Glucoside xylosyltransferase 2 as a diagnostic and prognostic marker in gastric cancer via comprehensive analysis

Figures & data

Table 1. The main information of GC expression analyses from TCGA data and GEO datasets

Data set	Platform	# GC	# Normal	Country
GSE13195	GPL5175	25	25	China
GSE13911	GPL570	38	31	Italy
GSE19826	GPL570	12	12	China
GSE27342	GPL5175	80	80	USA
GSE33335	GPL5175	25	25	China
GSE79973	GPL570	10	10	China
GSE63089	GPL5175	45	45	China
GSE64951	GPL570	63	31	USA
GSE65801	GPL14550	32	32	China
GSE54129	GPL570	111	21	China
GSE84787	GPL17077	10	10	China
GSE112369	GPL15207	36	17	Japan
GSE118916	GPL15207	15	15	China
TCGA	NR	375	32	USA

Figure 1. The expression level of GXYLT2 was significantly up-regulated in gastric cancer samples

(a) GXYLT2 was highly expressed in GC tissues based on TCGA database and GEO database (except GSE64951). The results of forest plot (b) and funnel plot (c) showed no significant publication bias was observed during the meta-analysis of GXYLT2 expression in GC tissues.

Figure 2. GXYLT2 could be used as a diagnostic marker for patients with gastric cancer based on the results of analysis of the diagnostic values

(a) Nine GEO datasets and TCGA data indicated that GXYLT2 might have strong diagnostic values for gastric cancer (P < 0.05). Forest plot of sensitivity (b), forest plot of specificity (c), forest plot of positive likelihood ratio (d), forest plot of negative likelihood ratio (e), forest plot of diagnostic odds ratio (f), plot of symmetric summary receiver operating curve (g) of GXYLT2 for GC diagnosis.

Table 2. Relationships between GXYLT2expression and clinicopathological parameters in GC patients

Clinicopathological	GXYLT2 expression		P-value
parameters	High(n = 146)	Low (n = 147)
Age
<65	68	58	0.239
≥65	78	89
Gender
Female	62	54	0.341
Male	84	93
Tumor grade
G1 & G2	44	63	0.029
G3	102	84
Pathological stage
I & II	56	79	0.036
III & IV	85	73
T classification
T1-T2	29	47	0.023
T3-T4	117	100
Distant metastasis
M0	137	136	0.817
M1	9	11
Lymph node metastasis
N0-N1	81	87	0.522
N2-N3	65	60

Figure 3. GXYLT2 might serve as a prognostic marker in patients with GC

(a) Based on TCGA data, the overall survival of GXYLT2 patients was significantly reduced in the GXYLT2 High expression group (n = 184) compared with the GXYLT2 Low expression group (n = 184). (b) Univariate Cox regression analysis showed that age, TNM stage, lymph node metastasis, and GXYLT2 expression were significant factors affecting the survival time of GC patients. (c) Multivariate Cox regression analysis showed that age, gender, and GXYLT2 expression were reliable prognostic indicators in GC patients.

Figure 4. GSEA identified GXYLT2-related tumor pathways and immune regulatory. The terms related to tumor pathways were ‘ECM Receptor Interaction,’ ‘Focal Adhesion,’ ‘Wnt Signaling Pathway,’ ‘Gap Junction,’ and ‘Cell Adhesion Molecules (CAMs)’. The terms related to immunity were ‘Complement and Coagulation Cascades,’ ‘Cytokine Receptor Interaction,’ ‘TGF Beta Signaling Pathway,’ and ‘Leukocyte Transendothelial Migration’

Figure 5. Identification of differences and correlation between the GXYLT2 expression level and tumor-infiltrating immune cells in in GC patients

(a) The correlation analysis demonstrated the relationship between the GXYLT2 expression and immune cells infiltration. The sizes of dots showed the relationship extents. The red dots indicated the positive correlations, and the blue dots indicated the negative correlations. The numbers indicated the coefficients of correlation between genes expression and cells infiltration. (b) Violin plots showed that the proportional differentiation of 22 tumor-infiltrating immune cells in GC specimens with high and low GXYLT2 expression relative to the median GXYLT2 expression was performed using Wilcoxon rank-sum. (c) Eleven tumor immune cells were significantly associated with GXYLT2. The GXYLT2 expression level was correlated positively with five tumor-infiltrating immune cells (resting dendritic cells, m2 macrophages, monocytes, active NK cells and resting mast cells), and was negatively correlated with six tumor-infiltrating immune cells (plasma cells, activated memory CD4 T cells, resting NK cells, activated dendritic cells, and activated neutrophils and mast cells).

Figure 6. Verification of upregulation of GXYLT2 expression levels in GC by qRT-PCR and Immunohistochemistry test. (a)The mRNA levels of GXYLT2 in normal gastric cell and gastric cancer cell lines. (b) Representative images of GXYLT2 immunohistochemistry in GC tissue

Data accessibility

Publicly available datasets were analyzed in this study, these can be found in GEO database (https://www.ncbi.nlm.nih.gov/geo), and The Cancer Genome Atlas (https://portal.gdc.cancer.gov). The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.