Sage Journals: Discover world-class research

Abstract

Gastric cancer (GC) is a common cancer with high mortality and morbidity rates worldwide. Although medical and surgical treatments have improved, the mechanisms of the progression of GC remain unclear. Platelet-derived growth factor receptor- $\beta$ (PDGFRB) plays a pivotal role in angiogenesis and tumor cell proliferation and has been suggested as a prognostic marker of cancer. This study aimed to explore the relationship of PDGFRB expression with clinicopathologic characteristics, immune cell infiltration status, and prognosis in GC. In this study, we visualized the expression and prognostic values of PDGFRB in GC using the Oncomine, UALCAN, GEPIA, and Kaplan-Meier Plotter databases. And then we explored the potential relationships between PDGFRB expression and the levels of immune cell infiltration using the TIMER, GEPIA databases and CIBERSORT algorithm. Furthermore, LinkedOmics analysis was performed to explore the functions for PDGFRB. The results showed close correlations between PDGFRB and immune cell infiltration especially M2 Macrophage infiltration in GC. High PDGFRB expression was related to poor outcomes in GC. High PDGFRB expression can negatively affect GC prognosis by promoting angiogenesis and modulating the tumor immune microenvironment. These results strongly suggest that PDGFRB can be used as a prognostic biomarker of GC and provide novel insights into possible immunotherapeutic targets.

Keywords

Gastric cancer PDGFRB bioinformatics prognostic biomarker

1. Introduction

Gastric cancer (GC) is one of the most common cancers worldwide [1]. Although medical and surgical treatments have improved, survival rates remain poor because in many cases, GC is already at an advanced stage when diagnosed [2]. Immunotherapy has emerged as a promising cancer treatment with major successes in breast cancer [3], prostate cancer [4] and melanoma [5]. Recent studies have reported that several types of tumor-infiltrating immune cells are associated with the disease outcomes of various human cancers [6, 7, 8]. Immune cell infiltration in human gastrointestinal cancers is associated with cancer progression and is a reliable prognostic predictor [9, 10, 11, 12, 13]. Both cellular composition and organization of tumor-infiltrating immune cells are crucial for inhibiting cancer progression and are implicated in the success of cancer immunotherapy [9, 14].

Platelet-derived growth factor receptor beta (PDGFRB) encoded by PDGFRB gene [15], is a typical transmembrane tyrosine kinase receptor that promotes blood vessel formation when activated by platelet-derived growth factor (PDGF) [16] and controls many important cellular processes such as growth, proliferation, movement and survival [17, 18, 19, 20]. Its dysregulation is closely related to carcinogenesis, as well as to cardiovascular diseases [20]. Raja et al. (2017) suggested that PDGFRB may serve as a prognostic factor in GC by immunohistochemical expression [21]. Wang et al. (2019) reported that NRP1 and PDGFRB were significantly correlated with tumor malignant phenotypes serving as potential prognostic biomarkers for GC patients [22]. Although a series works revealed that the overexpression of PDGFRB correlates with the progression of gastric carcinoma [23] and high expression of epithelial PDGFRB is associated with poor disease-free survival (DFS) and overall survival (OS) in GC patients [24], there was rare study to explore the relationship between PDGFRB expression and the tumor immune microenvironment. This study aimed to explore the relationship of PDGFRB expression with clinicopathologic characteristics, immune cell infiltration status, and prognosis in GC through a bioinformatics analysis.

2. Material and methods

2.1 Oncomine database analysis

Oncomine1

¹
Oncomine: https://www.oncomine.org/resource/login.html.

has powerful analytical capabilities for calculating gene expression characteristics, clusters, and genomic modules [25]. We used it to determine the expression levels of PDGFRB in GC. We set the threshold to a

p

-value of 0.01, a fold change of 2 and the gene rank of the top 10%.

2.2 GEPIA database analysis

GEPIA2

²
GEPIA: http://gepia.cancer-pku.cn/index.html.

[26] is an interactive web to analyze the RNA sequencing expression, including 9,736 tumors and 8587 normal samples from TCGA and the GTEx projects [27, 28]. It was used to evaluate the expression of PDGFRB in stomach adenocarcinoma (STAD) and to generate DFS and OS curves based on the expression of PDGFRB and a log-rank test. We also calculated the correlation between PDGFRB expression and gene markers of tumor-infiltrating immune cells using related modules in GEPIA. The gene markers of tumor-infiltrating immune cells, including T cells (general), CD8

{}^{+}

T cells, T-helper 1 (Th1) cells, Th2 cells, Th17 cells, Tregs, exhausted T cells, monocytes, tumor-associated macrophages (TAMs), M1 macrophages, M2 macrophages, neutrophils, NK cells, and dendritic cells (DCs), are referenced in prior studies [29, 30]. The correlation module in GEPIA generated the expression scatter plots between pairs of userdefined genes in GC and estimated statistical significance and Spearman’s correlation. We analyzed the relationship between PDGFRB expression and other gene markers in tumor and normal tissues.

2.3 UALCAN database analysis

UALCAN3

³
UALCAN: http://ualcan.path.uab.edu.

[31] uses TCGA RNA-seq data to analyze the expression differences between tumor and normal tissue samples, as well as between sub-groups, including cancer stage, patient gender, tumor grade and other clinical characteristics. We used it to compare the expression levels of PDGFRB between sub-groups of stomach adenocarcinoma (STAD).

2.4 Kaplan-Meier plotter database analysis

Kaplan Meier plotter4

⁴
Kaplan Meier plotter: http://kmplot.com/analysis/.

is capable to assess the effect of 54k genes (mRNA, miRNA, protein) on survival in 21 cancer types including breast (

n=

6,234), ovarian (

n=

2,190), lung (

n=

3,452), and gastric (

n=

1,440) cancer, using both TCGA RNA-sequencing datasets and genechip datasets [32, 33]. We used it to calculate the correlation between PDGFRB expression and survival in GC. We set the threshold to a log-rank test

p

-value of 0.05.

2.5 TIMER database analysis

TIMER5

⁵
TIMER: https://cistrome.shinyapps.io/timer/.

database [34] is designed to analyze immune cell infiltrates in multiple cancers. The correlation between PDGFRB expression and immune cell infiltration in GC was analyzed using a gene module in the database. The correlations between the infiltrating levels of subsets of immune cells and PDGFRB expression were also calculated.

Figure 1.

Expression levels of PDGFRB in gastric cancer. (A) PDGFRB expression in different cancers compared to normal tissues in Oncomine. (B–H) Differential expression (DE) of PDGFRB in Oncomine. (I) DE of PDGFRB in GEPIA using TCGA datasets. (J) DE of PDGFRB in microarray data with accession number of GSE79973. (K) DE of PDGFRB in microarray data with accession number of GSE118916.

Figure 2.

PDGFRB transcription in various subgroups of GC patients (UALCAN). (A) Normal and STAD tissue samples. (B) Normal samples of both genders and male and female STAD patients. (C) Normal samples of any age and STAD patients aged 21–40, 41–60, 61–80, and 81–100 years. (D) Normal samples and STAD patients with tumor grades 1, 2, and 3. (E) Normal samples and African American, Caucasian, and Asian STAD patients. (F) Normal samples and STAD patients with regional lymph node metastasis status N0, N1, N2, and N3. (G) Normal samples and STAD patients with cancer stages 1, 2, 3, and 4. (H) Normal samples and patients with and without H. pylori infection. (I) Normal samples and patients with and without TP53 mutation ( ${}^{*}p<$ 0.01, ${}^{**}p<$ 0.01, ${}^{***}p<$ 0.001).

Figure 3.

Comparison of Kaplan-Meier and GEPIA survival curves of high and low PDGFRB expression in STAD. Low expression was associated with favorable OS in Kaplan-Meier Plotter using (A) GeneChip and (B) RNA-seq data. (C) Low expression was associated with favorable progression-free survival in Kaplan-Meier Plotter using GeneChip data. (D) Low expression was associated with favorable OS in GEPIA.

2.6 Quantification of tumor infiltration immune cells by CIBERSOFT algorithm

The CIBERSOFT method quantifies the relative scores for 22 human immune cell types with gene-based deconvolution algorithm and the characteristics of 547 marker genes. To enhance the robustness of the results, the algorithm applies Monte Carlo sampling to obtain the deconvoluted $P$ value for each sample [35]. The filter criteria of each sample is set as $p<$ 0.05, indicating that the predicted proportion of each infiltrating immune cell subtype is fairly accurate and suitable for further analysis.

2.7 Function exploration for PDGFRB

LinkedOmics6

⁶
www.linkedomics.org/login.php.

database is a web-based platform aiming at analyzing 32 TCGA cancer associated datasets [36]. As a module in LinkOmics, LinkFinder was applied to identify genes correlated to PDGFRB in TCGA STAD (

n=

415) using Pearson correlation coefficients. Then the top 50 significantly and positively correlated genes were used to perform the gene functional enrichment analysis by R package cluterProfiler [37].

P

value

<

0.05 was set as the significance criteria.

2.8 Validation datasets analysis

External datasets with accession numbers of GSE118916 and GSE79973 were downloaded from the Gene Expression Omnibus (GEO) database. Differentially expressed genes were identified using R package limma [38] with fold change (FC) $>$ 2 and $p<$ 0.01.

3. Results

3.1 Elevated expression of PDGFRB in GC

We searched Oncomine to study the differences in PDGFRB expression between different tumor tissues and normal tissues. The results showed significantly low expression levels of PDGFRB in bladder, cervical, ovarian and prostate cancer and sarcoma but significantly high expression levels of PDGFRB in colorectal, esophageal, pancreatic, head and neck cancer, GC, and leukemia (Fig. 1A). There were 7 datasets supporting the expression elevation for PDGFRB in GC (Fig. 1B–H). Then GEPIA with the RNA-seq datasets (Fig. 1I) and GEO with two independent microarray datasets of GSE79973 and GSE118916 (Fig. 2J and K) were used to evaluate differences for PDGFRB expression between GC and normal tissues. All the results showed significant overexpression of PDGFRB in GC. Furthermore, subgroup analysis of multiple clinicopathologic features of TCGA STAD samples in UALCAN consistently showed elevated transcriptional levels of PDGFRB. The expression of PDGFRB was significantly higher in STAD than in normal tissues in subgroup comparisons by gender, age, race, disease stage, tumor grade, Helicobacter pylori infection, and TP53 mutation (Fig. 2).

3.2 PDGFRB is a prognosis predictor in GC

The link between PDGFRB expression and patient outcomes in GC was investigated using Kaplan-Meier Plotter in which the genechip data (Fig. 3A and C) and RNA-seq data (Fig. 3B) were applied. Furthermore, GEPIA was also used to validate the survival analysis results using TCGA datasets (Fig. 3D). The results showed that high PDGFRB expression was associated with poor survival.

3.3 Correlation of PDGFRB expression with clinicopathological features of GC

High PDGFRB expression was found to be associated with poor prognosis in GC. The underlying mechanisms were explored using Kaplan-Meier Plotter. The correlations between PDGFRB and patient clinicopathologic characteristics were calculated. The results showed that PDGFRB upregulation was associated with worse OS and progression-free survival (PFS) in male and female patients, certain stages and TNM stages (stage N categories represent lymph node involvement), and two types of Lauren classification and differentiation ( $p<$ 0.05; Table 1). Specifically, high PDGFRB mRNA expression correlated with worse OS and PFS in GC stages 2–4 but not stage 1 (OS HR $=$ 1.97, $p=$ 0.23; PFS HR $=$ 0.5, $p=$ 0.27) or stage M1 (OS HR $=$ 1.62, $p=$ 0.1; PFS HR $=$ 1.27, $p=$ 0.43).

Table 1
Correlation of PDGFRB mRNA expression and clinical prognosis in gastric cancer with different clinicopathological factors by Kaplan-Meier Plotter

Clinicopathological characteristics	Overall survival (875)			Progression-free survival (640)
	N	Hazard ratio	$p$ value	N	Hazard ratio	$p$ value
Sex
Male	544	2.03 (1.57–2.62)	3.4e-08	437	1.92 (1.51–2.44)	5.9e-08
Female	236	2.65 (1.63–4.32)	4.8e-05	201	2.37 (1.52–3.71)	9.5e-05
Stage
1	67	1.97 (0.63–6.18)	0.23	60	0.5 (0.14–1.74)	0.27
2	140	1.93 (1.05–3.55)	0.032	131	1.77 (0.96–3.26)	0.063
3	305	2.21 (1.52–3.23)	2.20E-05	186	2.18 (1.5–3.18)	3.1e-05
4	148	1.77 (1.19–2.63)	0.0043	141	1.59 (1.08–2.34)	0.018
Stage T
2	241	2.46 (1.61–3.77)	1.90E-05	239	2.29 (1.51–3.47)	5.8e-05
3	204	1.59 (1.11–2.28)	0.011	204	1.48 (1.05–2.1)	0.026
4	38	3.61 (1.42–9.14)	0.0042	39	3.61 (1.54–8.44)	0.0016
Stage N
0	74	3.86 (1.14–13.06)	0.02	72	3.67 (1.08–12.42)	0.026
1	225	2.75 (1.76–4.29)	3.50E-06	222	2.62 (1.71–4.02)	4.5e-06
2	121	2.21 (1.41–3.49)	0.00044	125	2.22 (1.44–3.42)	0.00022
3	76	1.95 (1.13–3.34)	0.014	76	2.05 (1–4.2)	0.046
1 $+$ 2 $+$ 3	422	2.3 (1.74–3.05)	2.7e-09	423	2.04 (1.58–2.64)	3.2e-08
Stage M
0	444	2.27 (1.68–3.05)	3.3e-08	443	2.06 (1.57–2.7)	8.5e-08
1	56	1.62 (0.91–2.91)	0.1	56	1.27 (0.7–2.28)	0.43
Lauren classification
Intestinal	320	2.45 (1.77–3.41)	3.3e-08	263	2.38 (1.57–3.62)	2.4e-05
Diffuse	241	2.24 (1.53–3.29)	2.2e-05	231	2.1 (1.48–2.99)	2.1e-05
Differentiation
Poor	165	1.45 (0.97–2.18)	0.072	121	1.58 (0.95–2.65)	0.077
Moderate	67	1.55 (0.79–3.06)	0.2	67	1.78 (0.93–3.39)	0.075

${}^{\#}$ Bold values indicate $P<$ 0.05.

3.4 PDGFRB functional analysis results

We investigated the functions and pathways in which PDGFRB might participate. LinkedOmics was used to identify significantly and positively correlated genes correlated to PDGFRB using 415 STAD samples. The top 50 correlated genes including LRRC32, FBN1, COL6A3, COL8A1 and CDH11, were listed (Fig. 4A). Functional enrichment analysis for the top 50 genes showed that they were involved in extracellular matrix organization, cell adhesion, blood vessel development, angiogenesis, PI3K-Akt signaling pathway, protein digestion and absorption and pathways in cancer (Fig. 4B).

Figure 4.

Genes in significantly positive correlation with PDGFRB in gastric cancer. (A) Heatmap showing genes positively correlated with PDGFRB in STAD (top 50). (B) Functional enrichment for the top 50 coexpressed genes.

Figure 5.

Correlations between PDGFRB expression and (A) B cell, (B) CD8 ${}^{+}$ T cell, (C) CD4 ${}^{+}$ T cell, (D) macrophage, (E) neutrophil, and (F) DC immune infiltration levels in STAD in the TIMER database.

Figure 6.

Cumulative survival analysis for (A) B cells, (B) CD8 $+$ T cells, (C) CD4 $+$ T cells, (D) macrophages, (E) neutrophils, (F) DCs, and (G) PDGFRB.

3.5 Correlations of PDGFRB transcription levels with tumor immune infiltration and cumulative survival in GC from TIMER

We examined whether the transcription levels of PDGFRB in gastric tumors correlated with immune infiltration. TIMER was used to analyze the correlations between PDGFRB expression and STAD. The results showed that PDGFRB did not correlate with B cells but significantly positively correlated with the levels of immune infiltration of CD8 ${}^{+}$ T cells, CD4 ${}^{+}$ T cells, macrophages, neutrophils, and DCs (Fig. 5). Multivariable hazard models were used to evaluate the impact of PDGFRB expression on cumulative survival in the presence of several tumor-infiltrating immune cells using TIMER. PDGFRB and macrophages showed significant prognostic efficacy for GC, indicating that high PDGFRB and macrophage infiltration could predict poor survival (Fig. 6).

3.6 Correlation analysis of PDGFRB expression and immune marker sets

To investigate the relationships between PDGFRB and tumor-infiltrating immune cells, we calculated the correlations between PDGFRB and the immune marker sets of CD8 ${}^{+}$ T cells, CD4 ${}^{+}$ T cells, B cells, monocytes, TAMs, M1 and M2 macrophages, neutrophils, NK cells, and DCs in STAD in both TIMER and GEPIA (Table 2, Fig. 7). We also analyzed functional T cells Th1, Th2, Tfh, Th17, Tregs, and exhausted T cells. After correlation adjustment by purity, the results showed that PDGFRB expression significantly correlated with most immune marker sets of monocytes, TAMs, M2 macrophages, and Tregs (Table 2). Chemokine (C-C motif) ligand (CCL-2), CD68, and IL10 of TAMs and CD163, VSIG4, and MS4A4A of the M2 phenotype significantly correlated with PDGFRB expression in STAD ( $P<$ 0.0001; Fig. 7). We further analyzed the correlations between PDGFRB expression and the markers of monocytes and TAMs in GEPIA. The results were similar to those obtained from TIMER (Table 3). These findings suggest that PDGFRB may regulate macrophage polarization in STAD. High PDGFRB expression correlated with high infiltration levels of DCs in STAD. DC markers such as HLA-DPB1, CD1C, ITGAX, and NRP1 also showed significant correlations with PDGFRB expression, indicating the close relationship between PDGFRB and DC infiltration.

Figure 7.

Scatterplots of the correlations between PDGFRB expression and gene markers of (A) monocytes, (B) TAMs, (C) M1 macrophages, and (D) M2 macrophages.

Table 2

Correlation analysis between PDGFRB and related gene markers of immune cells in TIMER

Description	Gene marker	None		Purity
		Rho	$p$	Rho	$p$
CD8 $+$ T cell	CD8A	0.333	3.41E-12	0.319	2.03E-10
	CD8B	0.162	9.51E-04	0.152	3.01E-03
T cell	CD2	0.331	4.52E-12	0.326	7.78E-11
	CD3D	0.259	8.39E-08	0.244	1.59E-06
	CD3E	0.297	6.79E-10	0.286	1.41E-08
B cell	CD19	0.274	1.33E-08	0.271	8.50E-08
	CD79A	0.282	5.30E-09	0.257	3.77E-07
Monocyte	CD86	0.489	2.77E-26	0.482	1.89E-23
	CSF1R	0.617	7.66E-45	0.606	2.73E-39
TAM	CCL2	0.493	7.60E-27	0.48	3.18E-23
	CD68	0.367	1.04E-14	0.345	4.83E-12
	IL10	0.499	1.76E-27	0.492	1.91E-24
M1 macrophage	NOS2	0.069	0.162	0.072	0.159
	PTGS2	0.247	3.30E-07	0.246	1.22E-06
	IRF5	0.375	2.82E-15	0.363	2.87E-13
M2 macrophage	CD163	0.556	4.40E-35	0.539	5.53E-30
	MS4A4A	0.559	1.87E-35	0.549	3.26E-31
	VSIG4	0.537	2.17E-32	0.536	1.63E-29
Neutrophils	CCR7	0.381	9.02E-16	0.374	4.70E-14
	CEACAM8	0.006	0.901	0.03	0.566
	ITGAM	0.562	7.17E-36	0.554	6.27E-32
Natrual killer cells	KIR2DL1	0.163	8.35E-04	0.172	7.92E-04
	KIR2DL3	0.082	0.095	0.071	0.17
	KIR2DL4	0.022	0.65	0.001	0.986
	KIR2DS4	0.089	0.069	0.077	0.136
	KIR3DL1	0.13	7.94E-03	0.119	0.021
	KIR3DL2	0.153	1.81E-03	0.149	3.61E-03
	KIR3DL3	$-$ 0.068	0.168	$-$ 0.052	0.316
Dendritic cell	HLA-DPA1	0.305	2.15E-10	0.292	6.68E-09
	HLA-DPB1	0.358	4.98E-14	0.346	4.51E-12
	HLA-DQB1	0.24	7.84E-07	0.218	1.94E-05
	HLA-DRA	0.286	2.78E-09	0.276	4.56E-08
	CD1C	0.36	3.57E-14	0.352	1.64E-12
	ITGAX	0.565	2.35E-36	0.56	1.16E-32
	NRP1	0.737	3.56E-72	0.723	1.78E-62
Th1	IFNG	0.079	0.106	0.084	0.103
	STAT1	0.197	5.35E-05	0.199	9.86E-05
	STAT4	0.371	5.06E-15	0.373	6.36E-14
	TBX21	0.344	5.51E-13	0.348	3.25E-12
	TNF	0.195	6.18E-05	0.178	5.04E-04
Th2	GATA3	0.343	6.92E-13	0.337	1.54E-11
	STAT5A	0.52	4.35E-30	0.517	2.87E-27
	STAT6	0.249	2.74E-07	0.251	7.32E-07
	IL13	0.165	7.49E-04	0.204	6.14E-05
Tfh	BCL6	0.464	1.50E-23	0.44	2.08E-19
	IL21	0.131	7.32E-03	0.128	0.013
Th17	IL17A	$-$ 0.137	5.34E-03	$-$ 0.147	4.08E-03
	STAT3	0.503	5.51E-28	0.491	1.93E-24
Treg	CCR8	0.509	1.02E-28	0.511	1.45E-26
	FOXP3	0.421	2.88E-19	0.413	4.39E-17
	STAT5B	0.578	2.01E-38	0.576	6.40E-35
	TGFB1	0.684	1.35E-58	0.676	5.90E-52
T cell exhaustion	CTLA4	0.247	3.58E-07	0.242	1.87E-06
	GZMB	0.169	5.34E-04	0.147	4.25E-03
	HAVCR2	0.528	3.46E-31	0.525	2.99E-28
	LAG3	0.227	3.01E-06	0.217	2.05E-05
	PDCD1	0.286	3.11E-09	0.286	1.36E-08

Bold values indicate $P<$ 0.01.

Table 3

Correlation analysis between PDGFRB and related gene markers of monocyte and macrophages in GEPIA

Description	Gene markers	Tumor		Normal
		STAD
		R	$P$ value	R	$P$ value
Monocyte	CD86	0.51	1.5e-28	$-$ 0.095	5.8E-01
	CD115(CSF1R)	0.65	1.4e-49	0.28	9.7E-02
TAM	CCL2	0.48	1.8e-25	0.53	9.4E-04
	CD68	0.39	2.8e-16	$-$ 0.32	6.2E-02
	IL10	0.55	5.2e-33	0.11	5.1E-01
M1 macrophage	INOS (NOS2)	0.082	9.6E-02	0.35	3.9E-02
	IRF5	0.39	6e-16	$-$ 0.36	2.9E-02
	COX2 (PTGS2)	0.29	1.9e-09	0.77	4.1e-07
M2 macrophage	CD163	0.5	2.1e-27	0.72	1.7e-06
	VSIG4	0.55	8.1e-34	0.54	7.6E-04
	MS4A4A	0.59	4.2e-40	0.58	2e-04

3.7 Tumor immune infiltration cell components prediction

Using CIBERSORT, the 22 immune cell components for GC samples in TCGA were infered (Fig. 8A). The results showed that macrophages M2 was the most abundant immune cell across the samples (Fig. 8B). PDGFRB expression had correlation with M2 macrophages ( $r=$ 0.27, $p=$ 0.0009) (Fig. 8C) and the results also revealed that M2 macrophages were significantly upregulated in the group with high expression of PDGFRB ( $p=$ 4.721e-06) (Fig. 8D), indicating that PDGFRB may influence the polarization of M2 macrophages.

Figure 8.

The immune cell components predicted by CIBERSORT. (A) The barplot of the immune cell components. (B) The boxplot of the immune cell components. (C) Macrophages M2 components are plotted according to PDGFRB expression level. (D) The scatterplot of the PDGFRB expression and the macrophages M2 components.

4. Discussion

Although Researchers have been trying to explore the biomarkers in GC [39, 40, 41, 42], the prognosis biomarker of GC patients remains poor Platelet-derived growth factor receptors play a critical role in angiogenesis and tumor cell proliferation and metastasis [43]. PDGFRB promotes proliferation and migration of cancer cells and angiogenesis, and has been used as a target in several cancer treatments [19, 44]. In this study, differential expression of PDGFRB between cancer and normal tissues was observed in GC. The results demonstrated that PDGFRB expression was remarkably increased in GC tissues. Abnormal gene expression may be associated with tumorigenesis and prognosis of patients [45].

GEPIA and Kaplan-Meier Plotter analyses revealed that high PDGFRB expression correlates with poor prognosis in GC. Both of them showed that high PDGFRB expression correlates with a high HR for OS and PFS in GC. High PDGFRB expression correlates with poor prognosis in GC stages 2–4, T2–T4, and N1–N3. The highest HRs for OS and PFS were observed when PDGFRB was highly expressed. High PDGFRB expression is most likely to impact the prognosis of GC patients with lymph node metastasis. This finding is in line with a previous study reporting that PDGFRB overexpression correlate with cancer progression and lymphogenous metastasis of gastric carcinoma [46], suggesting that PDGFRB may serve as an oncogene to promote gastric carcinogenesis. To study the molecular functions and underlying mechanisms of PDGFRB in gastric carcinogenesis, we performed coexpression analysis for PDGFRB in GC. Functional analysis of PDGFRB and its positively coexpressed genes showed that they are involved in extracellular matrix organization, cell adhesion, blood vessel development and angiogenesis which play important roles in carcinogenesis and cancer progression [47]. Controlled remodeling of the extracellular matrix is essential for the growth, invasion, and metastasis of malignant tumors [48]. Tumorigenesis-related pathways such as the “PI3K-Akt signaling pathway” was also identified which was widely reported to play a crucial role in cancer occurrence by promoting cell proliferation and inhibiting apoptosis [49].

The analysis of the correlation between immune infiltration levels, various immune marker sets, and PDGFRB expression provides insights into the potential significance of PDGFRB for tumor immunology and its use as a cancer biomarker. Gene markers of M1 macrophages, such as PTGS2 and NOS2, showed weak correlations with PDGFRB expression, whereas M2 markers, such as CD163, VSIG4, and MS4A4A, showed strong correlations ( $r>=$ 0.5, $p<$ 0.01; Table 2). Similar results were found by the correlation analysis between PDGFRB expression level and the macrophages M2 components using TCGA data ( $r>=$ 0.5, $p<$ 0.01; Table 3). All these analyses suggested that in GC microenvironment, macrophages exhibit M2 characteristics and may promote angiogenic activity, tumor growth and metastasis [48, 50, 51, 52]. It has been reported that M2 macrophages are independent prognostic factors for colorectal cancer [53, 54] and that high levels are associated with poor prognosis of breast cancer [55]. We used multivariable hazard models to investigate the impact of PDGFRB expression on OS in the presence of several tumor-infiltrating immune cells. The results showed that besides PDGFRB, high macrophage infiltration is associated with poor survival, implying that macrophages are another prognostic factor for GC. Furthermore, the results of CIBERSORT analysis showed that the proportion of M2 macrophages in the high PDGFRB expression group was remarkably increased compared with that in the low PDGFRB expression group (Fig. 8D) and there is a significant positive correlation between the proportion of M2 macrophages and the PDGFRB expression level (Fig. 8C). These results indicate the potential role of PDGFRB in modulating macrophage activation and polarization into type M2 TAMs, which contribute to tumor progression. Genin et al. revealed that M2 macrophages played a protective role in cancer cell apoptosis induced by etoposide, contributing to decreasing cancer cell apoptosis [56]. And preventing the differentiation from M0 macrophages to M2 macrophages exerted anti-tumor effects via inhibiting the invasion of breast cancer cells induced by macrophages [57].

Our results also indicate that PDGFRB has the potential to activate Tregs and induce T cell exhaustion. PDGFRB upregulation positively correlates with the expression of Treg and T cell exhaustion markers (FOXP3, CCR8, STAT5B, TIM-3, PD-1, CTLA4, and LAG3). TIM-3, also called HAVCR2, a crucial exhausted T cell surface protein, highly correlates with PDGFRB expression in STAD, playing a crucial role in tumor immunological tolerance [58]. Furthermore, we found significant correlations between PDGFRB expression and the regulation of several markers of Th cells (Th1, Th2, Tfh, and Th17). These correlations could be indicative of a potential mechanism whereby PDGFRB regulates T cell functions in STAD. Our study provides insights into possible mechanisms that might explain why PDGFRB expression correlates with immune cell infiltration and poor prognosis in GC. Although this study indicated that PDGFRB plays an important role in the recruitment and regulation of infiltrating immune cells in GC, the immune reaction is a complex and dynamic process. Therefore, further study is necessary to explore the role of PDGFRB involvement in the immune response in GC.

5. Conclusion

In summary, we demonstrated that overexpression of PDGFRB was associated with clinicalpathological parameters and predicted a poor prognosis. High PDGFRB expression was closely associated with increased immune infiltration levels of CD8 ${}^{+}$ and CD4 ${}^{+}$ T cells, macrophages, neutrophils, and DCs in GC. Moreover, PDGFRB expression possibly participates in the regulation of TAMs, Tregs, and T cell exhaustion. Our study revealed the potential role of PDGFRB in tumor immunology and its prognostic value. PDGFRB might be used as a potential prognostic biomarker and therapeutic target in GC.

Authors’ contributions

Conception: Baohong Liu, Yongliang Lou and Lingling Zhao.

Interpretation or analysis of data: Baohong Liu, Xingxing Xiao and Ziqin Lin.

Preparation of the manuscript: Baohong Liu.

Revision for important intellectual content: Yongliang Lou and Lingling Zhao.

Supervision: Yongliang Lou and Lingling Zhao.

Availability of data and materials

The analyzed data sets generated during the study are available from the corresponding author on reasonable request.

Funding

This work was supported by the National Natural Science Foundation of China [grant number 82002117]; open fund of Wenzhou Key Laboratory of Sanitary Microbiology [grant number ZD202003KF02].

Declaration of competing interest

The authors declare that they have no competing interests.

References

Bray

et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA Cancer J Clin 68(6) (2018), 394–424.

Abbas

et al., Current and future biomarkers in gastric cancer, Biomed Pharmacother 103 (2018), 1688–1700.

Nathan

M.R.

and Schmid

, The emerging world of breast cancer immunotherapy, Breast 37 (2018), 200–206.

Thomas-Kaskel

A.K.

et al., Immunotherapy with dendritic cells for prostate cancer, Int J Cancer 121(3) (2007), 467–473.

Vesely

M.D.

and Chen

, Normalization cancer immunotherapy for melanoma, J Invest Dermatol 140(6) (2020), 1134–1142.

Zhang

et al., CCR5 is associated with immune cell infiltration and prognosis of lung cancer, J Thorac Oncol 14(5) (2019), e102–e103.

Muppa

et al., Immune cell infiltration may be a key determinant of long-term survival in small cell lung cancer, J Thorac Oncol 14(7) (2019), 1286–1295.

Desmedt

et al., Immune infiltration in invasive lobular breast cancer, J Natl Cancer Inst 110(7) (2018), 768–776.

Chen

L.J.

et al., B7-H4 expression associates with cancer progression and predicts patient’s survival in human esophageal squamous cell carcinoma, Cancer Immunol Immunother 60(7) (2011), 1047–1055.

10.

Chen

L.J.

et al., Higher numbers of T-bet(+) intratumoral lymphoid cells correlate with better survival in gastric cancer, Cancer Immunol Immunother 62(3) (2013), 553–561.

11.

et al., Higher numbers of T-bet+ tumor-infiltrating lymphocytes associate with better survival in human epithelial ovarian cancer, Cell Physiol Biochem 41(2) (2017), 475–483.

12.

Zhu

et al., T-bet and eomesodermin are required for T cell-mediated antitumor immune responses, J Immunol 185(6) (2010), 3174–3183.

13.

Pimenta

E.M.

and Barnes

B.J.

, Role of tertiary lymphoid structures (TLS) in anti-tumor immunity: potential tumor-induced cytokines/chemokines that regulate TLS formation in epithelial-derived cancers, Cancers (Basel) 6(2) (2014), 969–997.

14.

Singer

et al., A distinct gene module for dysfunction uncoupled from activation in tumor-infiltrating T cells, Cell 171(5) (2017), 1221–1223.

15.

Steller

E.J.

et al., PDGFRB promotes liver metastasis formation of mesenchymal-like colorectal tumor cells, Neoplasia 15(2) (2013), 204–217.

16.

Andrae

Gallini

and Betsholtz

, Role of platelet-derived growth factors in physiology and medicine, Genes Dev 22(10) (2008), 1276–1312.

17.

Donnem

et al., Co-expression of PDGF-B and VEGFR-3 strongly correlates with lymph node metastasis and poor survival in non-small-cell lung cancer, Ann Oncol 21(2) (2010), 223–231.

18.

Calzolari

et al., Tumor progression and oncogene addiction in a PDGF-B-induced model of gliomagenesis, Neoplasia 10(12) (2008), 1373–1382, following 1382.

19.

Lev

D.C.

et al., Inhibition of platelet-derived growth factor receptor signaling restricts the growth of human breast cancer in the bone of nude mice, Clin Cancer Res 11(1) (2005), 306–314.

20.

Ustach

and Kim

H.R.

, Platelet-derived growth factor signaling and human cancer, J Biochem Mol Biol 36(1) (2003), 49–59.

21.

Raja

U.M.

et al., Immunohistochemical expression and localization of cytokines/chemokines/growth factors in gastric cancer, Cytokine 89 (2017), 82–90.

22.

Wang

et al., Hypomethylated gene NRP1 is co-expressed with PDGFRB and associated with poor overall survival in gastric cancer patients, Biomed Pharmacother 111 (2019), 1334–1341.

23.

Chen

et al., Candidate genes in gastric cancer identified by constructing a weighted gene co-expression network, PeerJ 6 (2018), e4692.

24.

Wang

J.X.

et al., GLI2 induces PDGFRB expression and modulates cancer stem cell properties of gastric cancer, Eur Rev Med Pharmacol Sci 21(17) (2017), 3857–3865.

25.

Rhodes

D.R.

et al., Oncomine 3.0: genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles, Neoplasia 9(2) (2007), 166–180.

26.

Tang

et al., GEPIA: A web server for cancer and normal gene expression profiling and interactive analyses, Nucleic Acids Res 45(W1) (2017), W98–W102.

27.

Carithers

L.J.

and Moore

H.M.

, The genotype-tissue expression (GTEx) project, Biopreserv Biobank 13(5) (2015), 307–308.

28.

Consortium

G.T.

, The genotype-tissue expression (GTEx) project, Nat Genet 45(6) (2013), 580–585.

29.

Pan

J.H.

et al., LAYN is a prognostic biomarker and correlated with immune infiltrates in gastric and colon cancers, Front Immunol 10 (2019), 6.

30.

Xiao

et al., TGFbeta2 is a prognostic-related biomarker and correlated with immune infiltrates in gastric cancer, J Cell Mol Med 24(13) (2020), 7151–7162.

31.

Chandrashekar

D.S.

et al., UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses, Neoplasia 19(8) (2017), 649–658.

32.

Szasz

A.M.

et al., Cross-validation of survival associated biomarkers in gastric cancer using transcriptomic data of 1,065 patients, Oncotarget 7(31) (2016), 49322–49333.

33.

Nagy

et al., Validation of miRNA prognostic power in hepatocellular carcinoma using expression data of independent datasets, Sci Rep 8(1) (2018), 9227.

34.

et al., TIMER: A web server for comprehensive analysis of tumor-infiltrating immune cells, Cancer Res 77(21) (2017), e108–e110.

35.

Newman

A.M.

et al., Robust enumeration of cell subsets from tissue expression profiles, Nat Methods 12(5) (2015), 453–457.

36.

Vasaikar

S.V.

et al., LinkedOmics: Analyzing multi-omics data within and across 32 cancer types, Nucleic Acids Res 46(D1) (2018), D956–D963.

37.

et al., clusterProfiler: An R package for comparing biological themes among gene clusters, OMICS 16(5) (2012), 284–287.

38.

Smyth

G.K.

, Linear models and empirical bayes methods for assessing differential expression in microarray experiments, Stat Appl Genet Mol Biol 3 (2004), Article3.

39.

Xia

et al., Combined detection of exosome concentration and tumor markers in gastric cancer, J Biomed Nanotechnol 16(2) (2020), 252–258.

40.

et al., Identifying multiple collagen gene family members as potential gastric cancer biomarkers using integrated bioinformatics analysis, PeerJ 8 (2020), e9123.

41.

D.M.

et al., Significant function and research progress of biomarkers in gastric cancer, Oncol Lett 19(1) (2020), 17–29.

42.

Yang

C.M.

et al., Circular RNAs in gastric cancer: Biomarkers for early diagnosis, Oncol Lett 20(1) (2020), 465–473.

43.

Tsao

A.S.

et al., Immunohistochemical overexpression of platelet-derived growth factor receptor-beta (PDGFR-beta) is associated with PDGFRB gene copy number gain in sarcomatoid non-small-cell lung cancer, Clin Lung Cancer 12(6) (2011), 369–374.

44.

Ostman

and Heldin

C.H.

, PDGF receptors as targets in tumor treatment, Advances in Cancer Research 97 (2007), 247.

45.

Sheng

et al., Overexpression of KIF2A is suppressed by miR-206 and associated with poor prognosis in ovarian cancer, Cell Physiol Biochem 50(3) (2018), 810–822.

46.

Guo

et al., Clinicopathological significance of platelet-derived growth factor B, platelet-derived growth factor receptor-beta, and E-cadherin expression in gastric carcinoma, Contemp Oncol (Pozn) 17(2) (2013), 150–155.

47.

Smirnova

O.V.

Ostroukhova

T.Y.

and Bogorad

R.L.

, JAK-STAT pathway in carcinogenesis: is it relevant to cholangiocarcinoma progression, World J Gastroenterol 13(48) (2007), 6478–6491.

48.

et al., Tumor-associated macrophages induce invasion and poor prognosis in human gastric cancer in a cyclooxygenase-2/MMP9-dependent manner, Am J Transl Res 11(9) (2019), 6040–6054.

49.

Martini

et al., PI3K/AKT signaling pathway and cancer: An updated review, Ann Med 46(6) (2014), 372–383.

50.

Riabov

et al., Role of tumor associated macrophages in tumor angiogenesis and lymphangiogenesis, Front Physiol 5 (2014), 75.

51.

Galdiero

M.R.

et al., Tumor associated macrophages and neutrophils in cancer, Immunobiology 218(11) (2013), 1402–1410.

52.

Tang

, Tumor-associated macrophages as potential diagnostic and prognostic biomarkers in breast cancer, Cancer Lett 332(1) (2013), 3–10.

53.

et al., Profiles of immune cell infiltration and immune-related genes in the tumor microenvironment of colorectal cancer, Biomed Pharmacother 118 (2019), 109228.

54.

Xiong

et al., Profiles of immune infiltration in colorectal cancer and their clinical significant: A gene expression-based study, Cancer Med 7(9) (2018), 4496–4508.

55.

Ali

H.R.

et al., Patterns of immune infiltration in breast cancer and their clinical implications: A gene-expression-based retrospective study, PLoS Med 13(12) (2016), e1002194.

56.

Genin

et al., M1 and M2 macrophages derived from THP-1 cells differentially modulate the response of cancer cells to etoposide, BMC Cancer 15 (2015), 577.

57.

Hodges

et al., Mast cells, disease and gastrointestinal cancer: A comprehensive review of recent findings, Transl Gastrointest Cancer 1(2) (2012), 138–150.

58.

Cheng

et al., Expression of Tim-3 in gastric cancer tissue and its relationship with prognosis, Int J Clin Exp Pathol 8(8) (2015), 9452–9457.

PDGFRB is a potential prognostic biomarker and correlated with immune infiltrates in gastric cancer

Abstract

Keywords

1. Introduction

2. Material and methods

2.1 Oncomine database analysis

1 Oncomine: https://www.oncomine.org/resource/login.html.

2 GEPIA: http://gepia.cancer-pku.cn/index.html.

3 UALCAN: http://ualcan.path.uab.edu.

4 Kaplan Meier plotter: http://kmplot.com/analysis/.

5 TIMER: https://cistrome.shinyapps.io/timer/.

2.7 Function exploration for PDGFRB

6 www.linkedomics.org/login.php.

3. Results

3.1 Elevated expression of PDGFRB in GC

3.2 PDGFRB is a prognosis predictor in GC

3.3 Correlation of PDGFRB expression with clinicopathological features of GC

Table 1 Correlation of PDGFRB mRNA expression and clinical prognosis in gastric cancer with different clinicopathological factors by Kaplan-Meier Plotter

3.6 Correlation analysis of PDGFRB expression and immune marker sets

5. Conclusion

Authors’ contributions

Availability of data and materials

Funding

Declaration of competing interest

References

¹
Oncomine: https://www.oncomine.org/resource/login.html.

²
GEPIA: http://gepia.cancer-pku.cn/index.html.

³
UALCAN: http://ualcan.path.uab.edu.

⁴
Kaplan Meier plotter: http://kmplot.com/analysis/.

⁵
TIMER: https://cistrome.shinyapps.io/timer/.

⁶
www.linkedomics.org/login.php.

Table 1
Correlation of PDGFRB mRNA expression and clinical prognosis in gastric cancer with different clinicopathological factors by Kaplan-Meier Plotter