Transcript level differences of ITGBs in GC patients
Currently, a total of eight ITGBs are found to be widely existing in mammalian cells. However, they are differentially expressed in various tumor tissues. We compared the differences in transcript levels of ITGBs in different tumors as well as in the corresponding normal tissues using the Oncomine database (Fig. 1). The results showed that the mRNA expression levels of ITGB1/2/3/5/8 were upregulated in GC patients while the mRNA expression level of ITGB7 was downregulated. Furthermore, for the expression levels of ITGB3 and ITGB6, there was no statistical significance between GC samples and their corresponding normal samples. In the dataset of Chen[24], D'Errico[25] and Cho[26], the ITGB1 transcript levels in GC samples were dramatically higher than in normal tissues, with P-values varying from 4.08E-4 to 8.56E-17. Three datasets include Chen dataset, D'Errico dataset and Wang dataset[27] showed ITGB2 expression levels were also remarkably elevated in GC patients. In the Cho dataset, ITGB4 expression was markedly increased in both subgroups of GC compared to normal samples, with fold changes of 2.105 (P = 0.001) and 2.008 (P = 4.87E-5), respectively. In the D'Errico dataset, the expression of ITGB5 was dramatically elevated in GC with a fold change of 3.241 (P = 4.89E-9). However,in two datasets (D'Errico and Cho), the expression of ITGB7 was downregulated in GC. For the expression of ITGB8, several datasets showed its high expression in tumor tissues (Table 1).
Table 1
The signifcant changes of ITGBs in transcription level (ONCOMINE database)
Gene name
|
Types of GC vs. normal
|
Fold change
|
t-test
|
P-value
|
Ref
|
ITGB1
|
Gastric intestinal Adenocarcinoma
|
2.130
|
10.951
|
8.56E-17
|
[24]
|
|
Gastric Mixed Adenocarcinoma
|
3.145
|
5.242
|
4.08E-4
|
[24]
|
|
Diffuse Gastric Adenocarcinoma
|
2.123
|
4.241
|
9.32E-4
|
[25]
|
|
Gastric Intestinal Type Adenocarcinoma
|
2.028
|
3.570
|
5.05E-4
|
[26]
|
|
Gastric Mixed Adenocarcinoma
|
2.789
|
3.731
|
9.79E-4
|
[26]
|
|
Diffuse Gastric Adenocarcinoma
|
-4.325
|
-3.451
|
0.004
|
[25]
|
ITGB2
|
Gastric Mixed Adenocarcinoma
|
2.398
|
4.235
|
5.20E-4
|
[24]
|
|
Diffuse Gastric Adenocarcinoma
|
2.033
|
3.845
|
4.45E-4
|
[24]
|
|
Diffuse Gastric Adenocarcinoma
|
2.318
|
3.624
|
0.003
|
[25]
|
|
Gastric Cancer
|
2.020
|
3.101
|
0.002
|
[27]
|
ITGB3
|
NA
|
NA
|
NA
|
NA
|
NA
|
ITGB4
|
Gastric Intestinal Type Adenocarcinoma
|
2.105
|
4.038
|
1.41E-4
|
[26]
|
|
Gastric Mixed Adenocarcinoma
|
2.008
|
3.242
|
0.003
|
[26]
|
ITGB5
|
Gastric Intestinal Type Adenocarcinoma
|
3.241
|
6.783
|
4.89E-9
|
[25]
|
ITGB6
|
NA
|
NA
|
NA
|
NA
|
NA
|
ITGB7
|
Gastric Mixed Adenocarcinoma
|
-2.723
|
-5.811
|
2.08E-5
|
[26]
|
|
Diffuse Gastric Adenocarcinoma
|
-2.157
|
-5.450
|
8.74E-7
|
[26]
|
|
Gastric Adenocarcinoma
|
-2.449
|
-4.872
|
0.003
|
[26]
|
|
Gastric Intestinal Type Adenocarcinoma
|
-2.318
|
-3.981
|
1.48E-4
|
[25]
|
ITGB8
|
Gastric Cancer
|
4.543
|
4.333
|
1.05E-4
|
[27]
|
|
Gastric Mixed Adenocarcinoma
|
5.245
|
7.495
|
6.93E-5
|
[25]
|
|
Gastric Intestinal Type Adenocarcinoma
|
3.133
|
6.845
|
3.88E-9
|
[25]
|
|
Gastric Mixed Adenocarcinoma
|
2.092
|
4.403
|
7.71E-4
|
[24]
|
NA, not available. |
Association between mRNA expression levels of ITGBs and clinicopathological parameters of GC patients
With the help of the GEPIA database and UALCAN database, we compared the transcript level differences of ITGBs between GC samples and their corresponding normal samples. The results from GEPIA database showed that the transcript levels of ITGB1/2/4/8 in GC tissues were markedly higher than their corresponding normal tissues, while the transcript levels of ITGB3/5/6/7 were not significantly contrasted between GC tissues and their corresponding normal samples (Fig. 2). Meanwhile,the results of the UALCAN database indicated that the transcript levels of ITGB1/2/4/5/6/8 were markedly greater in GC tissues than in their corresponding normal tissues, whereas the transcript levels of ITGB3/7 were not significantly contrasting in GC tissues and their corresponding normal samples (Fig. 3). In addition,we used the GEPIA database to compare the differences of mRNA expression of ITGBs in GC patients at different tumor stages, the results showed that the expression levels of ITGB2 and ITGB7 are notably related to the tumor stage of GC patients, while the expression levels of ITGB1/3/4/5/6/8 are not related to the tumor stage of GC patients (Fig. 4).
To verify the above results, we examined the mRNA expression level of ITGBs in GC cell lines (AGS, MGC803, HGC27) and stomach normal cell lines (GES-1) using qRT-PCR. The results indicated that the expression level of ITGB1/2/4/5/6/7/8 was highly expressed in GC cell lines compared with normal cell lines, whereas the expression level of ITGB3 was lower in the former than in the latter (P < 0.05, Fig. 5).
The prognostic value of ITGBs in GC
We investigated the prognostic significance of ITGBs transcription levels in GC using Kaplan–Meier plotter. As shown in Fig. 6, GC patients with higher expression levels of ITGB2/6/7 were remarkly related to better RFS. In contrast, GC patients with low expression levels of ITGB3/4/5 were significantly correlation with better RFS, while expression levels of ITGB1/8 were not related to RFS in GC patients. Besides,as shown in Fig. 7, GC patients with high expression of ITGB1/6/7/8 had longer OS,while patients with high expression of ITGB3/4/5 had shorter OS, only ITGB2 transcription levels were not OS related in GC patients. In summary, the findings suggest that high expression of ITGB1/6/7/8 may be a risk factor, while high expression of ITGB3/4/5 may be a protective factor for GC prognosis.
Gene–gene interaction network and PPI Analyses of ITGBs in GC Patients
We used the GeneMANIA database to build a gene–gene interaction network of ITGBs and then study their functions. As shown in Fig. 8a, eight ITGBs are surrounded by 20 related genes, these 20 genes stand for genes that are highly linked to ITGBs in terms of shared protein domains, physical interactions, co-localization, co-expression, prediction, genetic interactions and pathways. The five most relevant genes for ITGBs are ITGBL1(integrin subunit beta like 1), WIF1(WNT inhibitory factor 1), EGFL7(EGF like domain multiple 7), ATRNL1(attractin like 1) and TENM4(teneurin transmembrane protein 4). Among them, ITGB3 is related to TENM4 and ITGB8 in terms of co-expression and has a physical interaction and prediction with ITGB1. ITGB6 is related to ITGB1 and ITGB4 is related to ITGB5 in terms of co-localization. ITGBL1 and ATRNL1 share the same protein structural domain with all ITGBs family members. Meanwhile,WIF1,EGFL7 and TENM4 exhibit the same protein structural domains with the whole ITGBs members except ITGB2. In addition, functional analysis has shown that these genes are closely associated with protein complex involved in cell adhesion, receptor complex, rntegrin complex, cell-substrate adhesion, integral component of plasma membrane, plasma membrane signaling receptor complex and gastrulation.
Using STRING, we further investigated the protein-protein interaction network of 8 core genes (ITGB1/2/3/4/5/6/7/8) and their 20 functional partners with the top concordance scores. The most five relevant proteins with the ITGBs comprised PLEC (Plectin), ITGA5 (Integrin alpha-5), ITGAL (Integrin alpha-L), ITGAV (Integrin alpha-V) and ITGA2B (Integrin alpha-IIb) (Fig. 8b).
Functional enrichment analysis of ITGBs in GC patients
The Metascape database was adopted for functional annotation and pathway enrichment analysis of ITGBs and their neighboring genes. Among the top 20 items enriched by GO, 13 were biological processes(BP), 3 were molecular functions(MF), and 4 were cellular components(CC) (Figs. 9a, c and Table 2). The top three items are integrin-mediated signaling pathways, integrin complexes, and cell adhesion molecule binding, respectively. Among the top 11 items enriched by KEGG, as shown in Figs. 9b, d and Table 3, focal adhesion, cell adhesion molecules, proteoglycans in cancer, small cell lung cancer, rap1 signaling pathway, intestinal immune network for IgA production, and microRNAs in cancer were reported to be related to the tumorigenesis and progression of GC.
Table 2
GO function enrichment analysis of ITGBs and adjacent genes in GC (Metascape).
Category
|
Term ID
|
Description
|
LogP
|
Count
|
Biological Processes
|
GO:0007229
|
integrin-mediated signaling pathway
|
54.880
|
24/108
|
Cellular Components
|
GO:0008305
|
integrin complex
|
52.469
|
19/31
|
Molecular Functions
|
GO:0050839
|
cell adhesion molecule binding
|
34.739
|
22/541
|
Biological Processes
|
GO:0050900
|
leukocyte migration
|
22.464
|
17/511
|
Biological Processes
|
GO:0034446
|
substrate adhesion-dependent cell spreading
|
15.081
|
9/108
|
Biological Processes
|
GO:0043113
|
receptor clustering
|
10.969
|
6/52
|
Biological Processes
|
GO:0001706
|
endoderm formation
|
19.185
|
6/53
|
Molecular Functions
|
GO:0043236
|
laminin binding
|
7.739
|
4/29
|
Cellular Components
|
GO:0031253
|
cell projection membrane
|
26.639
|
7/338
|
Cellular Components
|
GO:0098552
|
side of membrane
|
8.410
|
8/613
|
Biological Processes
|
GO:0033622
|
integrin activation
|
11.963
|
3/25
|
Biological Processes
|
GO:1900024
|
regulation of substrate adhesion-dependent cell spreading
|
4.605
|
3/57
|
Biological Processes
|
GO:0060284
|
regulation of cell development
|
12.286
|
5/486
|
Molecular Functions
|
GO:1990782
|
protein tyrosine kinase binding
|
15.044
|
3/98
|
Biological Processes
|
GO:0072593
|
reactive oxygen species metabolic process
|
10.018
|
4/281
|
Biological Processes
|
GO:0006935
|
chemotaxis
|
6.712
|
5/644
|
Biological Processes
|
GO:0007565
|
female pregnancy
|
9.880
|
3/177
|
Biological Processes
|
GO:0043588
|
skin development
|
11.408
|
4/415
|
Cellular Components
|
GO:0030426
|
growth cone
|
11.729
|
3/178
|
Biological Processes
|
GO:0001817
|
regulation of cytokine production
|
12.708
|
3/180
|
Table 3
KEGG function enrichment analysis of ITGBs and adjacent genes in GC (Metascape).
Category
|
Term ID
|
Description
|
LogP
|
Count
|
KEGG Pathway
|
hsa04510
|
Focal adhesion
|
36.971
|
20/199
|
KEGG Pathway
|
hsa04514
|
Cell adhesion molecules (CAMs)
|
20.246
|
12/144
|
KEGG Pathway
|
ko04640
|
Hematopoietic cell lineage
|
11.288
|
7/97
|
KEGG Pathway
|
ko05205
|
Proteoglycans in cancer
|
10.759
|
8/203
|
KEGG Pathway
|
hsa05222
|
Small cell lung cancer
|
9.678
|
6/84
|
KEGG Pathway
|
ko04015
|
Rap1 signaling pathway
|
8.920
|
7/210
|
KEGG Pathway
|
ko05418
|
Fluid shear stress and atherosclerosis
|
4.938
|
4/142
|
KEGG Pathway
|
hsa04672
|
Intestinal immune network for IgA production
|
4.803
|
3/49
|
KEGG Pathway
|
ko04610
|
Complement and coagulation cascades
|
9.863
|
3/79
|
KEGG Pathway
|
hsa04933
|
AGE-RAGE signaling pathway in diabetic complications
|
6.794
|
3/107
|
KEGG Pathway
|
hsa05206
|
MicroRNAs in cancer
|
7.931
|
3/307
|
The correlation between the expression level of ITGBs and the immuno-infiltrative level in GC
Immune cells have been reported to be closely associated with tumor proliferation and development. So we assessed the association between the expression levels of ITGBs and the immune infiltration degree of GC using the TIMER 2.0 database. The results demonstrated that ITGBs are participating in the body's immune response and immune cell infiltration, which in turn affect the prognosis of GC patients (Fig. 10).
The expression of ITGB1 and ITGB4 was strongly associated with the infiltration of CD4 + T cells, macrophages and dendritic cells, while the expression of ITGB2 and ITGB3 was positively linked to the infiltration of six immune cells. ITGB5 and ITGB6 expressions were positively correlated with the infiltration of B cells,CD4 + T cells and macrophages. ITGB7 expression was strongly associated with the infiltration of all six immune cells except B cells, while ITGB8 expression was only weakly correlated with the infiltration of B-cells.
Then we explored the clinical relevance of GC immune subset using Survival module in TIMER2.0, with the flexibility to correct for multiple co-variables in a multi-variate Cox proportional hazards model. The output of the Cox model showed that infiltration of macrophages (P = 0.017) and ITGB4 (P = 0.026) was statistically associated with clinical outcomes in GC patients (Table 4).
Table 4
The cox proportional hazard model of ITGBs and six tumor-infiltrating immune cells in GC (TIMER).
|
coef
|
HR
|
95%CI_l
|
95%CI_u
|
P value
|
sig
|
B_cell
|
3.887
|
48.779
|
0.525
|
4530.990
|
0.093
|
|
CD8_T cell
|
-1.104
|
0.332
|
0.017
|
6.378
|
0.464
|
|
CD4_T cell
|
-3.803
|
0.022
|
0.000
|
3.420
|
0.139
|
|
Macrophage
|
4.146
|
63.195
|
2.112
|
1891.057
|
0.017
|
*
|
Neutrophil
|
0.867
|
2.381
|
0.011
|
529.976
|
0.753
|
|
Dendritic
|
0.662
|
1.938
|
0.092
|
41.035
|
0.671
|
|
ITGB1
|
-0.034
|
0.967
|
0.688
|
1.360
|
0.847
|
|
ITGB2
|
0.008
|
1.008
|
0.781
|
1.301
|
0.951
|
|
ITGB3
|
0.059
|
1.061
|
0.838
|
1.344
|
0.622
|
|
ITGB4
|
-0.194
|
0.823
|
0.694
|
0.977
|
0.026
|
*
|
ITGB5
|
0.268
|
1.308
|
0.944
|
1.812
|
0.107
|
|
ITGB6
|
0.081
|
1.084
|
0.952
|
1.236
|
0.224
|
|
ITGB7
|
-0.014
|
0.987
|
0.759
|
1.282
|
0.919
|
|
ITGB8
|
-0.075
|
0.928
|
0.761
|
1.130
|
0.456
|
|
*P < 0.05. |