing (WGS) was found to be 30% to 60% in bladder cancer, melanoma, pleural mesothelioma, head and neck cancer, glioblastoma, and esophageal squamous cell cancer (ESCC), with an average frequency of 13% (1384/10967) in pan-cancer datasets in The Cancer Genome Atlas (TCGA) (Supplemental Figure S1A) (Mermel et al. 2011; Cerami et al. 2012; Gao et al. 2013; Song et al. 2014; The Cancer Genome Atlas Research Network 2017; Cui et al. 2020). CDKN2A deep-deletion is associated with downregulation of CDKN2A gene expression, while CDKN2A amplification is associated with upregulation of CDKN2A gene expression in Pan-TCGA cancers (Supplemental Figure S1B).
It is well known that genetic CDKN2A inactivation contributes to malignant transformation, cancer metastasis, and therapy sensitivity of cancers to drugs, including CDK4/6 inhibitors and their combination with PD-1 blockades (Deng et al. 2018; Jerby-Arnon et al. 2018; Zhang et al. 2018; Yu et al. 2019). However, current gene copy number detection methods, including fluorescence-in-situ hybridization (FISH) and WGS, are not sensitive enough or are too costly for routine clinical use. While the amplification of oncogenes (such as EGFR, c-ERBB2, c-MYC, and c-MET) are increasingly driving decision-making for precise cancer treatments, clinical applications of somatic copy deletions of tumor suppressor genes, including CDKN2A, remain rare because of the lack of a feasible detection assay (Patel et al. 2014).
RESULTS
Characterization of a CDKN2A common deletion region (CDR) in human cancers
It has been previously reported that a homozygous deletion of approximately 170 kilobase pairs (kb), including the CDKN2A locus, can be detected in human cancers by microsatellite analyses (Cairns et al. 1995). To characterize the base-resolution genomic coordinates of CDKN2A deletions in cancers, we extracted sequence information of interstitial CDKN2A deletions from available published articles (Supplemental Table S1). We found a 5.1-kb CDR (chr9: 21,970,277 - 21,975,386, hg19) that spanned from the P16INK4a promoter to intron-2 in 83 (90%) of 92 reported cancer cell lines or tissue samples containing interstitial CDKN2A deletions (Figure 1). This CDR sequence is exactly the same as the CDKN2A deletion fragment in the HCC193 lung cancer cell line (Sasaki et al. 2003). The CDR coordinates were also confirmed in our WGS datasets (sequencing depth, 36×) of 18/18 gastric carcinomas (GC) (Xing et al. 2019), in which interstitial CDKN2A deletions were identified (Figure 1; Supplemental Table S1).
Base resolution of the genomic coordinates of CDKN2A deletion in human cancers. The top chart displays the coordinates of whole deleted fragments by Sanger sequencing of chromosome 9p21 in CDKN2A-deleted cancer cell lines and tissue samples (n=92, deep-blue lines), and in gastric carcinomas (GC, n=18, deep-red lines) by whole genome sequencing (WGS). The sample ID is labeled under each column. The bottom chart displays an amplified view of the CDKN2A gene, where the 5.1-kb common deletion region (CDR) occurred from the P16INK2a promoter to intron-2 (highlighted with a red dashed line rectangle), which is exactly the same region as the deleted CDKN2A fragment in the HCC193 lung cancer cell line (highlighted with a black line). The detailed deletion coordinates for each sample are listed in Supplemental Table S1. Each line represents a CDKN2A deletion fragment. The locations of P16INK4a and P14ARF (green lines) and exon-1α/2/3 (black dots) are also labeled as landmarks.
It is well known that germline CDKN2A inactivation can lead to a high predisposition for melanoma and pancreatic cancer (Hussussian et al. 1994; Freedberg et al. 2008; Harinck et al. 2012). Interestingly, we found that 14 of 15 CDKN2A allelic variants in the Online Mendelian Inheritance in Man (OMIM) database are located within the CDR sequence (Supplemental Figure S2) (Hamosh et al. 2005; Amberger et al. 2009). These phenomena indicate that both inherited and somatic defects of the CDKN2A CDR region are cancer drivers.
Both P16INK4a and P14ARF mRNAs are transcribed from the human CDKN2A gene at chromosome 9p21, but with different transcription start sites (Stone et al. 1995); they share the same exon-2 but have different translation reading frames. Because CDKN2A exon-2 is the essential exon for coding P16INK4a and P14ARF proteins and because it located within the CDR, our above findings indicate that P16INK4a and P14ARF are coinactivated in 87% (96/110) of human cancer cell lines and tissues containing CDKN2A CDR deletion (Figure 1).
The P16INK4a and P14ARF coinactivation promotes cell proliferation and migration and inhibits apoptosis
To study whether P16INK4a and P14ARF coinactivation plays a larger role in cancer development than individual P16INK4a or P14ARF inactivation alone, we knocked out P16INK4a-specific exon-1a (P16-KO), P14ARF-specific exon-1b (P14-KO), and P16INK4a and P14ARF-shared exon-2 (P14&P16-DKO) with CRISPR/Cas9 in P16INK4a and P14ARF-expressing human non-tumor embryo kidney HEK293T cells (Figure 2A). Two KO subclones for each genotype were obtained and pooled for the following experiments.
Comparison of behavioral alterations of HEK293T cells with various CDKN2A knockout (KO) genotypes. (A) KO of CDKN2A exon-1b, exon-1a, and exon-2 in cells by CRISPR/Cas9 and corresponding single guide RNA (sgRNA). Locations of sgRNAs and exons are labeled, and the 5.1-kb common deletion region is highlighted with a pink shadow (top chart). (B) The proliferation, migration, and invasion of pooled clones with various CDKN2A inactivation genotypes was assessed in long-term dynamic IncuCyte analyses. Each point represents the average value of 9 or 12 wells. The SD value is also displayed. *: P<0.01 compared with the CDKN2A wild-type control cells. (C) The proportions of apoptotic and dead cells with various CDKN2A inactivation genotypes were detected in FACS analyses using annexin V-isothiocyanate (FITC, FL1-A) and propidine iodide (PI, FL2-A) staining. The values in parentheses indicate the total proportions of early and late apoptotic cells with various CDKN2A KO genotypes. (D) The amounts of total RB1 and phosphorylated RB1 (pRB1) proteins in cells with various CDKN2A KO genotypes was determined by Western blot analyses.
As expected, the proliferation and migration of P14&P16-DKO HEK293T cells were highest among the cells with different genotypes, and the invasion of P14-KO, P16-KO, and P14&P16-DKO cells were similarly increased, as shown by long-term dynamic IncuCyte analyses (Figure 2B). The proportions of these HEK293T cells that were undergoing apoptosis or death were similarly decreased (Figure 2C).
P16INK4a and P14ARF proteins play crucial roles in cell senescence, apoptosis, and cell cycle arrest, preventing cell replicative stress via P16INK4a-CDK4-RB1 and P14ARF-MDM2-P21CIP1-CDK2-RB1 pathways (Serrano et al. 1993 and 1996; Kamijo et al. 1997; Chen et al. 2009). The amount of phosphorylated RB1 (pRB1) protein was highest in the P14&P16-DKO clones by Western blot analysis (relative to both GAPDH and RB1; Figure 2D). These results indicate that P16INK4a and P14ARF coinactivation leads to a more dramatic effect on cell proliferation than individual inactivation. This may account for the phenomena that most CDKN2A genetic diseases (mainly cancers) in the OMIM database are related to exon-2 variations (12/15=80%; Supplemental Figure S2).
Establishment of a quantitative PCR assay (P16-Light) to detect somatic CDKN2A CDR deletion
To study the application potential of somatic copy number variations (SCNVs) of the CDKN2A gene, we designed and experimentally evaluated a set of multiplex quantitative PCR assays and finally established a CDKN2A CDR-specific quantitative multiplex PCR assay for detecting the copy number of a 129-bp amplicon within CDKN2A/P16INK4a intron-2 (P16-Light, Figure 3A), which covers 86% (94/110) of known CDKN2A deletion fragments (Figure 1). Using genomic DNA from human A549 cells (with a homozygous CDKN2A deletion) and RKO cells (with 2 wild-type CDKN2A alleles) as CDKN2A CDR deletion-positive and deletion-negative controls, respectively, the proportions of CDKN2A CDR copy number were linearly correlated with the ratios (0 - 100%) of RKO cell DNA and A549 cell DNA in the input mixtures when the A549 DNA was spiked in at different proportions for the P16-Light analyses (Figure 3B). Furthermore, there was a high reproducibility when DNA with homozygous deletion of CDKN2A was present in ≥20% of the cells; results were verified in ten experimental repeats that were performed on different days (Figure 3C). Thus, when the proportion of CDKN2A copy number is significantly decreased or increased in a tumor sample relative to the paired blood DNA sample (p<0.05) in the P16-Light analyses, the tumor is defined as somatic CDKN2A deletion-positive or amplification-positive.
Detection of the copy number of CDKN2A intron-2 with quantitative gene-specific multiplex PCR (P16-Light). (A) The location of the 129-bp amplicon within the common deletion region (CDR) and its host genes. (B) The linear relationship between the proportion of CDKN2A copy number deletion and ratios of RKO cells (with two wild-type CDKN2A alleles) spiked with A549 cells (with a homozygous CDKN2A deletion). (C) Stability of the proportion of the CDKN2A copy number deletion by P16-Light during ten experiments over different days. The RKO cell DNA templates were spiked with 0, 10%, 20%, 25%, and 30% A549 cell DNA. Each column represents the average proportion of CDKN2A copy number deletion in triplicate. Exp. 1 - 10: the results of 10 repeated experiments performed on different days. *: compared to the 100% RKO control in chi-square test, p<0.05.
Somatic CDKN2A CDR deletion blocks regression of esophageal squamous cell dysplasia (ESCdys)
Recently, a panel of SCNV biomarkers has been reported to assess the progression potential of Barrett’s esophagus to esophageal adenocarcinoma (Killcoyne et al. 2020). P16INK4a is frequently inactivated by DNA methylation in both this cancer and this precancer. Our previous work, along with others, demonstrates that P16INK4a methylation is significantly associated with malignant transformation of low-grade gastric dysplasia, oral epithelial dysplasia and Barrett’s esophagus (Sun et al. 2004; Hall et al. 2008; Jin et al. 2009; Liu et al. 2015). This encouraged us to study whether somatic CDKN2A deletion could also be used to assess the prognosis of ESCdys, from which most ESCCs develop.
We are performing a long-term prospective multicentre endoscopic screening program among residents aged 40-69 yrs in populations in rural areas of China in which residents had a high risk of ESCC. 32 In this program, patients with severe ESCdys or ESCC are sent for endoscopic or surgical therapy, patients with mild or moderate (m/M) ESCdys are scheduled for repeat endoscopy within 3 years, and patients with inflammation (esophagitis) or normal mucosa are excluded from the follow-up cohort, according to the National Esophageal Cancer Screening guideline (Wang et al. 2020). For the current evaluation, we divided patients with baseline m/M ESCdys into three groups based on the results of their first repeat endoscopy (a median of 37.4 months after baseline): patients (n=74) whose worst squamous diagnosis progressed to severe ESCdys or ESCC; patients (n=95) who had persistent m/M ESCdys; and patients (n=94) who regressed to inflammation or normal (Supplemental Figure S3). Information on CDKN2A SCNVs was obtained by P16-Light analysis in the baseline m/M ESCdys lesions from 78.0% (205/263) patients (Figure 4, Supplemental Table S2). Interestingly, both somatic CDKN2A deletion and amplification were prevalent in these ESCdys lesions (33.7% for deletion and 27.3% for amplification, relative to the gene copy number in the paired blood DNA samples). More CDKN2A amplification was detected in patients with low annual income than those with high income, and in patients drinking tea than those who did not (Supplemental Table S3). Because no significant difference in somatic CDKN2A deletion or amplification was observed between the ESCdys patients in the progression and persistent groups (p=0.1934; Supplemental Table S4), we merged these two groups together in further frequency comparisons. No significant difference was observed between the merged progression & persistent group and the regression group in baseline grades of ESCdys, location of ESCdys within the esophagus, sex or age of the patients, or the patients’ city or county of residence (Table 1).
Participant flow diagram
Comparison of the prevalence of somatic CDKN2A gene copy number variations (SCNVs) in baseline esophageal squamous cell dysplasia biopsy samples from patients in the progression & persistent group and patients in the regression group *Linzhou, Yangzhong and Feicheng
Notably, the positive rate of somatic CDKN2A deletion was much lower in the baseline ESCdys lesions from the 70 patients in the regression group than in the baseline ESCdys lesions from the 135 patients in the progression & persistent group, and the positive rate of somatic CDKN2A amplification was much higher in the baseline ESCdys lesions in the regression group than in the baseline ESCdys lesions of the progression & persistent group (p<0.0001). These significant differences also remained in most strata of baseline grades of ESCdys, location of ESCdys within the esophagus, sex and age of the patients, and the patients’ city or county of residence (Table 1).
The cumulative incidence of regression of ESCdys with CDKN2A deletion was significantly lower than those without (Fine-Gray univariate analysis: p<0.0164) while the cumulative incidence of progression of ESCdys with CDKN2A deletion was significantly higher than those without (p=0.0071; Figure 5A). In the multivariate Fine-Gray analysis, CDKN2A deletion was still a significant independent regression predictor (p=0.0140; Table 2). In additon, the marital status, annual income, and tea consumption were also significantly correlated to regression of ESCdys. Thus, we used these significant factors to construct a monogram for predicting regression of m/M ESCdys (Figure 5B). The area under the receiver operating characteristic (ROC) curve (AUC) to predict regression of m/M ESCdys within 3 years was 72.4% (95% confidence interval: 64.0%–80.7%) (Figure 5C). The calibration curves at 3 and 5 years showed good agreement between the estimations with the nomogram and actual observations (Figure 5D). In addition, a significant majority of the patients with baseline somatic CDKN2A deletions still had somatic CDKN2A deletions in the follow-up biopsy samples (n=170) (Supplemental Figure S4).
Comparison of cumulative incidence function (CIF) for somatic CDKN2A copy number deletion in human esophageal mucosa biopsy samples from patients in the follow-up study. (A) Cumulative incidence of regression and progression of esophageal squamous cell dysplasia (ESCdys) with and without CDKN2A deletion using the competing risk analysis. The progression of ESCdys was used as a competing risk. The p-values in the Fine-Gray univariate analysis are labled. (B) Competing-risk nomogram including CDKN2A SCNV, the marrial status, annual income, and tea consumption to predict regression of ESCdys following start of follow-up in the presence of the competing risk progression of ESCdys. (C) The receiver operating characteristic (ROC) curve for regression of ESCdys within 3 years, calculated using the prediction model constructed with CDKN2A SCNV, the marrial status, annual income, and drinking tea. (D) The calibration curves at 3 and 5 years between predicted and actual observed probability of regression of ESCdys
The results of univariate and multivariate Fine-Gray regression analyses*
Taken together, these findings indicate that somatic CDKN2A deletions are usually persistent, and they may block regression and promote progression of m/M ESCdys.
DISCUSSION
ESCC is one of the main causes of cancer death in China (Bray et al. 2018). Most ESCCs progress from ESCdys lesions. While a panel of gene SCNVs has been used to predict the malignant transformation of Barrett’s esophagus within the lower esophagus (Killcoyne et al. 2020), a clinical biomarker is not yet available to predict the prognosis of ESCdys into ESCC. Somatic CDKN2A copy number deletion is a frequent event in both ESCdys and ESCC (Song et al. 2014; Liu et al. 2017; Cui et al. 2020). In the present study, we characterized a 5.1-kb CDR sequence at base resolution within the CDKN2A gene in various human cancers, and confirmed this finding using WGS datasets for all of 18 gastric cancers (Xing et al. 2019). We also established a convenient quantitative PCR assay, P16-Light, to detect CDKN2A SCNVs, and found that both somatic CDKN2A copy number deletion and amplification were prevalent in ESCdys lesions. We also discovered that CDKN2A SCNVs were significantly and consistently associated with the prognosis of ESCdys years before regression and progression. Using CDKN2A SCNVs and other 3 risk factors to construct a monogram to predict regression of ESCdys within 3 follow-up years, the AUC was 72.4% (95% confidence interval: 64.0%-80.7%).
It is well known that amplifications of oncogenes drive cancer development and progression. Here, we found that the amplification of the tumor suppressor gene CDKN2A is positively associated with the regression of ESCdys. To best of our knowledge, this is the first report that the amplification of a tumor suppressor gene can decrease cancer risk in patients with precancer. Also, the fact that CDKN2A deletion is associated with progression/persistence of ESCdys suggests that CDKN2A inactivation may play a crucial role in the initiation stage during malignant transformation of normal squamous cells in the esophageal mucosa and maintenance of ESCdys. Among subjects under a normal environment, two CDKN2A alleles are functionally enough for diploid cells to physiologically maintain. However, among patients with ESCdys in high cancer risk areas, it is unknown whether two CDKN2A alleles are functionally enough because environmental factors may cause epigenetic inactivation of the CDKN2A/P16INK4a gene. CDKN2A amplification is consistently associated with the upregulation of gene transcription in cancer cells. It is worth studying whether CDKN2A amplification may favor the recovery of function of this gene in precancer cells, and subsequently contribute to regression of ESCdys lesions.
The driver function of the CDKN2A gene in cancer development is enigmatic. P16ink4a inactivation contributes less than P19arf (the murine counterpart of human P14ARF) inactivation to cancer development in mice while P16INK4a inactivation contributes more than P14ARF inactivation to cancer development in humans (Peter et al. 2008; Li et al. 2009). The exact mechanisms leading to the difference among species is not clear. Here, we reported that aproximately 87% of genetic P16INK4a inactivation is accompanied by P14ARF inactivation in human cancer cell lines or tissues. This may account for the species-related functional difference of the CDKN2A gene. This explanation is also supported by the report that knocking out both p16ink4a and p19arf leads to more cancer development than individual inactivation in mice (Sharpless et al. 2004).
In conclusion, we have, for the first time, found that there is a 5.1-kb CDR region within the CDKN2A gene, and that most CDKN2A deletions lead to P16INK4a and P14ARF coinactivation in human cancers. Using the CDR as a target sequence, we developed a convenient quantitative multiplex PCR assay, the P16-Light to detect CDKN2A SCNVs in clinical practice. Both CDKN2A deletion and amplification were prevalent in ESCdys lesions and were significantly associated with progression or persistence and regression, respectively, of mild or moderate ESCdys in a population-based follow-up study. CDKN2A SCNVs are potential biomarkers for predicting the regression of precancer in esophageal squamous cells. In addition, we observed a similar CDR region within other tumor suppressor genes such as ATM, FAT1, miR31HG, PTEN, and RB1 in the SNP array-based TCGA SCNV datasets (Supplemental Figure S5), suggesting that our strategy to detect CDKN2A SCNVs may be suitable for the establishment of detection methods for other genes.
METHODS
Specimens and DNA preparation
Since 1999, the National Cancer Center of the Chinese Academy of Medical Science has conducted a multicentre prospective ESCC endoscopic screening project among 22696 residents aged 40-69 yrs in Cixian County (Hebei Province), Yanting County (Sichuan Province), Linzhou City (Henan Province) and Yangzhong City (Jiangsu Province) and Feicheng City (Shandong Province), China, rural areas in which residents have a high risk of ESCC (Project Registration no. NCT02094105, ChiCTR-EOC-17010553). In baseline exams, mild and moderate (m/M) esophageal squamous cell dysplasia (ESCdys) was pathologically diagnosed in 3612 patients. From 2010 to 2015, a repeat endoscopic screening was offered to these eligible subjects to re-evaluate the status of baseline ESCdys lesions and to look for the development of new lesions, and 2147 (59.4%) underwent these repeat endoscope exams. As in the baseline exams, endoscopic biopsy specimens were taken from all lesions visible in the esophagus after mucosal iodine staining. Peripheral blood samples were also taken from each patient before this repeat endoscopic examination. The biopsies were fixed immediately in buffered formalin and then embedded, sectioned, and stained with hematoxylin and eosin. The microscopic slides were read by a panel of three senior pathologists as we previously described (Wei et al. 2015). The presence or absence of ESCC, ESCdys (mild, moderate, severe), inflammation (mild, moderate, severe), and other pathological changes were recorded for each biopsy, and a global diagnosis was made in each case representing the most advanced lesion.
Subjects (n=74) with a diagnosis of m/M ESCdys at the baseline who progressed to severe ESCdys or ESCC during the follow-up were included in the progression group. Subjects remaining as mild or moderate ESCdys were included in the persistent group. Subjects who regressed to inflammation or normal mucosa were included in the regression group (Supplemental Figure S6). m/M ESCdys samples in the persistent group (n=95) and the regression group (n=94) were selected from the tissue block archive, matched to the progressing subjects by age, sex, and village/county/city.
All baseline biopsy samples of mild or moderate ESCdys in the progression group were used, if sections were available from archived paraffin blocks, for P16-Light analysis. Baseline m/M ESCdys biopsy samples were collected from 263 patients and used for P16-Light analysis. Finally, information on the status of CDKN2A SCNVs was obtained for 205 patients with sufficient DNA extracted from their baseline biopsies to complete the P16-Light analysis (Supplemental Figure S3). The peripheral blood samples were selected for all these cases from the National Cohort of Esophageal Cancer (NCEC) biobank. Information including demographic characteristics and risk factors by questionnaire were also collected for each case. Genomic DNA was extracted from these samples with a phenol/chloroform method in a standard fashion. These studies were approved by the Institutional Review Board of the China Cancer Foundation and the National Cancer Center, Cancer Hospital of the Chinese Academy of Medical Sciences and Peking Union Medical College (Approval No. 16-171/1250), and all of the patients provided written informed consent.
Optimized quantitative multiplex PCR assay (P16-Light) to detect CDKN2A copy number
A number of multiplex primer and probe combinations were designed based on the best multiplex primer probe scores for the CDR in the CDKN2A gene and GAPDH sequences by Bacon Designer 8 software. Multiplex PCR assays were established according to the Applied Biosystems (ABI) TaqMan universal PCR master mix manual. The performances of these assays for the detection of CDKN2A copy numbers were compared with each other. Finally, a multiplex primer and probe combination was selected (Supplemental Table S5) and their components’ concentrations were optimized. Briefly, each multiplex PCR assay was carried out in a total 20 μL volume that included 5-10 ng of input DNA, 10 μM of forward and reverse primers and probe for CDKN2A intron-2, 10 μM forward and reverse primers and probe for GAPDH, and 10 μL of 2 x TaqMan Universal Master Mix II with uracil-N-glycosylase (Kit-4440038, ABI, Lithuania). The PCRs were performed in triplicate in a MicroAmp Fast Optical 96-Well Reaction Plate with a barcode (0.1 mL; ABI, China) with an ABI 7500 Fast Real-Time PCR System. The specific conditions of the PCR were as follows: initial incubation for 10 min at 95°C, followed by 40 cycles of 95°C for 20 sec and 58°C for 60 sec. When Ct value for GAPDH input for a sample was 34 or less cycles, this sample was considered as CDKN2A SCNV informative.
Definitions of CDKN2A CDR deletion-positive and amplification-positive
We used the genomic DNA from A549 cells containing no CDKN2A allele to dilute genomic DNA from RKO cells containing 2 wild-type CDKN2A alleles, and then we set the standard curve according to the relative copy number of the CDKN2A gene at different dilution concentrations. The ΔCt value and relative copy number for the CDKN2A gene were calculated using the GAPDH as the reference. When the CDKN2A copy number in the A549-diluted template was consistently lower than the copy number in the RKO control template, and the difference was statistically significant (p<0.05), it was judged that the lowest dilution concentration was the detection limit of CDKN2A deletion (the difference in CDKN2A copy number between the 100% RKO template and 80% RKO template spiked with 20% A549 DNA). When the CDKN2A relative copy number in a tissue sample was significantly lower or higher than that of the paired blood sample, the sample was defined as somatic CDKN2A CDR deletion-positive or amplification-positive, respectively. The 100% A549, 100% RKO, and 20% A549 + 80% RKO DNA mix controls were analyzed for each experiment.
Call for CDKN2A interstitial deletion in the GC WGS datasets
We used Meerkat (http://compbio.med.harvard.edu/Meerkat/) 23 to predict somatic SVs and their breakpoints in WGS datasets (accession numbers, EGAD00001004811 with 36× of sequencing depth) for gastric adenocarcinoma samples from 168 patients using the suggested parameters (Xing et al. 2020). This method used soft-clipped and split reads to identify candidate breakpoints, and precise breakpoints were refined by local alignments. CDKN2A deletion information was obtained from WGS datasets for 157 GC samples.
Cell lines and cultures
The human cell line HEK293T (kindly provided by Professor Yasuhito Yuasa of Tokyo Medical and Dental University) and the P16 allele homogygously deleted cell line A549 (kindly provided by Dr. Zhiqian Zhang of Peking University Cancer Hospital and Institute) were grown in RPMI-1640 medium, and the RKO cell line containing two wild type P16 alleles was purchased from American Type Culture Collection and grown in DMEM media. The medium was supplemented with 10% (v/v) fetal bovine serum (FBS). These cell lines were tested and authenticated by Beijing JianLian Genes Technology Co., Ltd. before they were used in this study. STR patterns were analyzed using a Goldeneye™ 20A STR Identifiler PCR Amplification kit.
Cell proliferation, migration, and invasion assays using IncuCyte
Cells were seeded into 96-well plates (2,000 cells/well, 10 wells/group) and cultured for at least 96 hr to determine the proliferation curves. The cells were photographed every 6 hr using a long-term dynamic observation platform (IncuCyte, Essen, MI, USA). The cell confluence was analyzed using IncuCyte ZOOM software (Essen, Ann Arbor, MI, USA). For continuous observation of cell migration and invasion, the cells were seeded into 96-well plates at a density of 25,000 cells/well and then were cultured for 24 hr. After a wound scratch was established, the cells were washed three times with PBS. For the invasion test, after PBS washed, 50 µL Matrigel (BD Bioscience, San Jose, CA) diluted with RPMI 1640 Medium at ratio of 1:8 was added, and the cells were cultured for 30 min at 37 °C. The cells were then regularly cultured and photographed every 6 hr for at least 96 hr. The relative wound width was calculated using the same software.
Knockout of CDKN2A exon-1a, exon-1b, and exon-2 with CRISPR/Cas9
A single gRNA approach was used to knock out target sequences in the CDKN2A gene using the CRISPR/Cas9 system. The sgRNAs were designed using online software found at the website (http://crispr.mit.edu), and they were synthesized by Thermo Scientific, Inc., Rockford, IL, USA (Supplemental Table S5). The sgRNAs were cloned into the Lenti-CRISPR-V2 vector expressing Cas9 (Plasmid #52961, Addgene, Inc.) at the BsmBI restriction site. Then, the lentiviral plasmid expressing gRNA and Cas9 was introduced into HEK293FT cells. The viral supernatants were collected and filtered through a 0.45 μm PVDF filter (Millipore, USA) 72 hr after transfection, and then the viruses were used to infect HEK293T cells. Three days later, the infected cells were subjected to puromycin selection for one week, and genomic DNA from the surviving cells was isolated for PCR amplification and sequencing using the primers listed in Table S5. Then, the cells were seeded into 96-well plates to select the monoclonal cells. Cells transfected with an empty Lenti-CRISPR-V2 control vector were used as a wild-type (WT) control.
Western blot
Cells were collected and lysed to obtain protein lysate. The resulting proteins were electrophoresed through a 10% SDS-PAGE gel and then were transferred onto a PVDF membrane. After blocking with 5% fat-free milk overnight at 4 °C, the membrane was incubated with primary antibodies (anti-P16, Abcam, ab81278 UK; anti-P14, Abcam, ab185620 UK; anti-RB1, Abcam, ab181616 UK; anti-Phospho-RB1 (Ser807/811), Cell Signaling Technology, #8516, USA; anti-GAPDH, Protein Tech, 60004-1, China) for 1 hr at room temperature. The membrane was then washed 3 times with PBST (PBS with 0.1% Tween 20). After washing, the membrane was incubated with the corresponding horseradish peroxidase-conjugated goat anti-goat or anti-mouse IgG at room temperature for 1 hr. The signals were visualized using an Immobilon Western Chemiluminescent HRP Substrate kit (WBKLS0500, Millipore, Billerica, USA).
Cell apoptosis and death analyses
Cells were seeded in six-well plates (2 × 105/well). After 48 hr, the cells were treated with trypsin and washed twice with cold PBS. They were labeled with annexin V-FITC and propidium iodide (PI) according to the manufacturer’s protocol (Dojindo, Japan). Then, the cells were analyzed with a BD Accuri C6 flow cytometer (BD Biosciences, USA). The percentages of cells in early apoptosis (annexin V-positive, PI-negative) and late apoptosis or necrosis (annexin V- and PI-positive) were calculated using BD Accuri C6 Software.
Statistical analysis
Chi-square or Fisher’s exact tests were used to compare the proportion of somatic CDKN2A CDR deletion or amplification between different groups of tissue samples. Student t-test was used to compare differences of the density or confluence of cells with different CDKN2A genotypes and the proportion of CDKN2A gene copy number between genomic DNA samples. Univariate competing risk analysis was used to identify risk factors for regression of ESCdys, and cumulative regression rate was calculated using the cumulative incidence function method with progression of ESCdys as a competing risk. Factors with p<0.05 in the univariate analysis were incorporated into the multivariate analysis, and used to construct a competing-risk nomogram. The discrimination of the nomogram was evaluated by AUC of the ROC curve. The calibration, which compares estimations with actual observations, was graphically assessed with a calibration curve. All statistical tests were two-sided, and p value less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS software (version 16.0) and R software (version 4.0.5).
Finding
This work was supported by the Beijing Natural Science Foundation (grant number 7181002 to D.J.D.); Capital’s Funds for Health Improvement and Research (grant number 2018-1-1021 to D.J.D.); and by the National Key R&D Program of China (grant number 2016YFC0901404 to W.Q.W.).
Data Availability
Detailed information on the base-resolution interstitial common deletion region in CDKN2A gene in 110 cancer samples was listed in Supplemental Table S1. Deidentified individual participant data that underlie the reported results was attached in the manuscript as Supplemental Table S2. Individual participant data will not be shared.
Competing Insterest statement
The authors have nothing to disclose.
Author contributions
D.J.D. and W.Q.W. designed the study. D.J.D. performed bioinformatics analyses, characterized base resolution CDKN2A CDR coordinates, and wrote the manuscript draft. W.Q.W, Z.Y.F., and Y.Q. organized the follow-up study. J.Z., Y.T., Z.J.L., and L.K.G. developed the P16-Light assay and analyzed CDKN2A SCNVs in samples. R.X., Y.Y.L., J.F.J., provided the CDKN2A CDR coordinates from whole genome sequencing datasets. J.L.Q. performed gene KO experiments. S.M.D., critically reviewed the manuscript. All authors approved the final manuscript.
Data access
Detailed information on the base-resolution interstitial common deletion region in CDKN2A gene in 110 cancer samples was listed in Supplemental Table S1. Deidentified individual participant data that underlie the reported results was attached in the manuscript as Supplemental Table S2. Individual participant data will not be shared.
SUPPLEMENTARY FIGURE LIST
Supplemental Figure S1. Prevalence of CDKN2A deep-deletion and the levels of gene expression in 10967 samples from cancer patients in Pan-TCGA studies. (A) Prevalence of CDKN2A deep deletion detected by Affymetrix SNP6.0 microarray. The number of total cancer cases and cases with CDKN2A deep deletion are listed for each kind of cancer. (B) The levels of P16INK4a mRNA determined by RNA sequencing in cancers with various CDKN2A genetic changes. The charts for patients (n=10953) in 32 Pan-TCGA studies were adapted from a graphic view at the cBioPortal Cancer Genomics website (www.cbioportal.org).
Supplemental Figure S2. Distribution pattern of the Online Mendelian Inheritance in Man (OMIM) allelic variants within the CDKN2A common deletion region (CDR, highlighted in blue shadow). 12 allelic variants are located in CDKN2A exon-2, 2 allelic variants are located in CDKN2A exon-1α, and 1 allelic variant is located in CDKN2A exon-1β. This chart was adapted from the UCSC Genome Browser on March 10, 2021.
Supplemental Figure S3. The prevalence of CDKN2A SCNVs in baseline follow-up esophageal mucosal biopsy samples from patients with baseline mild or moderate esophageal squamous cell dysplasia (ESCdys) and different follow-up experiences, including progression to severe ESCdys or ESCC (orange color), persistence as mild or moderate ESCdys (yellow color), or regression to inflammation or normal (green color). Deletion (blue), somatic CDKN2A deletion; Amplification (red), somatic CDKN2A amplification; Diploid (grey), no SCNV; not informative (white), the amount of genetic DNA from biopsies was not enough for P16-Light analysis
Supplemental Figure S4.The proportion of various CDKN2A SCNVs in the followup esophageal mucosa biopsy samples from baseline ESCdys patients with and without CDKN2A SCNVs. No statistically significant difference of the proportion of CDKN2A SCNVs was observed between the baseline Diploid group and baseline Amplification group. The exact number of samples in each subgroup is labeled. The chi-square values and p-values are also listed between groups.
Supplemental Figure S5. Approximate locations of the estimated common deletion fragment within tumor suppressor genes ATM, FAT1, RB1, PTEN, and miR31HG in TCGA pan-cancers (by Affymetrix SNP6.0 microarray).
Supplemental Figure S6. Images of pathological lesions in esophageal mucosa biopsied from two representative patients with mild ESCdys at baseline and during the followup in the regressive, persistent, and progressive groups. H.E. staining; X100
SUPPLEMENTARY TABLE LIST
Supplemental Table S1. Base-resolution coordinates for CDKN2A deletions in 110 cancer samples
Supplemental Table S2. Clinicalpathological characteristics of all 205 patients with mild or moderate ESCdys at baseline and the status of CDKN2A SCNV in baseline ESCdys
Supplemental Table S3. Comparison of the prevalence of CDKN2A SCNVs between patients with various environmental and personal factors
Supplemental Table S4. Comparison of the prevalence of somatic CDKN2A gene copy number variations (SCNVs) in baseline esophageal squamous cell dysplasia biopsy samples from patients in the progression group and the persistent group in the multicenter followup study
Supplemental Table S5. Oligo sequences
Acknowledgements
We sincerely thank Dr. Guohui Song from Cixian County Cancer Institute, Dr. Changqing Hao from Linzhou Cancer Hospital, Dr. Zhaolai Hua from Yanzhong Cancer Hospital, Dr. Jun Li from Yanting Cancer Hospital, and Dr. Yanyan Li from Feicheng People’s Hospital. Thanks to all cooperating demonstration centers and their staff whose hard work in follow-up made this study possible. We also than Miss Gina Mckeown in New York, USA for English language editing.
REFERENCES
