1. APOA2 is up-regulated in STAD tissues
In the GEPIA database, all tumor samples and corresponding normal tissue expression profiles, we find APOA2 expression is significantly up-regulated in STAD tumor tissues. Through analysis of TCGA database samples, APOA2 expression is found to be significantly elevated (p < 0.05) in tumor samples (n = 373) compared to normal tissue samples (n = 32) (Figure1A, B). Clinically, STAD is diagnosed base on the expression of APOA2, and its ROC curve (AUC = 0.69, p < 0.05) also confirms that APOA2 is up-regulated in gastric cancer tissues (Figure1C)
2. APOA2 overexpression is correlated with STAD clinical stage grade
The OS threshold of 373 STAD patients is used as the cut-off point of APOA2 expression, which is divided into the group with high APOA2 expression and low APOA2 expression, and the clinical significance of APOA2 expression is analyzed according to the corresponding indicators (Table 4). We respectively analyze age, gender, family history, TNM stage, histological grade, residual tumor, survival status, and disease status and found that there is no significant difference between high and low APOA2 expression in these parameters (p > 0.05). We further analyzed the clinical parameters significantly associated with APOA2 overexpression and found that in TNM stage, there were significant differences in APOA2 overexpression between TNM stage I and TNM stage II, TNM stage I and TNM stage III, TNM stage I and TNM stage IV, TNM stage II and TNM stage III, and TNM stage III and TNM stage IV. In histological grade, G1 grade versus G2 grade, G2 grade versus G3 grade, there is a significant difference in APOA2 overexpression (p < 0.05).We confirm that APOA2 overexpression is significantly different in TNM stage, histological grade, survival status, and disease status (Figure 2).
Table 4 Relationship between APOA2 expression and clinical parameters in patients with TCGA hepatocellular carcinoma
APOA2 Expression
|
|
Total
|
High
|
Low
|
P-value
|
|
(N=338)
|
(N=143)
|
(N=195)
|
|
Age (year)
|
|
|
|
|
< 65
|
143 (42.3%)
|
57 (39.9%)
|
86 (44.1%)
|
0.686
|
≥ 65
|
192 (56.8%)
|
85 (59.4%)
|
107 (54.9%)
|
|
Unknown
|
3 (0.9%)
|
1 (0.7%)
|
2 (1.0%)
|
|
Gender
|
|
|
|
|
Male
|
217 (64.2%)
|
87 (60.8%)
|
130 (66.7%)
|
0.323
|
Female
|
121 (35.8%)
|
56 (39.2%)
|
65 (33.3%)
|
|
Family history of cancer
|
|
|
|
|
NO
|
257 (76.0%)
|
106 (74.1%)
|
151 (77.4%)
|
0.439
|
YES
|
15 (4.4%)
|
5 (3.5%)
|
10 (5.1%)
|
|
Unknown
|
66 (19.5%)
|
32 (22.4%)
|
34 (17.4%)
|
|
TNM stage
|
|
|
|
|
I
|
48 (14.2%)
|
27 (18.9%)
|
21 (10.8%)
|
0.193
|
II
|
109 (32.2%)
|
43 (30.1%)
|
66 (33.8%)
|
|
III
|
147 (43.5%)
|
58 (40.6%)
|
89 (45.6%)
|
|
IV
|
34 (10.1%)
|
15 (10.5%)
|
19 (9.7%)
|
|
Histologic grade
|
|
|
|
|
G1–G2
|
127 (37.6%)
|
54 (37.8%)
|
73 (37.4%)
|
0.899
|
G3–G4
|
203 (60.1%)
|
85 (59.4%)
|
118 (60.5%)
|
|
Unknown
|
8 (2.4%)
|
4 (2.8%)
|
4 (2.1%)
|
|
Residual tumor
|
|
|
|
|
R0
|
282 (83.4%)
|
113 (79.0%)
|
169 (86.7%)
|
0.0973
|
R1-R2
|
28 (8.3%)
|
13 (9.1%)
|
15 (7.7%)
|
|
Unknown
|
28 (8.3%)
|
17 (11.9%)
|
11 (5.6%)
|
|
Living status
|
|
|
|
|
Alive
|
204 (60.4%)
|
81 (56.6%)
|
123 (63.1%)
|
0.279
|
Dead
|
134 (39.6%)
|
62 (43.4%)
|
72 (36.9%)
|
|
Disease status
|
|
|
|
|
NO
|
173 (51.2%)
|
66 (46.2%)
|
107 (54.9%)
|
0.23
|
YES
|
37 (10.9%)
|
19 (13.3%)
|
18 (9.2%)
|
|
Unknown
|
128 (37.9%)
|
58 (40.6%)
|
70 (35.9%)
|
|
3.APOA2 overexpression predicts OS and RFS in STAD
To evaluate the prognostic value of APOA2 expression in STAD, the OS threshold is used as a cut-off point for APOA2 expression and divide into groups with high APOA2 expression and low APOA2 expression. We found that patients with high APOA2 expression have significantly lower OS and RFS survival within 6 years than those with low APOA2 expression (p < 0.05) (Figure 3). We use Cox regression analysis to further investigate the relationship between APOA2 expression in each index and OS, RFS. In univariate analysis, TNM stage, residual tumor, and age are correlated with OS (p < 0.05); histological grade is correlated with RFS (p < 0.05). To further exclude the influence of other factors, multivariate analysis find that APOA2 overexpression is an independent indicator in OS and RFS. APOA2 overexpression has a regulatory effect on other indicators in gastric cancer (Tables 5 and Tables 6). In TNM stage subunits, TNM stage I compare with TNM II, and TNM stage III compare with TNM stage IV, it is found that the OS and RFS survival rates of patients with high MEGEA6 expression are significantly lower than those of patients with low APOA2 expression within 6 years (p < 0.05) (Figure 4).
Table 5 , Cox proportional hazards regression model analysis of overall survival rate
|
Univariate analysis
|
Multivariate analysis
|
Variables
|
HR (95% CI)
|
p
|
HR (95% CI)
|
p
|
Age (≥ 65 vs. < 65)
|
1.6(1.12,2.29)
|
0.010
|
1.65(1.14,2.38)
|
0.008
|
Gender (Female vs. Male)
|
0.76(0.53,1.1)
|
0.151
|
-
|
-
|
Family history of cancer (YES vs. NO)
|
1.05(0.51,2.17)
|
0.885
|
-
|
-
|
TNM stage (II vs. I)
|
1.55(0.78,3.07)
|
0.211
|
1.53(0.77,3.05)
|
0.227
|
TNM stage (III vs. I)
|
2.42(1.28,4.58)
|
0.007
|
2.36(1.23,4.49)
|
0.009
|
TNM stage (IV vs. I)
|
4.08(1.98,8.43)
|
<0.001
|
3.62(1.67,7.83)
|
0.001
|
Histologic grade (G3–G4 vs. G1–G2)
|
1.4(0.97,2.02)
|
0.068
|
-
|
-
|
Residual tumor (R1–R2 vs. R0)
|
3.12(1.92,5.07)
|
<0.001
|
2.26(1.32,3.87)
|
0.003
|
APOA2 (high vs. low)
|
1.02(0.98,1.06)
|
0.324
|
-
|
-
|
Table 6, Cox proportional hazards regression model analysis of relapse-free survival
|
Univariate analysis
|
Multivariate analysis
|
Variables
|
HR (95% CI)
|
p
|
HR (95% CI)
|
p
|
Age (≥ 65 vs. < 65)
|
1.13(0.59,2.16)
|
0.717
|
-
|
-
|
Gender (Female vs. Male)
|
0.55(0.27,1.15)
|
0.112
|
-
|
-
|
Family history of cancer (YES vs. NO)
|
0.88(0.21,3.68)
|
0.860
|
-
|
-
|
TNM stage (II vs. I)
|
2.03(0.72,5.7)
|
0.179
|
-
|
-
|
TNM stage (III vs. I)
|
1.57(0.57,4.33)
|
0.387
|
-
|
-
|
Histologic grade (G3–G4 vs. G1–G2)
|
2.18(1.03,4.63)
|
0.043
|
2.18(1.03,4.63)
|
0.043
|
Residual tumor (R1–R2 vs. R0)
|
1.12(0.15,8.22)
|
0.913
|
-
|
-
|
APOA2 (high vs. low)
|
1.08(1,1.17)
|
0.052
|
-
|
-
|
4. Validation of the prognostic value of APOA2 in gastric cancer
To validate the prognostic value of APOA2 in gastric cancer, we use APOA2 expression and TNM stage as independent indicators by multivariate analysis and constructed nomogram prognostic analysis to predict the 1-year, 3-year, and 5-year survival rates of OS and RFS. By constructing a line graph, we find that the survival correction values predict by the prognostic analysis of gene-clinical TNM staging nomogram are more consistent with the actual survival observation values. Confirm that by multivariate analysis, we can predict the prognosis and survival of gastric cancer patients more accurately (Figure 5).
5. APOA2 is upregulated in STAD by miR-27b-3p downregulation
We explore the mechanism by which APOA2 is upregulated in STAD from a genetic perspective. We evaluate the relationship between APOA2 expression and APOA2 methylation,and found that APOA2 expression is negatively correlated with APOA2 methylation (p < 0.05), illustrating that methylation can inhibit APOA2 expression. We in turn explore potential miRNAs capable of regulating MAGEA in STAD and select miR-27b-3p as a miRNA regulating APOA2. Through the GCTA database, we find that miR-27b-3p is significantly down-regulated (p < 0.05) in gastric cancer tissues (n = 434) compare with normal tissues (n = 41), and miR-27b-3p is negatively correlated with APOA2 (p < 0.05). Analysis of the relationship between miR-27b-3p down-regulation and OS, RFS reveal that low miR-27b-3p expression causes high APOA2 expression which decreases OS and RFS. We find that miR-27b-3p downregulation is significantly associated with decreased OS and RFS in STAD within 4 years (p < 0.05) (Figure 6).
6.APOA2 Enrichment Analysis
We select the top 50 differentially expressed genes with high enrichment in gastric cancer according to the degree of enrichment and we construct a PPI network map and find that APOA2 is directly associated with two genes, AFP and ALB. The matascape database is used to perform GO enrichment analysis of APOA2 to investigate the role of APOA2 in molecular function (MF), cellular component (CC), and biological process (BP) in STAD. We find that APOA2 is significantly enriched in endoplasmic reticulum lumen in CC, steroid metabolic process; negative regulation of catabolic process; negative regulation of hydrolase activity in BP (p < 0.05) (Figure 7). We also use GSEA to investigate the role of APOA2 upregulation in KEGG pathways and find that APOA2 is significantly enriched in "DNA replication", "cell cycle" (p < 0.05) (Figure 8).
7.APOA2 Expression Results
We obtained the Cq values of APOA2 and β-actin in gastric cancer tissues and normal gastric tissues by qPCR experiments, and the smaller the Cq, the higher the expression of this gene in the samples. We used 2- ∆∆ Ct as a parameter for gene expression and found that APOA2 expression was significantly higher in gastric cancer tissues than in normal gastric tissues by independent samples t-test (p < 0.05) (Figure 9).