Patient characteristics
The RNA sequencing data of 378 samples from TCGA database and detailed clinical prognostic information were included in the analysis. Patients were divided into a low expression group (n =169) and a high expression group (n =169). Age, gender, lymphatic metastasis, metastasis did not differ significantly between the high and low expression groups (P = 0.5002, 0.7321, 0.5366, 0.6563). Grade, tumor size and stage was significantly different between the high and low expression groups (P = 0.0383, 1.00E-04, 0.017) (Table 1).
Table. 1
Correlation between AEBP1 expression in tumor and clinicopathological characteristics of patients with colorectal cancer.
Covariates
|
Type
|
Total
|
High
|
Low
|
Pvalue
|
Age
|
<=65
|
155(45.86%)
|
82(48.52%)
|
73(43.2%)
|
0.5002
|
Age
|
>65
|
179(52.96%)
|
87(51.48%)
|
92(54.44%)
|
|
Age
|
unknow
|
4(1.18%)
|
0(0%)
|
4(2.37%)
|
|
Grade
|
G1
|
8(2.37%)
|
3(1.78%)
|
5(2.96%)
|
0.0383
|
Grade
|
G2
|
123(36.39%)
|
51(30.18%)
|
72(42.6%)
|
|
Grade
|
G3
|
198(58.58%)
|
110(65.09%)
|
88(52.07%)
|
|
Grade
|
unknow
|
9(2.66%)
|
5(2.96%)
|
4(2.37%)
|
|
M
|
M0
|
301(89.05%)
|
151(89.35%)
|
150(88.76%)
|
0.6563
|
M
|
M1
|
19(5.62%)
|
8(4.73%)
|
11(6.51%)
|
|
M
|
unknow
|
18(5.33%)
|
10(5.92%)
|
8(4.73%)
|
|
N
|
N0
|
103(30.47%)
|
53(31.36%)
|
50(29.59%)
|
0.5366
|
N
|
N1
|
87(25.74%)
|
40(23.67%)
|
47(27.81%)
|
|
N
|
N2
|
71(21.01%)
|
33(19.53%)
|
38(22.49%)
|
|
N
|
N3
|
69(20.41%)
|
39(23.08%)
|
30(17.75%)
|
|
N
|
unknow
|
8(2.37%)
|
4(2.37%)
|
4(2.37%)
|
|
T
|
T1
|
18(5.33%)
|
1(0.59%)
|
17(10.06%)
|
1.00E-04
|
T
|
T2
|
67(19.82%)
|
29(17.16%)
|
38(22.49%)
|
|
T
|
T3
|
160(47.34%)
|
81(47.93%)
|
79(46.75%)
|
|
T
|
T4
|
93(27.51%)
|
58(34.32%)
|
35(20.71%)
|
|
T
|
unknow
|
0(0%)
|
0(0%)
|
0(0%)
|
|
Gender
|
female
|
118(34.91%)
|
57(33.73%)
|
61(36.09%)
|
0.7321
|
Gender
|
male
|
220(65.09%)
|
112(66.27%)
|
108(63.91%)
|
|
Stage
|
stage I
|
1(0.3%)
|
0(0%)
|
1(0.59%)
|
0.017
|
Stage
|
Stage I
|
45(13.31%)
|
13(7.69%)
|
32(18.93%)
|
|
Stage
|
Stage II
|
108(31.95%)
|
62(36.69%)
|
46(27.22%)
|
|
Stage
|
Stage III
|
145(42.9%)
|
76(44.97%)
|
69(40.83%)
|
|
Stage
|
Stage IV
|
29(8.58%)
|
13(7.69%)
|
16(9.47%)
|
|
Stage
|
unknow
|
10(2.96%)
|
5(2.96%)
|
5(2.96%)
|
|
expression
|
High
|
169(50%)
|
169(100%)
|
-
|
0
|
expression
|
Low
|
169(50%)
|
-
|
169(100%)
|
|
methylation
|
High
|
169(50%)
|
80(47.34%)
|
89(52.66%)
|
0.3841
|
methylation
|
Low
|
169(50%)
|
89(52.66%)
|
80(47.34%)
|
|
AEBP1 expression is higher in normal tissues than in tumor samples
Analysis of the mRNA expression levels of AEBP1 in TCGA samples showed that AEBP1 expression was lower in tumor samples than in normal tissues (P < 0.022) (Figure. 1A). Analysis of the mRNA expression levels of AEBP1 in TCGA samples showed that AEBP1 expression was higher in tumor samples than in normal tissues (Fig. 1B).
Transcriptional Expression and Epigenetic Regulation of AEBP1 Across Various Clinicopathological Parameters
The expression of AEBP1 in GC was analyzed based on the different clinicopathological parameters like sample type, individual cancer stage, cancer grades, patient’s sex and age, histological subtype, nodal metastasis status, and TP53 mutation status using the UALCAN server (Figure. 2 and Table. 2). The results support the inference depicted in the earlier section by demonstrating that AEBP1 expression is higher in the colorectal cancer tissue at different clinical stages than in normal tissue (Figure. 2A-F). It tends to increase the expression of AEBP1 at advanced stages of cancer (Stage 4 > Stage 3 > Stage 1) (Figure. 2B) and decrease along with the increase in the age group of patients (Figure 2C). It was also found that the expression of the COL11A1 gene increases along with the nodal metastasis status (N3 > N2 > N1) (Figure. 2E).
Table. 2
The AEBP1 mRNA expression for gastric cancer based on different clinicopalthoogical parameters using UALCAN.
Variables
|
Different stages
|
N
|
Comparisons
|
Statistical significance
|
Statistical significance
|
Normal
|
41
|
Normal vs. primary tumor
|
3.64E-05
|
Primary tumor
|
286
|
Individual cancer stages
|
Stage 1
|
34
|
Normal-vs-Stage1
|
1.02E-02
|
Stage 2
|
123
|
Normal-vs-Stage2
|
2.47E-06
|
Stage 3
|
169
|
Normal-vs-Stage3
|
1.64E-04
|
Stage 4
|
41
|
Normal-vs-Stage4
|
2.38E-03
|
Patients age
|
21-40 yrs
|
34
|
Normal-vs-Age(21-40Yrs)
|
8.88E-01
|
41-60 yrs
|
4
|
Normal-vs-Age(41-60Yrs)
|
1.54E-06
|
61-80 yrs
|
128
|
Normal-vs-Age(61-80Yrs)
|
1.21E-04
|
81-100 yrs
|
253
|
Normal-vs-Age(81-100Yrs)
|
3.83E-01
|
Histological subtype
|
Adenocarcinoma NOS
|
155
|
Normal-vs-Adenocarcinoma(NOS)
|
2.01E-06
|
Adenocarcinoma Diffuse
|
69
|
Normal-vs-Adenocarcinoma(Diffuse)
|
6.91E-07
|
Adenocarcinoma SignetRing
|
12
|
Normal-vs-Adenocarcinoma(Signet Ring)
|
7.74E-02
|
IntestinalAdenocarcinoma NOS
|
73
|
Normal-vs-IntestinalAdenocarcinoma(NOS)
|
5.13E-02
|
IntestinalAdenocarcinoma Tubular
|
76
|
Normal-vs-IntestinalAdenocarcinoma(Tubular)
|
6.85E-01
|
IntestinalAdenocarcinoma Mucinous
|
20
|
Normal-vs-IntestinalAdenocarcinoma(Mucinous)
|
8.20E-04
|
IntestinalAdenocarcinoma Papillary
|
7
|
Normal-vs-IntestinalAdenocarcinoma(Papillary)
|
7.73E-01
|
Nodal metastasis status
|
N0-No regional lymph node metastasis
|
123
|
Normal-vs-N0
|
9.88E-05
|
N1- Metastases in 1 to 3 axillary lymph nodes
|
112
|
Normal-vs-N1
|
8.79E-04
|
N2-Metastases in 4 to 9 axillary lymph nodes
|
79
|
Normal-vs-N2
|
2.92E-03
|
Metastases in 10 or more axillary lymph nodes
|
82
|
Normal-vs-N3
|
1.38E-04
|
TP53 mutation status
|
TP53 mutation status
|
178
|
Normal-vs-TP53-Mutant
|
2.58E-03
|
TP53 non-mutant
|
235
|
Normal-vs-TP53-NonMutant
|
2.21E-06
|
DNA methylation is relatively associated with the development of cancer within the human body[28]. From our data, it was evident that the promoter methylation of the AEBP1 gene is overexpressed in the colon cancer tissue than that of the normal tissue, and is negatively regulated for all other clinicopathological parameters (Figures. 2G–L and Table. 3). It is reflected that along with the development of cancer stages and nodal metastasis status, the expression of promoter methylation decreases in the tissues (N0 > N1 > N3) (Figures. 2K). These results indicate that the promoter methylation is negatively associated with the expression of AEBP1 mRNA, and the hypermethylation of the promoter of AEBP1 may inhibit AEBP1 in upgrading cancer development.
Table. 3
The AEBP1 promoter methylation for gastric cancer based on different clinicopathological parameters using UALCAN.
Variables
|
Different stages
|
N
|
Comparisons
|
Statistical significance
|
Statistical significance
|
Normal
|
2
|
Normal vs. primary tumor
|
3.64E-05
|
Primary tumor
|
235
|
Individual cancer stages
|
Stage 1
|
34
|
Normal-vs-Stage1
|
6.43E-01
|
Stage 2
|
123
|
Normal-vs-Stage2
|
7.49E-01
|
Stage 3
|
169
|
Normal-vs-Stage3
|
7.67E-01
|
Stage 4
|
41
|
Normal-vs-Stage4
|
9.80E-01
|
Patients age
|
21-40 yrs
|
4
|
Normal-vs-Age(21-40Yrs)
|
2.77E-01
|
41-60 yrs
|
127
|
Normal-vs-Age(41-60Yrs)
|
8.62E-01
|
61-80 yrs
|
237
|
Normal-vs-Age(61-80Yrs)
|
6.98E-01
|
81-100 yrs
|
22
|
Normal-vs-Age(81-100Yrs)
|
6.49E-01
|
Histological subtype
|
Grade1
|
9
|
Normal-vs-Grade1
|
6.32E-01
|
Grade2
|
142
|
Normal-vs-Grade2
|
7.18E-01
|
Grade3
|
235
|
Normal-vs-Grade3
|
7.81E-01
|
Nodal metastasis status
|
N0-No regional lymph node metastasis
|
124
|
Normal-vs-N0
|
7.72E-01
|
N1- Metastases in 1 to 3 axillary lymph nodes
|
101
|
Normal-vs-N1
|
6.94E-01
|
N2-Metastases in 4 to 9 axillary lymph nodes
|
79
|
Normal-vs-N2
|
9.36E-01
|
Metastases in 10 or more axillary lymph nodes
|
28
|
Normal-vs-N3
|
5.99E-01
|
TP53 mutation status
|
TP53 mutation status
|
170
|
Normal-vs-TP53-Mutant
|
5.87E-01
|
TP53 non-mutant
|
225
|
Normal-vs-TP53-NonMutant
|
8.82E-01
|
Gastric cancer AEBP1 expression analysis
The expression of AEBP1 in GC was confirmed by immunohistochemistry (Figure. 3A−F). The results confirmed that AEBP1 expression was higher in GC tissues than in normal gastric tissues.
Higher AEBP1 mRNA expression in GC is associated with shorter OS
According to the KM chart, GC cases with higher AEBP1 mRNA expression had a shorter OS (P = 0.007, 0.003, 0.011) (Figure. 4A-C), methylation sites cg27493928, cg25289803, cg08739576, cg06128448, cg02126753 GC cases with lower AEBP1 mRNA expression had a shorter OS (P = 0.007, 0.003, 0.04, 0.046, 0.034) (Figure. 4D), methylation sites cg12978582 GC cases with higher AEBP1 mRNA expression had a shorter OS (P = 0.043), methylation sites cg1083171, cg25289803, cg27493928 GC cases with higher AEBP1 mRNA expression had a shorter progression-free survival (P = 0.038, 0.017, 0.040) in the test cohort (Figure. 4E). The results showed that the expression of AEBP1 was lower in gastric cancer than in normal gastric tissues. A meta-analysis of the above three datasets was performed to evaluate the correlation between overall survival (OS) and expression AEBP1 gene and obtain more objective conclusions. Because there was no statistically significant difference between the three datasets (P = 0.54, I2 = 0%), a fixed effects model was used to evaluate the combined hazard ratio (HR) and 95% confidence interval (CI). A relatively high expression of the AEBP1 gene was significantly correlated with poor OS (HR = 1.20, 95% CI: 1.10–1.30, P < 0.0001; Figure. 4G), indicating that AEBP1 may be a predictor of poor OS.
According to the ROC curve analysis and measurement based on the predictive power and accuracy of the risk characteristics expressed by AEBP1, the results show that the 1-year, 3-year, and 5-year predicted AUC are 0.540, 0.608, and 0.759, respectively (Figure. 4F). And accuracy, and check the forecasting ability and accuracy for 1, 3, and 5 years.
Multivariate Cox analysis and univariate Cox analysis are used to check whether AEBP1 can be used as an independent prognostic factor (Figure. 5). The results show that the p value of AEBP1 is less than 0.05, indicating that AEBP1 can be used as an independent prognostic factor. The hazard ratio is greater than 1, which is greater than the hazard ratio of the most common clinical features, indicating that the expression of base AEBP1 is highly predictive of risk, and the expression of base AEBP1 is more closely related to prognosis than common clinical features(Figure. 5A,B).
Correlation between AEBP1 expression and clinical characteristics
Analysis of the association between AEBP1 mRNA expression and clinicopathological parameters in GC patients showed that AEBP1 expression decreased with age (P = 0.043) (Figure. 6A). high expression of AEBP1 was associated with gastric cancer progression from G2 to G3 (P <0.0015) (Figure. 6A). In terms of tumor size, high expression of AEBP1 was associated with gastric cancer progression from T1 to T2, T1 to T3 and T1 to T4 has an important influence (P = 8.6e-07, 2.7e-07, 1.9e-07) (Figure. 6A). Regarding tumor stage, t high expression of AEBP1 was associated with gastric cancer progression from stage I to stage II, stage I to stage III, stage I to stage IV and stage II to stage III has an important influence (P = 5.5e-05, 0.0026, 0.023, 0.083) (Figure. 6A). The level of AEBP1 gene methylation was associated with T2 to T3, T2 to T4 (P=0.0064, 0.042). (Figure. 6B).
DNA methylation analysis
The degree of methylation of cg06852744 was the highest, followed from high to low by cg10480062, cg08495088, cg00009293, cg12955216, cg10873171, cg14249876, cg12978582, cg02126753, cg27493928, cg25289803, cg06128448, cg08739576 and cg01399219 (Figure. 7A). Among them, methylation sites such as cg00009293, cg06852744, cg08495088, cg10480062, cg12955216 and cg12978582 were positively correlated with high expression of AEBP1(Figure. 7C). The sites cg25289803, cg14249876, cg08739576, cg06128448, cg02126753, cg01399219, cg27493928, cg07476508, and cg10873171 methylation sites were negatively correlated with high expression of AEBP1 (Figure. 7B). The methylation level of AEBP1 was lower in GC than in the normal control group, indicating that changes in methylation are related to the abnormal expression of AEBP1. In addition, DNA methylation may be related to the molecular mechanism underlying the high expression level of AEBP1 in tumor tissues and the pathogenesis of GC.
Correlation analysis between infiltrating immune cells and AEBP1 expression
Tumor infiltrating lymphocytes affect the survival of various cancer patients. Therefore, we analyzed the correlation between AEBP1 expression and four infiltrating immune cells (neutrophils, CD8 + T cells, and CD4 + T cells) and three immune-related genes (CD2, CD3D, and CD3E). The results showed that the expression level of AEBP1 is comparable to neutrophils (r = 0.27, P = 8.76e-08), CD4+ T cells (r = 0.275, P = 5.34e-08), CD8 + T cells (r = 0.382, P = 1.18e-04), CD2 (r = 0.235, P = 1.25e-06), CD3D (r = 0.177, P = 2.94e-04) and CD3E (r = 0.227, P = 3.08e-06) (2021.8.17) infiltration levels are significantly positively correlated. P <0.05 was considered as the significance level (Figure. 8). These results indicate that high AEBP1 expression was related to the immunosuppressive microenvironment.
GO and KEGG enrichment analyses
GO enrichment analysis and KEGG pathway analysis were performed. Figure 9A shows the top 30 significantly enriched upregulated pathways. GO pathway analysis showed that relatively high expression of AEBP1 is associated with extracellular matrix organization (GO:0030198, P = 4.01e-60), extracellular structure organization (GO:0043062, P = 4.8e-60), collagen fibril organization (GO:0030199, P = 1.47e-27), and ossification (GO:0001503, P = 1.77e-20) (Figure. 9A). The KEGG pathway analysis identified many enriched pathways, including extracellular matrix organization, extracellular structure organization, collagen fibril organization, ossification and cell-substrate adhesion, and AEBP1 genes were strongly linked (Figure. 9B)