Significance of Chr9p22.1-p21.3 Deletion in Cancer Development: A Pan-cancer In Silico Analysis

PAOLA G. GONÇALVES1,2, RUI M. REIS1,3,4 and LUCAS T. BIDINOTTO1,2,5

1 Molecular Oncology Research Center, Barretos Cancer Hospital, Barretos, SP, Brazil;

2 Department of Pathology, Botucatu Medical School, São Paulo State University (UNESP), Botucatu, SP, Brazil;

3Life and Health Sciences Research Institute (ICVS), Medical School, University of Minho, Braga, Portugal;

43ICVS/3B’s-PT Government Associate Laboratory, Guimarães, Portugal;

5 Barretos School of Health Sciences, Dr Paulo Prata - FACISB, Barretos, SP, Brazil

Abstract. Background/Aim: Deletions in chr9p22.1-21.3 locus have been related to the development of several types of cancer, mainly due to the presence of CDKN2A and CDKN2B genes. However, there are several other genes in the region with potential importance in tumorigenesis. We, therefore, aimed to analyze in silico the potential prognostic significance of alterations in copy number and expression of genes present in the chr9p22.1-21.3 locus in 33 TCGA datasets (approximately 10,000 patients). Materials and Methods: We analyzed which of the 27 genes are expressed in the datasets. Additionally, we associated the deletion of the locus with survival (log rank analysis) and hazard ratio (HR) (univariate cox regression). Finally, we performed univariate, multivariate, and overall survival analyses in 13 datasets considering the expression of 10 genes present in the locus. Results: We identified 10 genes of the chr9p22.1- 21.3 locus expressed in the datasets (MLLT3, FOCAD, PTPLAD2, KLHL9, IFNE, MTAP, CDKN2A, CDKN2B, DMRTA1 and ELAVL2). Moreover, we found that deletion in at least 1 of these genes was associated with poor survival and increased HR in 13 datasets: adrenocortical carcinoma (ACC), glioblastoma (GBM), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), low- grade glioma (LGG), lung adenocarcinoma (LUAD), mesothelioma (MESO), pancreatic adenocarcinoma (PAAD),

Correspondence to: Dr. Lucas Tadeu Bidinotto, Molecular Oncology Research Center, Barretos Cancer Hospital, Rua Antenor Duarte Villela, 1331, CEP 14784 400, Barretos, Sao Paulo, Brazil. Tel/Fax: +55 1733216600 ext: 7350, e-mail: lucasbidinotto@gmail.com

Key Words: In silico analysis, TCGA, prognostic value, gene expression, 9p, 9p22.1 locus.

prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), sarcoma (SARC) and uterine corpus endometrial carcinoma (UCEC). Finally, we found an association of survival/HR and altered expression of MLLT3 in the MESO dataset, of FOCAD in the READ dataset, of PTPLAD2 in the KIRP dataset, of KLHL9 in the LGG and UCEC datasets, of IFNE in ACC, GBM, KIRC and LUAD datasets, of MTAP in LGG, LUAD and MESO datasets, of CDKN2A in the HNSC, KIRC and MESO datasets, of CDKN2B in the LGG and READ datasets, of DMRTA1 in SARC datasets and of ELAVL2 in the LGG dataset (p<0.01 for all associations). Conclusion: Besides CDKN2A and CDKN2B, numerous other genes are possibly related to cancer development, requiring further investigation.

Cancer is the second leading cause of death-related worldwide. According to the World Health Organization (WHO) 10 million people died of cancer in 2020 (1), while only in the USA 609,000 people are likely to die of cancer in 2022 (2). Its high mortality is due to several factors, including diagnosis in late stages, which was aggravated by the COVID-19 pandemic, and lack of effective treatments.

Treatment of cancer depends on tumor type, stage, and molecular profile, although the standard treatment for cancer is usually surgery, followed by chemotherapy and radiotherapy (3). Chemotherapy uses different types of agents to target cancer cells and induce cell death, however there is still a lack of chemotherapeutic treatments targeting specific molecules approved by the FDA (4). Therefore, there is a crucial need to identify molecular markers for cancer to be able to improve personalized treatment and diagnosis. In addition, being able to understand the underlying biology of a tumor can provide a new target therapy for that tumor.

The progression of cancer involves several steps related to genetic changes in multiple genes or chromosomes. One of the

common genetic alterations observed in cancer is homozygous deletions of recessive cancer genes and fragile sites (5). Our research group found that about 50% of glioblastoma patients have deletion of the 22.1-21.3 region of the short arm of chromosome 9 (6). This region has been associated with the development of several types of cancers. The deletion of genes on chromosome 9p is described as an early event in the development of cancers, and the frequency of loss is similar in both non-invasive and invasive tumors, which can indicate the presence of important tumor suppressor genes in the locus (7-35). This region harbors two important tumor suppressors: CDKN2A (p16) and CDKN2B (p15) that have been widely studied (7, 14, 25, 26, 28, 31, 33, 35-38). CDKN2A is known to be a tumor suppressor gene, which inhibits the formation of complexes CDK4 and CDK6, inducing cell-cycle arrest in G1 and G2 phases (35). Activation of CDK kinases is also prevented by p15 which is a cell growth regulator that controls cell-cycle G1 progression. Genetic and epigenetic changes in CKDN2A and CDKN2B have been described in the development of cancer, metastasis, recurrence, and poor prognosis in several types of tumors (35, 38). However, chr9p22.1-21.3 region harbors 25 more genes that are less studied and may have a role in cancer development. The Cancer Genome Atlas (TCGA) is an outstanding initiative of the National Institute of Health (NIH) aiming to describe the main genetic changes found in several types of cancers (39).

Therefore, the aim of the study was to identify the role of the genes present in the locus chr9p22.1-21.3 in 33 types of cancers from the TCGA, using in silico tools. Moreover, we intended to associate these data with clinicopathological features and describe potential new driver genes with clinical impact in carcinogenesis.

Materials and Methods

This study was approved by the Barretos Cancer Hospital Ethics Committee (protocol number 1394/2017) and all methods were performed according to relevant guidelines and regulations. The results published here are based on data generated by the TCGA Research Network: https://www.cancer.gov/tcga.

Data import and selection of potentially important genes. The overall design of in silico analyses is depicted on Figure 1. In order to test which genes of chr9p22.1-p21.3 locus are expressed in cancer, normalized RSEM (RNA-Seq by Expectation- Maximization) data of RNA sequencing of 27 genes in the chr9p22.1-p21.3 locus (MLLT3, FOCAD, SNORA30, PTPLAD2, IFNB1, IFNW1, IFNA21, IFNA4, IFNA10, IFNA16, IFNA7, IFNA17, IFNA14, IFNA5, KLHL9, IFNA6, IFNA13, IFNA2, IFNA8, IFNA1, IFNE, MTAP, C9orf53, CDKN2A, CDKN2B, DMRTA1, ELAVL2) were imported from 33 different TCGA datasets (9,659 samples) (Table I). These values were transformed to a logarithmic scale (log10+1). In order to select genes with potential impact on the datasets, the expression of genes present in the chr9p22.1-p21.3 locus across all patients was analyzed in the 33 datasets.

Figure 1. Experimental design depicting the in silico experiments performed using data from The Cancer Genome Atlas (TCGA).

27 genes 33 datasets

Genes expressed in the datasets

10 genes 33 datasets

Deletion No deletion

Copy number alterations

Univariate cox analysis Overall survival

10 genes 13 datasets

High expression Normal expression Low expression

RNA sequencing

Univariate cox analysis Multivariate cox analysis Overall survival

Copy number alteration analysis. Firstly, it was determined if the gene was expressed on the datasets or not. Then, the genes most prominently expressed in the datasets were selected. The 10 selected genes were MLLT3, FOCAD, PTPLAD2, KLHL9, IFNE, MTAP, CDKN2A, CDKN2B, DMRTA1 and ELAVL2. The GISTIC2 results from CGH Custom Microarray 2x415K data from TCGA were obtained to analyze the copy number variation in the chr9p22.1-p21.3 locus in 10,815 samples across the 33 datasets. The patients were distributed in 2 groups, depending on the presence of deletion in chr9p22.1-p21.3 locus. If there was at least 1 gene with homozygous deletion in the region, the patient was allocated in the “Deleted” group. Otherwise, the patient was considered to have no deletion. In order to determine the prognostic impact of the deletion in chr9p22.1- p21.3 locus among the cancer types, a univariate cox analysis was performed. The datasets presenting potential prognostic significance were considered for gene expression analyses. Then, overall survival analysis was performed in adrenocortical carcinoma (ACC), glioblastoma (GBM), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), low-grade glioma (LGG), lung adenocarcinoma (LUAD), mesothelioma (MESO), pancreatic adenocarcinoma (PAAD), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), sarcoma (SARC) and uterine corpus endometrial carcinoma (UCEC) datasets. Finally, the individual percentage of deletion of each gene of the locus was determined.

Gene expression analysis. The 13 datasets with potential prognostic significance in copy number analysis were selected for gene expression studies. The Z-score was calculated for all genes obtained of RNASeq analysis. Gene expression was considered High if Z-score ≥1.5 or Low if Z-score ≤-1.5. Otherwise, gene

Table I. Datasets from the The Cancer Genome Atlas (TCGA) showing the type of cancer and number of samples.
Cancer typeDatasetNumber of patients in aCGHNumber of patients in RNA sequencing
Adrenocortical carcinomaACC9079
Bladder urothelial carcinomaBLCA408408
Breast invasive carcinomaBRCA1,0791,093
Cervical squamous cell carcinoma and endocervical adenocarcinomaCESC295304
CholangiocarcinomaCHOL3636
Colon adenocarcinomaCOAD449298
Lymphoid neoplasm diffuse large B-cell lymphomaDLBC4848
Esophageal carcinomaESCA184184
GlioblastomaGBM573159
Head and neck squamous cell carcinomaHNSC522521
Kidney chromophobeKICH6666
Kidney renal clear cell carcinomaKIRC528533
Kidney renal papillary cell carcinomaKIRP288290
Acute myeloid leukemiaLAML191173
Brain lower grade gliomaLGG512515
Liver hepatocellular carcinomaLIHC370371
Lung adenocarcinomaLUAD516516
Lung squamous cell carcinomaLUSC501501
MesotheliomaMESO8787
Ovarian serous adenocarcinomaOV573304
Pancreatic adenocarcinomaPAAD184178
Pheochromocytoma and paragangliomaPCPG162179
Prostate adenocarcinomaPRAD492497
Rectum adenocarcinomaREAD16598
SarcomaSARC257259
Skin cutaneous melanomaSKCM367469
Stomach adenocarcinomaSTAD441418
Testicular germ cell tumorsTGCT134134
Thyroid carcinomaTHCA499501
ThymomaTHYM123120
Uterine corpus endometrial carcinomaUCEC539183
Uterine carcinosarcomaUCS5657
Uveal melanomaUVM8080

aCGH: Arrays comparative genomic hybridization.

expression was considered normal. Univariate cox analysis was performed in each gene and dataset. Those presenting p≤0.1 were selected for multivariate cox analysis per dataset. Finally, the genes/datasets with statistical significance after multivariate analysis were submitted to overall survival analysis.

All analyses were performed through the R v.4.0.2 software, using the packages RTCGAToolbox and TCGAbiolinks (40, 41). HR and 95% confidence interval (95%CI) were generated using Uni- and Multivariate cox analysis and survival analysis was performed constructing Kaplan-Meier curves using the Survminer package. The group contrasts were considered statistically significant when p<0.05.

Results

Ten genes of the chr9p22.1-p21.3 locus were found expressed among 33 TCGA datasets. In order to select genes, that are expressed in the studied tumors, the expression of each of the 27 genes present in the locus was analyzed across all datasets. Overall, among the genes analyzed, 10 genes were found

expressed across all datasets (MLLT3, FOCAD, PTPLAD2, KLHL9, IFNE, MTAP, CDKN2A, CDKN2B, DMRTA1 and ELAVL2) (Figure 2). The remaining genes presented number of reads close to 0 in most patients and datasets; therefore, they were excluded from subsequent analyses.

Deletion in chr9p22.1-p21.3 locus is a prognostic factor in 13 TCGA datasets. Once expressed genes present in all datasets were selected, we analyzed the frequency of homozygous deletion of the chr9p22.1-p21.3 region across the 33 datasets. We found 7 datasets with more than 30% of patients presenting deletion in the chr9p22.1-p21.3 locus: BLCA (33.3%), DLBC (31.3%), ESCA (41.8%), GBM (58.6%), HNSC (32.2%), MESO (47.1%) and SKCM (33.2%) (Figure 3).

Univariate cox analysis of copy number results points to a potential prognostic significance of chr9p22.1-p21.3 deletion in 13 datasets (ACC, GBM, HNSC, KIRC, KIRP,

Figure 2. Expression of the 27 genes located in chr9p22.1-21.3 locus using RNA sequencing data of 9,659 cancer patients from The Cancer Genome Atlas (TCGA).

4

3

Expression

N

..

..

1

..

-

..

..

0

MLLT3

KIAA1797

SNORA30

PTPLAD2

IFNB1

IFNW1

IFNA21

IFNA4

IFNA10

IFNA16

IFNA7

IFNA17

IFNA14

IFNA5

KLHL9

IFNA6

IFNA13

IFNA2

IFNA8

IFNA1

IFNE

MTAP

C9orf53

CDKN2A

CDKN2B

DMRTA1

ELAVL2

Gene

LGG, LUAD, MESO, PAAD, PRAD, READ, SARC and UCEC). Interestingly, LGG results showed a 7.93-higher chance of death in the patients presenting the deletion (p<0.0001, Figure 4).

Overall survival analysis shows the deletion associated with poorer prognosis in ACC (p=0.012), GBM (p=0.006), HNSC (p=0.0021), KIRC (p<0.0001), KIRP (p<0.0001), LGG (p<0.0001), LUAD (p=0.016), MESO (p<0.0001), PAAD (p=0.094), PRAD (p=0.0094), READ (p=0.02), SARC (p=0.0021) and UCEC (p=0.027) (Figure 5).

We further evaluated which genes present in the locus are most frequently deleted in the region. Except for UCEC, in most patients of the other datasets, the deletion in the region harbors mainly MTAP, CDKN2A, and CDKN2B genes (Figure 6).

Gene expression analysis shows potential prognostic significance of chr9p22.1-p21.3 gene expression mainly in LGG and MESO. Uni- and multivariate cox analyses point to

potential prognostic significance of 10 genes across 11 datasets, mainly low-grade gliomas (4 genes), and mesotheliomas (3 genes). Low expression of MLLT3, FOCAD, and PTPLAD2 is associated with decreased survival in MESO, READ, and KIRP patients, respectively. Multivariate analysis showed that high KLHL9 expression is associated with decreased survival in LGG and UCEC. High IFNE expression is associated with decreased survival in GBM, KIRC, and LUAD. Low MTAP expression is associated with decreased survival in LUAD, whereas high MTAP expression is associated with decreased survival in LGG, LUAD, and MESO. High CDKN2A expression is associated with decreased survival in KIRC and MESO, whereas low expression is associated with increased survival in HNSC. High expression of CDKN2B is associated with decreased survival in LGG and READ. Low expression of DMRTA1 is associated with decreased survival in SARC. Finally, high expression of ELAVL2 is associated with increased survival in LGG.

Figure 3. Percentage of deletion of chr9p22.1-21.3 locus in the 33 The Cancer Genome Atlas (TCGA) datasets. A patient was considered as having the deletion if there weas at least one gene deleted in the locus.

1.0

☐ No deletion

☒ Deletion

0.8

0.6

0.4

0.2

0.0

ACC

BLCA

BRCA

CESC

CHOL

COAD

DLBC

ESCA

GBM

HNSC

KICH

KIRC

KIRP

LAML

LGG

LIHC

LUAD

LUSC

MESO

OV

PAAD

PCPG

PRAD

READ

SARC

SKCM

STAD

TGCT

THCA

THYM

UCEC

UCS

UVM

Accordingly, overall survival analysis showed that 4 genes have potential prognostic importance in LGG (KLHL9, MTAP, CDKN2B and ELAVL2, p<0.0001) whereas 3 genes present potential prognostic importance in MESO (MLLT3, MTAP and CDKN2A, p<0.001). Two genes were associated with survival in READ (FOCAD, CDKN2B, p<0.05), KIRC (IFNE, CDKN2A, p<0.0001) and LUAD (MTAP,IFNE, p=0.05), while in ACC (IFNE, p<0.0001), KIRP (PTPLAD2, p<0.0001), GBM (IFNE, p<0.0001), HNCS (CDKN2A p<0.01), SARC (DMRTA1 p<0.001) and UCEC (KLHL9 p<0.05) only one gene was associated with survival (Figure 7).

Discussion

In the present study, we inquired the deletion and expression of genes present on chr9p22.1-21.3 locus in 33 different types of cancer across TCGA datasets. Overall, we found a potential prognostic significance of chr9p22.1-21.3 deletion in 13 datasets (ACC, GBM, HNSC, KIRC, KIRP, LGG,

LUAD, MESO, PAAD, PRAD, READ, SARC and UCEC) and that differential expression of 10 genes (MLLT3, FOCAD, PTPLAD2, KLHL9, IFNE, MTAP, CDKN2A, CDKN2B, DMRTA1 and ELAVL2) is present in at least one of these datasets. Moreover, we found a potential prognostic significance of genes KLHL9, MTAP, MLLT3 and ELAVL2 in cancer, mainly in LGG and MESO.

Deletion of chr9p22.1-21.3 is associated with poor survival in 13 datasets. Our results demonstrated that deletion of chr9p22.1-21.3 was associated with a poor prognosis in 13 datasets. To the best of our knowledge, loss of heterozygosity (LOH) of the locus has been more frequently analyzed (7, 8, 10, 11, 21, 23, 27, 33, 34, 36, 42-44), explaining the differences of frequency between other studies and ours. The loss of chr9p22.1-21.3 region is the most described genetic alteration for HNSCC (32, 34). One study found the LOH of 9p in 72% patients with HNSCC (11). In our analysis, we found 32.2% of patients with homozygous

0.001

0.004

0.016

0.062

0.250

1.000

4.000

Deletion (Deaths)No deletion (Deaths)p-Value
ACC5280.0171
BLCA631170.7448
BRCA91400.5581
CESC1670.5929
CHOL4140.7136
COAD01010.9947
DLBC360.7447
ESCA29470.3244
GBM2821920.0062
HNSC861340.0023
KICH19
KIRC121630
KIRP6382e-04
LAML21250.0996
LGG35910
LIHC111180.26
LUAD441410.0169
LUSC651540.1055
MESO40340
OV173150.1036
PAAD34660.0148
PCPG060.9988
PRAD180.0357
READ1270.0491
SARC20780.0026
SKCM721220.6717
STAD241500.5696
THCA0160.9977
THYM180.219
UCEC3870.0377

Figure 4. Hazard ratio (HR) and p-values of univariate cox analysis of deletion in the chr9p22.1-21.3 locus of The Cancer Genome Atlas (TCGA) datasets.

deletion in the HNSCC dataset, and a poorer survival of these patients. In regards to lung cancer, Panani et al. using fluorescence in situ hybridization (FISH) showed that 73% of small cell carcinomas, 84% of adenocarcinomas and 100% large cell carcinomas had the loss present (25). We found that 18.2% of lung adenocarcinoma patients with homozygous deletion in this locus, also presented a poorer survival. Another study that used FISH to detect deletion of the locus was that of Luo et al. (45), where they found a loss of 50% in patients with pancreatic adenocarcinoma. Another study using single cell comparative genomic hybridization (SCOMP) found a loss of the locus in 31% of patients (42), whereas in our study the homozygous deletion frequency was 8.3%. For clear cell renal cell carcinoma (ccRCC), LOH was detected in 44% of ccRCCs patients and the loss was associated with higher stage, larger tumors, necrosis, micro vascular, and renal vein invasion (13). Moreover, a study showed that 88% of ccRCC patients without loss of the chromosome 9p survived 5 years after diagnosis vs. 43% of patients with 9p loss (8, 14). In the present study, we demonstrated that 3.0% of clear cell renal cell carcinoma

patients presented homozygous deletion in the locus, and less than 50% of those survived 50 months after diagnosis, whereas about 70% of patients without deletion survived this period. Considering kidney cancer, we also found 4.5% of kidney renal papillary cell carcinoma patients with homozygous deletion in the locus, with a worse survival compared to those with no deletion. Concerning mesotheliomas, Cheng et al. (1993) described loss of the region in 87% of patients and 83% of cell lines, being 43% deletions (11). We found deletion of the region in 45.1% of the patients. Interestingly, all patients presenting the deletion died before 60 months after diagnosis, whereas about 25.0% of the patients with no deletion survived longer. For sarcomas, our study found deletion in 14.8%, while in the literature a study described 15% copy number alterations in the locus, being 5 homozygous deletions and one hemizygous deletion, and the CNA was associated with poor survival (46). Finally, besides a low frequency, we found homozygous deletion in adrenocortical carcinoma, prostate adenocarcinoma, uterine corpus endometrial carcinoma and rectum adenocarcinoma, and this low frequency has been

Figure 5. Overall survival of patients according to deletion of chr9p22.1-21.3 locus in 13 The Cancer Genome Atlas (TCGA) datasets.

ACC

GBM

HNSC

KIRC

1.00

1.00

1.00

1.00

Overall survival

Overall survival

Overall survival

Overall survival

0.75

0.75

0.75

0.75

0.50

0.50

0.50

0.50

0.25

0.25

0.25

0.25

p = 0.012

p = 0.005

p = 0.0021

P< 0.0001

0.00

0.00

0

50

Months

100

150

0

25

50

75

125

0.00

0.00

Months

100

0

50

100

50

100

150

Number at risk

Number at risk

Number at risk

Months

150

200

0

Months

Deletion 7

3

0

0

Deletion 334

47

12

7

2

1

Deletion 167

23

6

3

0

Number at risk

Deletion 16

1

0

0

No

83

30

6

2

No

237

52

17

7

1

deletion

0

No

353

59

4

1

No

512

197

38

0

deletion

10

deletion

deletion

KIRP

LGG

LUAD

MESO

1.00

1.00

1.00

1.00

Overall survival

Overall survival

Overall survival

Overall survival

0.75

0.75

0.75

0.75

0.50

0.50

0.50

0.50

0.25

0.25

p<0.0001

0.25

p <0.0001

0.25

p = 0.016

P < 0.0001

0.00

0.00

0.00

0.00

0

50

100

Months

150

200

0

50

100

Months

150

200

0

50

100

150

Months

200

250

0

25

50

75

100

Number at risk

Deletion 13

2

0

0

0

Number at risk

Deletion 67

Number at risk

Number at risk

Months

2

0

0

0

Deletion 94

11

1

0

0

0

Deletion 41

5

0

0

0

No

274

68

9

1

deletion

0

No

443

81

21

6

1

No

deletion

413

60

15

6

3

0

No

deletion

44

22

7

3

0

deletion

1.00

PAAD

1.00

PRAD

1.00

READ

1.00

SARC

Overall survival

Overall survival

Overall survival

Overall survival

0.75

0.75

0.75

0.75

0.50

0.50

0.50

0.50

0.25

0.25

0.25

p=0.02

0.25

p = 0.014

p=0.0094

p = 0.0021

0.00

0.00

0.00

0.00

0

25

50

Months

75

100

0

40

80

Months

120

160

0

25

50

75

Months

100

125

0

50

100

150

200

Number at risk

Number at risk

Number at risk

Number at risk

Months

Deletion 52

4

0

0

0

Deletion 9

3

0

0

0

Deletion

1

0

0

0

0

0

Deletion 38

5

1

0

0

No

132

28

10

2

0

No

483

172

32

4

1

No

163

72

12

4

2

0

No

219

67

15

5

0

deletion

UCEC

deletion

deletion

deletion

1.00

Overall survival

0.75

0.50

+

Deletion

+ No deletion

0.25

p = 0.027

0.00

0

50

100

150

200

Number at risk

Months

Deletion 3

1

1

0

0

No

534

145

19

2

1

deletion

related in the literature (47-49). The loss in 0.6% of rectum adenocarcinoma was described for the first time in this study. Importantly, for gliomas, James et al. (1993) found loss of the region in 67% of glioma-derived cell lines and 37% of primary cell lines (18), suggesting that this loss is an early event in the development of gliomas. We found deletion in 13.1% of patients diagnosed with LGG and 58.7% of GBM, suggesting that this deletion may also be correlated to malignancy of this tumor type. Overall, these studies indicate

the presence of potential tumor suppressor genes in this region and, in fact, CDKN2A and CDKN2B are well known tumor suppressors genes (35, 38). Analysis of deletion of individual genes in the locus point to recurrent deletion of mainly CDKN2A and CDKN2B (corroborating the literature) and MTAP genes. However, the other genes presented in the region (MLLT3, FOCAD, PTPLAD2, KLHL9, IFNE, DMRTA1 and ELAVL2) are shown as deleted in at least 20-25% of the samples. Therefore, we analyzed the potential prognostic

Figure 6. Percentage of deletion of each gene present in the chr9p22.1-21.3 locus in 13 The Cancer Genome Atlas (TCGA) datasets.

Deletion

Genes

ACC-

7.8%

14.3%

28.6%

28.6%

28.6%

28.6%

85.7%

100.0%

71.4%

28.6%

28.6%

GBM -

58.6%

22.6%

29.8%

28.3%

45.2%

50.6%

89.6%

97.6%

95.2%

52.1%

34.5%

HNSC-

32.2%

8.9%

12.5%

14.3%

19.0%

20.8%

63.7%

98.2%

86.9%

26.8%

16.7%

KIRC-

3.0%

6.2%

12.5%

6.2%

12.5%

12.5%

50.0%

93.8%

93.8%

25.0%

6.2%

KIRP -

4.5%

7.7%

23.1%

23.1%

38.5%

53.8%

92.3%

100.0%

100.0%

53.8%

30.8%

Percentage of patients with deletion per gene

LGG-

13.1%

28.4%

32.8%

34.3%

50.7%

52.2%

86.6%

95.5%

97.0%

55.2%

40.3%

Tumor type

0.00

LUAD-

18.2%

24.5%

33.0%

35.1%

47.9%

54.3%

80.9%

97.9%

93.6%

66.0%

34.0%

0.25

0.50

MESO-

47.1%

22.0%

22.0%

22.0%

39.0%

46.3%

87.8%

97.6%

95.1%

61.0%

43.9%

0.75

1.00

PAAD -

28.3%

13.5%

26.9%

28.8%

53.8%

59.6%

88.5%

100.0%

96.2%

67.3%

30.8%

PRAD -

1.8%

33.3%

33.3%

33.3%

44.4%

44.4%

88.9%

88.9%

88.9%

55.6%

66.7%

READ-

0.6%

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

100.0%

0.0%

SARC-

14.8%

23.7%

15.8%

15.8%

34.2%

36.8%

78.9%

94.7%

94.7%

47.4%

28.9%

UCEC-

0.6%

66.7%

66.7%

33.3%

33.3%

33.3%

33.3%

33.3%

33.3%

33.3%

33.3%

Deletion -

MLLT3

FOCAD -

PTPLAD2

KLHL9

IFNE -

MTAP -

CDKN2A-

CDKN2B

DMRTA1

ELAVL2-

Genes

significance of the expression of these genes in order to investigate if they can also play a role as tumor suppressors.

Besides CDKN2A and CDKN2B, there are several genes of chr9p22.1-21.3 with potential prognostic value in several tumor types. By using our approach, we found two datasets, in which the differential expression of genes in chr9p22.1-21.3 may be more related to survival and death risk: LGG and MESO. Our data point to these datasets as good candidates for further studies considering the possibility of finding novel tumor suppressor genes in the 9p22.1-21.3 region.

Interestingly, besides being a well-recognized tumor suppressor gene, high expression of CDKN2A was correlated to worse survival in 2 datasets. This behavior has been shown in several studies, mainly studies related to viral infection and cancer development, showing potential of p16 (protein encoded by CDKN2A) to be a surrogate marker of human papillomavirus infection in cancer (50). This gene, implicated in the cyclin D1/retinoblastoma pathway, is shown to be disrupted in the majority of human hepatocellular carcinomas (51-53). Indeed, its methylation has been associated with several parameters of liver hepatocellular carcinoma (LIHC), including virus infection

Figure 7. Overall survival of patients according to expression of 10 genes across 13 The Cancer Genome Atlas (TCGA) datasets.

MLLT3 expression in MESO dataset

FOCAD expression in READ dataset

PTPLAD2 expression in KIRP dataset

KLHL9 expression in LGG dataset

1.00

1.00

1.00

1.00

Overall survival

0.75

p<0.0001

Overall survival

0.75

Overall survival

0.75

Overall survival

0.75

p<0.0001

0.50

0.50

0.50

0.50

0.25

0.25

0.25

p<0.0001

0.25

p=0.042

0.00

0.00

0.00

0.00

0

25

50

Months

75

100

0

25

50

75

100

125

0

50

100

150

Months

200

Months

0

50

100

Months

150

200

Number at risk

Number at risk

Number at risk

Number at risk

MLLT3.cat=Normal 65

26

7

3

0

KIAA1797.cat-Normal

90

47

9

4

2

2

PTPLAD2.cat=Normal 281

68

9

0

MLLT3.cat=Low 20

KIAA1797.cat-High

2

PTPLAD2.cat=High

1

KLHL9.cat=Normal 491

82

21

6

1

1

0

6

0

0

0

0

0

0

0

0

1

0

0

KLHL9.cat=High

KIAA1797.cal=Low 1

5

1

0 O 0

0

0

0

0

0

0

0

PTPLAD2.cal=Low 7

2

0

0

0

KUHL9.cat=Low

17

0

0

A

0

MTAP expression in LGG dataset

MTAP expression in LUAD dataset

MTAP expression in MESO dataset

KLHL9 expression in UCEC dataset

1.00

1.00

1.00

1.00

Overall survival

0.75

p<0.0001

Overall survival

0.75

p<0.0001

Overall survival

0.75

p<0.0001

Overall survival

0.75

0.50

0.50

0.50

0.50

0.25

0.25

0.25

0.25

p=0.011

0.00

0.00

0.00

0.00

0

50

100

Months

150

200

0

50

100

150

200

250

0

25

50

75

100

0

50

100

150

Number at risk

Number at risk

Months

Number at risk

Months

Number at risk

Months

MTAP.cat-Normal 475

83

21

6

1

MTAP.cat=Normal 459

MTAP,cat=High 5

68

16

6

3

0

MTAP.cat=Normal 56

24

7

3

0

KLHL9.cat=Normal 170

23

5

0

MTAP.cat=Low 38

0

0

0

0

0

D

MTAP.cat=Low 43

KLHL9.cal=High

1

0

0

0

0

0

0

3

0

0

0

0

MTAP.cat-Low

29

3

0

0

0

KLHL9.cat=Low

10

0

0

0

IFNE expression in ACC dataset

IFNE expression in GBM dataset

IFNE expression in KIRC dataset

IFNE expression in LUAD dataset

1.00

1.00

1.00

1.00

Overall survival

0.75

Overall survival

0.75

p<0.0001

Overall survival

0.75

Overall survival

0.75

p=0.017

0.50

0.50

0.50

0.50

0.25

0.25

0.25

0.25

p=0.00033

p<0.0001

0.00

0.00

0.00

0.00

0

50

40

Number at risk IFNE.cat=Normal 77

Months

100

150

0

20

60

80

0

50

Months

100

150

0

50

100

150

Months

200

250

Number at risk IFNE.calmNormal 156

Months

Number at risk

Number at risk

28

7

2

29

7

2

1

IFNE.cal=Normal 513

196

39

D

IFNE.cat-Normal 483

69

15

5

2

0

IFNE.cat=High

2

0

0

0

IFNE.cat=High 2

0

0

0

0

IFNE.cat=High

20

5

0

0

IFNE.cat-High 2

2

1

1

1

0

CDKN2A expression in HNSC dataset

CDKN2A expression in KIRC dataset

CDKN2A expression in MESO dataset

DMRTA1 expression in SARC dataset

1.00

1.00

1.00

1.00

Overall survival

0.75

Overall survival

0.75

Overall survival

0.75

p=0.00014

Overall survival

0.75

p=0.00031

0.50

0.50

0.50

0.50

0.25

0.25

p<0.0001

0.25

0.25

0.00

p=0.0015

0.00

0.00

0.00

0

50

100

150

200

0

50

100

150

0

25

50

75

100

0

50

100

150

200

Number at risk

Months

Number at risk

risk

Months

Number at risk

Months

Number at risk

Months

CDKN2A.cat=Normal 394

69

12

5

COKN2A.cat=Normal 462

163

0

CDKN2A.cat=Normal 79

27

DMRITA1.cal=Normal 227

67

5

COKN2A.cat=High 75

11

2

1

1

30

7

3

D

15

0

Q

COKN2A.cat=High 2

CDKN2A.cat=Low 50

0

Q

0

CDKN2A.car-High

2

0

0

DMRTA1.cat-High 1

1

0

0

4

2

1

0

0

CDKN2A.cat=Low

0

0

CDKN2A.cat=Low 69

38

9

0

4

0

0

0

0

DMRTA1.cal=Low 31

5

1

0

0

CDKN2B expression in LGG dataset

CDKN2B expression in READ dataset

ELAVL2 expression in LGG dataset

1.00

1.00

1.00

Overall survival

0.75

p<0.0001

Overall survival

0.75

Overall survival

0.75

p<0.0001

0.50

0.50

0.50

+

Normal

+

High

+

Low

0.25

0.25

0.25

p=0.05

0.00

0.00

0.00

0

50

100

150

200

0

25

50

75

125

0

50

100

150

200

Number at risk

Months

100

Number at risk CDKN2B.cat-Normal 92

Months

Number at risk

Months

CDKN2B.cat-Normal 483

81

21

6

1

46

9

4

2

2

ELAVL2.cat-Normal 184

21

3

0

0

(54), and CDKN2A expression was correlated to a poor prognosis, similarly, in the uterine cervix (55). Furthermore, CDKN2A is part of a six-gene signature in UCEC which is related to poor prognosis (56). Lamperska et al. studied using immunohistochemistry the expression of p16 in uveal melanoma (UVM), and their findings showed upregulation of p16, cyclin D1, cyclin 3, as well as abnormal pRB and E2F binding during the development of human UVM (57, 58). High expression of CDKN2A was associated with better overall survival in HNSC patients, corroborating the literature (59).

Of interest, low MTAP expression was associated with poor survival in 3 datasets (LGG, LUAD and MESO). In gliomas, one study found loss of MTAP expression in 27.8% of diffuse astrocytomas, 50.0% of anaplastic astrocytomas, 45.6% of adult glioblastomas and 54.8% of pediatric glioblastomas, however the loss was not associated with clinicopathological features and survival (60). A bioinformatics study using gliomas TCGA cases found the loss of MTAP of 25.8% in LGG and 60% in glioblastoma and the survival of patients with loss of the MTAP gene was shorter than that in other patients (61). Another study found loss of MTAP in 12.2% of grade I and in 62.5% of grade IV gliomas and the loss was correlated with shorter overall survival (p=0.011) and a shorter progression-free survival (p=0.016) (62). For lung adenocarcinoma, only one study described the loss of MTAP (12.9% of MTAP loss in LUAD), however the study did not associate the loss with survival (63). For MESO, MTAP loss was found in 76.2% of malignant MESO (64) and in 74% in malignant pleural MESO of which 37.5% were heterozygous and 12.5% homozygous deletions (65). In fact, studies for MESO and MTAP expression are not focused in survival, but in expression, since immunohistochemistry for MTAP is a reliable surrogate for CDKN2A fluorescence in situ hybridization for diagnosis of malignant MESO (66), and our results point that MTAP may be a surrogate marker of the locus loss in other cancer types.

We also found MLLT3, DMRTA1, and ELAVL2 as interesting subjects of study in specific datasets, since low expression was associated with poor survival or high expression was associated with good survival. MLLT3 gene, which encodes AF9 protein, is part of an elongation complex, required in increasing the catalytic rate of RNA polymerase II (67). This, in turn, it regulates the control checkpoint elongation, and its loss of function is associated with the development of hematologic cancers, and, in fact, we found that this loss led to a worse prognosis. The function of this gene has been well described in common rearrangements found in leukemia (68), important in the maintenance of intermediate precursor brain cells, preventing their premature cell-cycle exit through epigenetic modifications (69). We found low expression of this gene drastically associated with low survival in mesothelioma patients, pointing to a potential role as a tumor suppressor.

DMRTA1, a gene part of the DMRT family that encodes transcription factors involved in sexual development, controls

testicular development, including the differentiation of germ cells and Sertoli cells (70). Another study showed a loss of DMRTA1 expression in 27% of bladder cancer cell lines (71). Also, a distinctive intergenic insertion between DMRTA1 and LINC01239 had oncogenic effects through activation of the mammalian target of rapamycin (mTOR)/4EBP/S6K pathway in HBV-infected intrahepatic cholangiocarcinomas (72). The ELAVL family are RNA-binding proteins that bind with high specificity and to AU instability factors in the 3’UTR region of mRNAs from genes involved in cell growth (73). ELAVL2 binds with high specificity to the 3’UTR region of c-Myc mRNAs (74), HIF-1a, PTBP1 and VEGF (75). In addition, ELAVL2 binds to the 3’-UTR of the MycN mRNA and increases its stability in neuroblastoma (76). ELAVL2 also regulates several aspects of neuronal function, including neuronal excitability and synaptic transmission, both of which are critical for normal brain function in cognition and behavior (77). In cancer, ELAVL2 methylation was correlated with better rates of progression-free and overall survival in oropharyngeal cancer (78). ELAVL2 overexpression is a risk factor for poor response to chemotherapy in patients with esophageal squamous cell carcinoma (79). Of note, we found that low DMRTA1 expression was associated with poor survival in sarcoma, whereas high ELAVL2 expression was associated with increased survival in LGG.

Conclusion

Besides CDKN2A and CDKN2B, we discovered several other genes in the chr9p22.1-21.3 locus that are possibly associated with cancer development, namely MLLT3, DMRTA1, MTAP, and ELAVL2. Furthermore, these genes may be related to cell cycle, differentiation, and metabolism thus may be important for future studies, especially in gliomas, MESO, and sarcomas. These genes enrich the piece of evidence regarding cancer development and may contribute towards a better understanding of underlying molecular biology of cancer development.

Availability of Data and Materials

The datasets analyzed during the current study are available in The Cancer Genome Atlas repository [https://tcga-data.nci.nih.gov].

Conflicts of Interest

The Authors declare that they have no competing interests.

Authors’ Contributions

PGG contributed to the analysis and interpretation of the data and wrote the first draft. RMR contributed to conceptualization of the project, interpretation of data, writing reviewing and editing. LTB contributed to the analysis and interpretation of data, methodology,

conceptualization of the project, funding acquisition, final writing, reviewing, and editing.

Acknowledgements

This study was funded by the Fundação de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP), grant number 2016/21727-4 for LTB. PGG was funded by the Fundação de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP) fellowship number 2017/09749-5. Both contributed to the design of the study and collection, analysis, and interpretation of data.

References

1 Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A and Bray F: Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 71(3): 209-249, 2021. PMID: 33538338. DOI: 10.3322/caac.21660

2 Siegel RL, Miller KD, Fuchs HE and Jemal A: Cancer statistics, 2022. CA Cancer J Clin 72(1): 7-33, 2022. PMID: 35020204. DOI: 10.3322/caac.21708

3 Baskar R, Lee KA, Yeo R and Yeoh KW: Cancer and radiation therapy: current advances and future directions. Int J Med Sci 9(3): 193-199, 2012. PMID: 22408567. DOI: 10.7150/ijms.3635

4 Pérez-Herrero E and Fernández-Medarde A: Advanced targeted therapies in cancer: Drug nanocarriers, the future of chemotherapy. Eur J Pharm Biopharm 93: 52-79, 2015. PMID: 25813885. DOI: 10.1016/j.ejpb.2015.03.018

5 Cox C, Bignell G, Greenman C, Stabenau A, Warren W, Stephens P, Davies H, Watt S, Teague J, Edkins S, Birney E, Easton DF, Wooster R, Futreal PA and Stratton MR: A survey of homozygous deletions in human cancer genomes. Proc Natl Acad Sci USA 102(12): 4542-4547, 2005. PMID: 15761058. DOI: 10.1073/pnas.0408593102

6 Bidinotto LT, Torrieri R, Mackay A, Almeida GC, Viana- Pereira M, Cruvinel-Carloni A, Spina ML, Campanella NC, Pereira de Menezes W, Clara CA, Becker AP, Jones C and Reis RM: Copy number profiling of brazilian astrocytomas. G3 (Bethesda) 6(7): 1867-1878, 2016. PMID: 27172220. DOI: 10.1534/g3.116.029884

7 Brenner AJ and Aldaz CM: Chromosome 9p allelic loss and p16/CDKN2 in breast cancer and evidence of p16 inactivation in immortal breast epithelial cells. Cancer Res 55(13): 2892- 2895, 1995. PMID: 7796417.

8 Brunelli M, Eccher A, Gobbo S, Ficarra V, Novara G, Cossu- Rocca P, Bonetti F, Menestrina F, Cheng L, Eble JN and Martignoni G: Loss of chromosome 9p is an independent prognostic factor in patients with clear cell renal cell carcinoma. Mod Pathol 21(1): 1-6, 2008. PMID: 17906617. DOI: 10.1038/ modpathol.3800967

9 Cairns P, Shaw ME and Knowles MA: Preliminary mapping of the deleted region of chromosome 9 in bladder cancer. Cancer Res 53(6): 1230-1232, 1993. PMID: 8443801.

10 Chan AS, To KF, Lo KW, Ding M, Li X, Johnson P and Huang DP: Frequent chromosome 9p losses in histologically normal nasopharyngeal epithelia from southern Chinese. Int J Cancer 102(3): 300-303, 2002. PMID: 12397655. DOI: 10.1002/ ijc.10689

11 Cheng JQ, Jhanwar SC, Lu YY and Testa JR: Homozygous deletions within 9p21-p22 identify a small critical region of chromosomal loss in human malignant mesotheliomas. Cancer Res 53(20): 4761-4763, 1993. PMID: 8402655.

12 Devlin J, Elder PA, Gabra H, Steel CM and Knowles MA: High frequency of chromosome 9 deletion in ovarian cancer: evidence for three tumour-suppressor loci. Br J Cancer 73(4): 420-423, 1996. PMID: 8595153. DOI: 10.1038/bjc.1996.75

13 El-Mokadem I, Fitzpatrick J, Bondad J, Rauchhaus P, Cunningham J, Pratt N, Fleming S and Nabi G: Chromosome 9p deletion in clear cell renal cell carcinoma predicts recurrence and survival following surgery. Br J Cancer 111(7): 1381-1390, 2014. PMID: 25137021. DOI: 10.1038/bjc.2014.420

14 El-Mokadem I, Kidd T, Pratt N, Fleming S and Nabi G: Tumour suppressor gene (CDKNA2) status on chromosome 9p in resected renal tissue improves prognosis of localised kidney cancer. Oncotarget 7(45): 73045-73054, 2016. PMID: 27682877. DOI: 10.18632/oncotarget.12196

15 Farrell WE, Simpson DJ, Bicknell JE, Talbot AJ, Bates AS and Clayton RN: Chromosome 9p deletions in invasive and noninvasive nonfunctional pituitary adenomas: the deleted region involves markers outside of the MTS1 and MTS2 genes. Cancer Res 57(13): 2703-2709, 1997. PMID: 9205080

16 Gunawan B, Bergmann F, Höer J, Langer C, Schumpelick V, Becker H and Füzesi L: Biological and clinical significance of cytogenetic abnormalities in low-risk and high-risk gastrointestinal stromal tumors. Hum Pathol 33(3): 316-321, 2002. PMID: 11979372. DOI: 10.1053/hupa.2002.32216

17 Huang DP, Lo KW, van Hasselt CA, Woo JK, Choi PH, Leung SF, Cheung ST, Cairns P, Sidransky D and Lee JC: A region of homozygous deletion on chromosome 9p21-22 in primary nasopharyngeal carcinoma. Cancer Res 54(15): 4003-4006, 1994. PMID: 8033130

18 James CD, He J, Collins VP, Allalunis-Turner MJ and Day RS 3rd: Localization of chromosome 9p homozygous deletions in glioma cell lines with markers constituting a continuous linkage group. Cancer Res 53(16): 3674-3676, 1993. PMID: 8339274

19 M’soka TJ, Nishioka J, Taga A, Kato K, Kawasaki H, Yamada Y, Yu A, Komada Y and Nobori T: Detection of methylthioadenosine phosphorylase (MTAP) and p16 gene deletion in T cell acute lymphoblastic leukemia by real-time quantitative PCR assay. Leukemia 14(5): 935-940, 2000. PMID: 10803528. DOI: 10.1038/sj.leu.2401771

20 Mead LJ, Gillespie MT, Irving LB and Campbell LJ: Homozygous and hemizygous deletions of 9p centromeric to the interferon genes in lung cancer. Cancer Res 54(9): 2307-2309, 1994. PMID: 8162570.

21 Merlo A, Gabrielson E, Mabry M, Vollmer R, Baylin SB and Sidransky D: Homozygous deletion on chromosome 9p and loss of heterozygosity on 9q, 6p, and 6q in primary human small cell lung cancer. Cancer Res 54(9): 2322-2326, 1994. PMID: 8162574.

22 Middleton PG, Prince RA, Williamson IK, Taylor PR, Reid MM, Jackson GH, Katz F, Chessells JM and Proctor SJ: Alpha interferon gene deletions in adults, children and infants with acute lymphoblastic leukemia. Leukemia 5(8): 680-682, 1991. PMID: 1886420.

23 Olopade OI, Buchhagen DL, Malik K, Sherman J, Nobori T, Bader S, Nau MM, Gazdar AF, Minna JD and Diaz MO: Homozygous loss of the interferon genes defines the critical

region on 9p that is deleted in lung cancers. Cancer Res 53(10 Suppl): 2410-2415, 1993. PMID: 7683574.

24 Olopade OI, Jenkins RB, Ransom DT, Malik K, Pomykala H, Nobori T, Cowan JM, Rowley JD and Diaz MO: Molecular analysis of deletions of the short arm of chromosome 9 in human gliomas. Cancer Res 52(9): 2523-2529, 1992. PMID: 1568221.

25 Panani AD, Maliaga K, Babanaraki A and Bellenis I: Numerical abnormalities of chromosome 9 and p16CDKN2A gene deletion detected by FISH in non-small cell lung cancer. Anticancer Res 29(11): 4483-4487, 2009. PMID: 20032395.

26 Puig S, Ruiz A, Lázaro C, Castel T, Lynch M, Palou J, Vilalta A, Weissenbach J, Mascaro JM and Estivill X: Chromosome 9p deletions in cutaneous malignant melanoma tumors: the minimal deleted region involves markers outside the p16 (CDKN2) gene. Am J Hum Genet 57(2): 395-402, 1995. PMID: 7668266.

27 Sabah M, Cummins R, Leader M and Kay E: Loss of heterozygosity of chromosome 9p and loss of p16INK4A expression are associated with malignant gastrointestinal stromal tumors. Mod Pathol 17(11): 1364-1371, 2004. PMID: 15181453. DOI: 10.1038/modpathol.3800199

28 Stadler WM and Olopade OI: The 9p21 region in bladder cancer cell lines: large homozygous deletion inactivate the CDKN2, CDKN2B and MTAP genes. Urol Res 24(4): 239-244, 1996. PMID: 8873383. DOI: 10.1007/BF00295899

29 Stadler WM, Sherman J, Bohlander SK, Roulston D, Dreyling M, Rukstalis D and Olopade OI: Homozygous deletions within chromosomal bands 9p21-22 in bladder cancer. Cancer Res 54(8): 2060-2063, 1994. PMID: 7513608.

30 Taguchi T, Jhanwar SC, Siegfried JM, Keller SM and Testa JR: Recurrent deletions of specific chromosomal sites in 1p, 3p, 6q, and 9p in human malignant mesothelioma. Cancer Res 53(18): 4349-4355, 1993. PMID: 8364929.

31 Takita J, Hayashi Y, Kohno T, Yamaguchi N, Hanada R, Yamamoto K and Yokota J: Deletion map of chromosome 9 and p16 (CDKN2A) gene alterations in neuroblastoma. Cancer Res 57(5): 907-912, 1997. PMID: 9041193.

32 Tripathi A, Dasgupta S, Roy A, Sengupta A, Roy B, Roychowdhury S and Panda CK: Sequential deletions in both arms of chromosome 9 are associated with the development of head and neck squamous cell carcinoma in Indian patients. J Exp Clin Cancer Res 22(2): 289-297, 2003. PMID: 12866580.

33 Ueki K, Rubio MP, Ramesh V, Correa KM, Rutter JL, von Deimling A, Buckler AJ, Gusella JF and Louis DN: MTS1/CDKN2 gene mutations are rare in primary human astrocytomas with allelic loss of chromosome 9p. Hum Mol Genet 3(10): 1841-1845, 1994. PMID: 7849711. DOI: 10.1093/hmg/3.10.1841

34 van der Riet P, Nawroz H, Hruban RH, Corio R, Tokino K, Koch W and Sidransky D: Frequent loss of chromosome 9p21-22 early in head and neck cancer progression. Cancer Res 54(5): 1156- 1158, 1994. PMID: 8118798.

35 Zhao R, Choi BY, Lee MH, Bode AM and Dong Z: Implications of genetic and epigenetic alterations of CDKN2A (p16(INK4a)) in cancer. EBioMedicine 8: 30-39, 2016. PMID: 27428416. DOI: 10.1016/j.ebiom.2016.04.017

36 Piao Z, Park C, Lee JS, Yang CH, Choi KY and Kim H: Homozygous deletions of the CDKN2 gene and loss of heterozygosity of 9p in primary hepatocellular carcinoma. Cancer Lett 122(1-2): 201-207, 1998. PMID: 9464511. DOI: 10.1016/s0304-3835(97)00403-5

37 Usvasalo A, Savola S, Räty R, Vettenranta K, Harila-Saari A, Koistinen P, Savolainen ER, Elonen E, Saarinen-Pihkala UM and Knuutila S: CDKN2A deletions in acute lymphoblastic leukemia of adolescents and young adults: an array CGH study. Leuk Res 32(8): 1228-1235, 2008. PMID: 18328560. DOI: 10.1016/j.leukres.2008.01.014

38 Xia Y, Liu Y, Yang C, Simeone DM, Sun TT, DeGraff DJ, Tang MS, Zhang Y and Wu XR: Dominant role of CDKN2B/p15INK4B of 9p21.3 tumor suppressor hub in inhibition of cell-cycle and glycolysis. Nat Commun 12(1): 2047, 2021. PMID: 33824349. DOI: 10.1038/s41467-021-22327-5

39 Tomczak K, Czerwińska P and Wiznerowicz M: The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol (Pozn) 19(1A): A68-A77, 2015. PMID: 25691825. DOI: 10.5114/wo.2014.47136

40 Samur MK: RTCGAToolbox: a new tool for exporting TCGA Firehose data. PLoS One 9(9): e106397, 2014. PMID: 25181531. DOI: 10.1371/journal.pone.0106397

41 Colaprico A, Silva TC, Olsen C, Garofano L, Cava C, Garolini D, Sabedot TS, Malta TM, Pagnotta SM, Castiglioni I, Ceccarelli M, Bontempi G and Noushmehr H: TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res 44(8): e71, 2016. PMID: 26704973. DOI: 10.1093/nar/gkv1507

42 Luebke AM, Baudis M, Matthaei H, Vashist YK, Verde PE, Hosch SB, Erbersdobler A, Klein CA, Izbicki JR, Knoefel WT and Stoecklein NH: Losses at chromosome 4q are associated with poor survival in operable ductal pancreatic adenocarcinoma. Pancreatology 12(1): 16-22, 2012. PMID: 22487468. DOI: 10.1016/j.pan.2011.11.001

43 Tiago D. P. Ramos and Lidia M. F. Amorim: Molecular biology techniques for loss of heterozygosity detection: The glioma example. J Bras Patol Med Lab 51(3): 189-196, 2015. DOI: 10.5935/1676-2444.20150033

44 Tarmin L, Yin J, Zhou X, Suzuki H, Jiang HY, Rhyu MG, Abraham JM, Krasna MJ, Cottrell J and Meltzer SJ: Frequent loss of heterozygosity on chromosome 9 in adenocarcinoma and squamous cell carcinoma of the esophagus. Cancer Res 54(23): 6094-6096, 1994. PMID: 7954453.

45 Luo Y, Tian L, Feng Y, Yi M, Chen X and Huang Q: The predictive role of p16 deletion, p53 deletion, and polysomy 9 and 17 in pancreatic ductal adenocarcinoma. Pathol Oncol Res 19(1): 35-40, 2013. PMID: 22782330. DOI: 10.1007/s12253- 012-9555-3

46 Orlow I, Drobnjak M, Zhang ZF, Lewis J, Woodruff JM, Brennan MF and Cordon-Cardo C: Alterations of INK4A and INK4B genes in adult soft tissue sarcomas: effect on survival. J Natl Cancer Inst 91(1): 73-79, 1999. PMID: 9890173. DOI: 10.1093/jnci/91.1.73

47 Russell AJ, Sibbald J, Haak H, Keith WN and McNicol AM: Increasing genome instability in adrenocortical carcinoma progression with involvement of chromosomes 3, 9 and X at the adenoma stage. Br J Cancer 81(4): 684-689, 1999. PMID: 10574256. DOI: 10.1038/sj.bjc.6690748

48 Barros ÉAF, Pontes-Junior J, Reis ST, Lima AER, Souza IC, Salgueiro JL, Fontes D, Dellê H, Coelho RF, Viana NI, Leite KRM, Nahas WC and Srougi M: Correlation between chromosome 9p21 locus deletion and prognosis in clinically localized prostate cancer. Int J Biol Markers 32(2): e248-e254, 2017. PMID: 28058701. DOI: 10.5301/jbm.5000242

49 Fujino T, Risinger JI, Collins NK, Liu FS, Nishii H, Takahashi H, Westphal EM, Barrett JC, Sasaki H and Kohler MF: Allelotype of endometrial carcinoma. Cancer Res 54(16): 4294- 4298, 1994. PMID: 8044774

50 Jouhi L, Hagström J, Atula T and Mäkitie A: Is p16 an adequate surrogate for human papillomavirus status determination? Curr Opin Otolaryngol Head Neck Surg 25(2): 108-112, 2017. PMID: 28141601. DOI: 10.1097/MOO.0000000000000341

51 Azechi H, Nishida N, Fukuda Y, Nishimura T, Minata M, Katsuma H, Kuno M, Ito T, Komeda T, Kita R, Takahashi R and Nakao K: Disruption of the p16/cyclin D1/retinoblastoma protein pathway in the majority of human hepatocellular carcinomas. Oncology 60(4): 346-354, 2001. PMID: 11408803. DOI: 10.1159/000058531

52 Hui AM, Sakamoto M, Kanai Y, Ino Y, Gotoh M, Yokota J and Hirohashi S: Inactivation of p16INK4 in hepatocellular carcinoma. Hepatology 24(3): 575-579, 1996. PMID: 8781327. DOI: 10.1002/hep.510240319

53 Shim YH, Yoon GS, Choi HJ, Chung YH and Yu E: p16 Hypermethylation in the early stage of hepatitis B virus- associated hepatocarcinogenesis. Cancer Lett 190(2): 213- 219, 2003. PMID: 12565176. DOI: 10.1016/s0304- 3835(02)00613-4

54 Lv X, Ye G, Zhang X and Huang T: p16 Methylation was associated with the development, age, hepatic viruses infection of hepatocellular carcinoma, and p16 expression had a poor survival: A systematic meta-analysis (PRISMA). Medicine (Baltimore) 96(38): e8106, 2017. PMID: 28930859. DOI: 10.1097/MD.0000000000008106

55 Eleutério J Jr, Lima TS, Cunha MD, Cavalcante DI and Silva AM: Immunohistochemical expression of the tumor suppressor Protein p16INK4a in cervical adenocarcinoma. Rev Bras Ginecol Obstet 39(1): 21-25, 2017. PMID: 28245508. DOI: 10.1055/s-0037-1598042

56 Wang Y, Ren F, Chen P, Liu S, Song Z and Ma X: Identification of a six-gene signature with prognostic value for patients with endometrial carcinoma. Cancer Med 7(11): 5632-5642, 2018. PMID: 30306731. DOI: 10.1002/cam4.1806

57 De la Houssaye G, Vieira V, Masson C, Beermann F, Dufier JL, Menasche M and Abitbol M: ETS-1 and ETS-2 are upregulated in a transgenic mouse model of pigmented ocular neoplasm. Mol Vis 14: 1912-1928, 2008. PMID: 18958307.

58 Lamperska K, Mackiewicz K, Kaczmarek A, Kwiatkowska E, Starzycka M, Romanowska B, Heizman J, Stachura J and Mackiewicz A: Expression of p16 in sporadic primary uveal melanoma. Acta Biochim Pol 49(2): 377-385, 2002. PMID: 12362979.

59 Salazar CR, Anayannis N, Smith RV, Wang Y, Haigentz M Jr, Garg M, Schiff BA, Kawachi N, Elman J, Belbin TJ, Prystowsky MB, Burk RD and Schlecht NF: Combined p16 and human papillomavirus testing predicts head and neck cancer survival. Int J Cancer 135(10): 2404-2412, 2014. PMID: 24706381. DOI: 10.1002/ijc.28876

60 Menezes WP, Silva VAO, Gomes INF, Rosa MN, Spina MLC, Carloni AC, Alves ALV, Melendez M, Almeida GC, Silva LSD, Clara C, da Cunha IW, Hajj GNM, Jones C, Bidinotto LT and Reis RM: Loss of 5’-methylthioadenosine phosphorylase (MTAP) is frequent in high-grade gliomas; Nevertheless, it is not associated with higher tumor aggressiveness. Cells 9(2): , 2020. PMID: 32093414. DOI: 10.3390/cells9020492

61 Li Z, Jin Y, Zou Q, Shi X, Wu Q, Lin Z, He Q, Huang G and Qi S: Integrated genomic and transcriptomic analysis suggests KRT18 mutation and MTAP are key genetic alterations related to the prognosis between astrocytoma and glioblastoma. Ann Transl Med 9(8): 713, 2021. PMID: 33987411. DOI: 10.21037/atm-21-1317

62 Frazão L, do Carmo Martins M, Nunes VM, Pimentel J, Faria C, Miguéns J, Sagarribay A, Matos M, Salgado D, Nunes S, Mafra M and Roque L: BRAF V600E mutation and 9p21: CDKN2A/B and MTAP co-deletions - Markers in the clinical stratification of pediatric gliomas. BMC Cancer 18(1): 1259, 2018. PMID: 30558563. DOI: 10.1186/s12885-018-5120-0

63 Zhang C, Zhang J, Xu FP, Wang YG, Xie Z, Su J, Dong S, Nie Q, Shao Y, Zhou Q, Yang JJ, Yang XN, Zhang XC, Li Z, Wu YL and Zhong WZ: Genomic landscape and immune microenvironment features of preinvasive and early invasive lung adenocarcinoma. J Thorac Oncol 14(11): 1912-1923, 2019. PMID: 31446140. DOI: 10.1016/j.jtho.2019.07.031

64 Hiroshima K, Wu D, Hamakawa S, Tsuruoka S, Ozaki D, Orikasa H, Hasegawa M, Koh E, Sekine Y, Yonemori Y, Nabeshima K, Tsuji S, Miyagi Y and Imai K: HEG1, BAP1, and MTAP are useful in cytologic diagnosis of malignant mesothelioma with effusion. Diagn Cytopathol 49(5): 622-632, 2021. PMID: 32441895. DOI: 10.1002/dc.24475

65 Chapel DB, Dubuc AM, Hornick JL and Sholl LM: Correlation of methylthioadenosine phosphorylase (MTAP) protein expression with MTAP and CDKN2A copy number in malignant pleural mesothelioma. Histopathology 78(7): 1032-1042, 2021. PMID: 33387364. DOI: 10.1111/his.14324

66 Chapel DB, Schulte JJ, Berg K, Churg A, Dacic S, Fitzpatrick C, Galateau-Salle F, Hiroshima K, Krausz T, Le Stang N, McGregor S, Nabeshima K and Husain AN: MTAP immunohistochemistry is an accurate and reproducible surrogate for CDKN2A fluorescence in situ hybridization in diagnosis of malignant pleural mesothelioma. Mod Pathol 33(2): 245-254, 2020. PMID: 31231127. DOI: 10.1038/s41379-019-0310-0

67 Stelzer G, Rosen N, Plaschkes I, Zimmerman S, Twik M, Fishilevich S, Stein TI, Nudel R, Lieder I, Mazor Y, Kaplan S, Dahary D, Warshawsky D, Guan-Golan Y, Kohn A, Rappaport N, Safran M and Lancet D: The GeneCards suite: From gene data mining to disease genome sequence analyses. Curr Protoc Bioinformatics 54: 1.30.1-1.30.33, 2016. PMID: 27322403. DOI: 10.1002/cpbi.5

68 Schoch C, Schnittger S, Klaus M, Kern W, Hiddemann W and Haferlach T: AML with 11q23/MLL abnormalities as defined by the WHO classification: incidence, partner chromosomes, FAB subtype, age distribution, and prognostic impact in an unselected series of 1897 cytogenetically analyzed AML cases. Blood 102(7): 2395-2402, 2003. PMID: 12805060. DOI: 10.1182/blood-2003-02- 0434

69 Büttner N, Johnsen SA, Kügler S and Vogel T: Af9/Mllt3 interferes with Tbr1 expression through epigenetic modification of histone H3K79 during development of the cerebral cortex. Proc Natl Acad Sci USA 107(15): 7042-7047, 2010. PMID: 20348416. DOI: 10.1073/pnas.0912041107

70 Kikkawa T, Obayashi T, Takahashi M, Fukuzaki-Dohi U, Numayama-Tsuruta K and Osumi N: Dmrta1 regulates proneural gene expression downstream of Pax6 in the mammalian telencephalon. Genes Cells 18(8): 636-649, 2013. PMID: 23679989. DOI: 10.1111/gtc.12061

71 Nickerson ML, Witte N, Im KM, Turan S, Owens C, Misner K, Tsang SX, Cai Z, Wu S, Dean M, Costello JC and Theodorescu D: Molecular analysis of urothelial cancer cell lines for modeling tumor biology and drug response. Oncogene 36(1): 35- 46, 2017. PMID: 27270441. DOI: 10.1038/onc.2016.172

72 An J, Kim D, Oh B, Oh YJ, Song J, Park N, Kim HI, Kang HJ, Oh JH, Kim W, Lee E, Sung CO, Song GW, Kim DG, Yu E, Letouzé E, Zucman-Rossi J, Lee HC and Shim JH: Comprehensive characterization of viral integrations and genomic aberrations in HBV-infected intrahepatic cholangiocarcinomas. Hepatology 75(4): 997-1011, 2022. PMID: 34478159. DOI: 10.1002/hep.32135

73 Antic D, Lu N and Keene JD: ELAV tumor antigen, Hel-N1, increases translation of neurofilament M mRNA and induces formation of neurites in human teratocarcinoma cells. Genes Dev 13(4): 449-461, 1999. PMID: 10049360. DOI: 10.1101/ gad.13.4.449

74 Chagnovich D, Fayos BE and Cohn SL: Differential activity of ELAV-like RNA-binding proteins in human neuroblastoma. J Biol Chem 271(52): 33587-33591, 1996. PMID: 8969226. DOI: 10.1074/jbc.271.52.33587

75 Hatanaka T, Higashino F, Tei K and Yasuda M: The neural ELAVL protein HuB enhances endogenous proto-oncogene activation. Biochem Biophys Res Commun 517(2): 330-337, 2019. PMID: 31358321. DOI: 10.1016/j.bbrc.2019.07.089

76 Gupta A, Williams BR, Hanash SM and Rawwas J: Cellular retinoic acid-binding protein II is a direct transcriptional target of MycN in neuroblastoma. Cancer Res 66(16): 8100-8108, 2006. PMID: 16912187. DOI: 10.1158/0008-5472.CAN-05-4519

77 Berto S, Usui N, Konopka G and Fogel BL: Elavl2-regulated transcriptional and splicing networks in human neurons link neurodevelopment and autism. Hum Mol Genet 25(12): 2451- 2464, 2016. PMID: 27260404. DOI: 10.1093/hmg/ddw110

78 Kurokawa T, Nakagawa T, Matsusaka K, Fukuyo M, Mima M, Misawa K, Rahmutulla B, Ikeda JI, Hanazawa T, Okamoto Y and Kaneda A: Establishment of epigenetic markers to predict irradiation efficacy against oropharyngeal cancer. Cancer Sci 111(4): 1407-1416, 2020. PMID: 32012407. DOI: 10.1111/cas.14338

79 Zhao WS, Yan WP, Chen DB, Dai L, Yang YB, Kang XZ, Fu H, Chen P, Deng KJ, Wang XY, Xie XW, Chen HS and Chen KN: Genome-scale CRISPR activation screening identifies a role of ELAVL2-CDKN1A axis in paclitaxel resistance in esophageal squamous cell carcinoma. Am J Cancer Res 9(6): 1183-1200, 2019. PMID: 31285951.

Received July 6, 2022 Revised September 3, 2022 Accepted September 14, 2022