The Role of Early Growth Response 1 in the Prognostic Value and Cell Migration of Breast Cancer


 Background: Early growth response family members (EGRs), EGR1-4, have increasingly attracted attention in multiple cancers. However, the exact expression patterns and prognostic values of EGRs in the progress of breast cancer (BRCA) remain largely unknown. Methods: The mRNA expression and prognostic characteristics of EGRs were examined by the Cancer Genome Atlas (TCGA), Oncomine and Kaplan-Meier plotter. Enrichment analyses were conducted based on protein-protein interaction (PPI) network. The Tumor Immune Estimation Resource (TIMER) database and MethSurv were further explored. The protein expression level of EGR1 and cell migration were measured by Western blotting, immunohistochemistry, wound-healing assay and Boyden chamber assay in BRCA. Results: The transcriptional levels of EGR1/2/3 displayed significantly low expression in BRCA compared to that in normal tissues, while EGR4 was shown adverse expression pattern. Survival analysis revealed up-regulated EGR1-4 were remarkably associated with favorable relapse-free survival (RFS). A close correlation with specific tumor-infiltrating immune cells (TIICs) and several CpG sites of EGRs were exhibited. Immunohistochemistry assays showed that the protein expression of EGR1 was remarkably downregulated in BRCA compared to that in paracancerous tissues. Cell migration of MCF10A cells was increased after the silence of EGR1 by siRNA transfection.Conclusions: This study provides a novel insight to the role of EGR1 in the prognostic value and cell migration of BRCA.


Introduction
Breast cancer (BRCA) remains one of the widespread and main fatal malignancies in female diseases worldwide [1,2]. However, the overall survival (OS) and release-free survival (RFS) of patients with BRCA remain far from satisfaction [3]. Nevertheless, it is di cult for patients with high risk to be diagnosed timely in the early screen system and to be evaluated accurately before postoperative recurrence, owing to lack of reliable and e cient biomarkers [4]. Moreover, personalized treatments are increasingly concerned with the advent of precision medicine [5,6]. Therefore, the novel potential biomarkers for BRCA treatment need to pay more effort to explore.
Early growth response (EGR) gene family encompasses four family members: EGR1, EGR2, EGR3 and EGR4, locating on 5q31, 10q21, 8p21 and 2p13, respectively [7]. They are transcription factors that contain three highly conserved zinc nger domains in the C-terminus, which recognize GC-rich consensus sequences of the promoters of multiple target genes. Besides, four EGR proteins also contain a transcriptional activation domain in N-terminus [8].
EGR1 acts as an anti-oncogene engaging in multiple cancer processes, including cancer cell proliferation, apoptosis, migration and even affects tumor microenvironment [9,10]. EGR1 decreased cell growth through downregulating EPO-R transcription under hypoxia in non-small cell lung carcinoma [11]. EGR2 induces cell apoptosis via upregulating BNIP3L and BAK in a PETN-dependent manner [12]. EGR3 is also de ned as a tumor suppressor, which inhibits cell proliferation and induces apoptosis in hepatocellular carcinoma in vitro [13][14][15]. EGR4 is abundantly expressed in cholangiocarcinoma tissue and the low expression of EGR4 retards cell growth of cholangiocarcinoma [16].
Although a crowd of studies elucidate the mechanism of four members of the EGR family for plentiful types of cancers, the landscape of the prognostic value and role of EGR1 poorly explored in BRCA.
Currently, updated public databases based on integrative bioinformatics analysis of the Cancer Genome Atlas (TCGA) have signi cantly enhanced the e ciency of identi cation of biomarkers and functional genes in cancerous diseases [17][18][19]. Therefore, this study evaluates the transcriptional pro les and potential prognostic value of the EGR family by systematical bioinformatics analysis, and provides a novel role of EGR1 in the prognostic value and cell migration of BRCA.

Bioinformatics analysis
Oncomine analysis The mRNA expression of EGR1-4 of multiple cancers was retrieved from the Oncomine platform (https://www.oncomine.org/) [20]. The expression among different cancers could be presented on the heat map. The color presents mRNA expression of target genes with overexpression (red) or downexpression (blue).

TCGA data acquisition
The RNA-sequencing and clinical information of BRCA patients in TCGA dataset were downloaded from UCSC Xena (https://xena.ucsc.edu/). The level of gene expression was measured as log 2 (x+1) transformed RSEM normalized count. A total 1104 BRCA patients was included in our research. The relationship between EGRs expression and the clinical features were explored.

Kaplan-Meier plotter analysis
The prognostic value of EGR family members to RFS was analyzed by the Kaplan-Meier plotter (KM plotter) (http://www.kmplot.com/) [21]. The clinical outcome was displayed with hazard ratio (HR), 95% con dence interval (95%CI) and log-rank P value calculated by algorithms set in KM plotter.

Protein-protein interaction network construction and enrichment analysis
Protein-protein interaction (PPI) network was been constructed by GeneMANIA (https://genemania.org/) and visualized by Cytoscape 3.7.2 [22]. DAVID (https://david.ncifcrf.gov/) is a widely applied gene functional annotation tool [23]. In this study, DAVID was applied to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of EGRs and their co-operators. The human genome (Homo sapiens) was set as the background variables.

TIMER analysis
Tumor Immune Estimation Resource (TIMER, https://cistrome.shinyapps.io/timer/) is a bene cial tool to detect tumor-in ltrating immune cells (TIICs) via using the RNA-seq expression pro les, including B cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages and dendritic cells [24]. The association between immune in ltrates cells and the expression levels of EGR family members was detected through the TIMER platform, which was displayed by Pearson method.

MethSurv analysis
MethSurv (https://biit.cs.ut.ee/methsurv/) was used to explore the DNA methylation of EGR1-4 in TCGA [25]. The methylation levels and prognostic values of each CpG in EGR1-4 were analyzed. The patients were divided into low and high methylation group which were split at the best cut-off point. Western blotting analysis MCF10A, MDA-MD-231, MCF-7 and SUM1315 cell lines seeded into 60 mm dishes/24-well (Thermo Fisher) were washed with PBS and then lysed with 2 × SDS sample buffer. The lysates were harvested and abundant protein extracts were separated by 10% SDS-PAGE. The following antibodies were used anti-EGR1 (1:1000 dilution; catalog no. 55117-1-AP, Proteintech) and anti-β-actin (catalog no. AB21181, Bioworld). Protein levels were normalized to β-actin.

Wound healing assay
For wound healing analysis, MDA-MD-231 cells were seeded into 96-well plates (Lot No: 181122-078, BIOFIL) and grown to con uence. Then, the cells were transfected with siRNA using Lipofectamine 2000 Reagent for 6 h and were switched to fresh DMEM with 10%FBS to further cultured for 24-48 h. Next, the cell monolayers were wounded by a plastic pipette tip and rinsed with PBS to remove cell debris. The results were acquired at 0 h and 12 h after migration. The migratory area was counted by measuring the distance from the edge of the wound closure.

Boyden chamber assay
Cell migration was estimated in a modi ed Boyden chamber (Coster, Corning, NY), in which two chambers were separated by a polycarbonate membrane (8.0-μm pore diameter). The upper chamber membrane was rendered into single cell suspensions (1 × 10 5 cells) in serum-free DMEM supplied with 5 μg/mL BSA and the lower chamber was lled with DMEM with 10% FBS. The cells were allowed to migrate for 12 h at 37℃. Then, discard the medium, wash with PBS and xed the cells with 4% paraformaldehyde with PBS. The stationary upper-cells were dislodged with a cotton-tipped applicator and the lower chamber membrane was stained with 0.5% crystal violet. The approximate number of cells crossed over the membrane were counted by a microscope (Olympus Corporation, Tokyo, Japan).
The staining intensity was scored as 0-3: 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). The expression of EGR1 was assessed by immunoreactivity score (IRS) equaling to the percentages of positive cells multiplied with staining intensity. IRS was employed without prior knowledge of clinical response. Immunostained sections were scanned by a microscope (Olympus Corporation, Tokyo, Japan).

Statistical analysis
A series of statistical analyses were conducted through the bioinformatics database online. The GraphPad Primes 8.0 was used to analyze the TCGA data. Student's t-test and one-way ANOVA were used for the EGRs mRNA expression levels. Scatter plot charts show scatter plots and means ± SEM. Differences were considered signi cant if P values were less than 0.05 in all circumstances.

Results
The mRNA expression levels of EGR family across various cancers For the sake of understanding a pan-cancer view of EGRs' expression, the mRNA expression levels of EGR1-4 on the Oncomine were analyzed. The expressions of EGR1 and EGR3 in 20 different types of human cancers were downregulated compared to that in normal tissues, including BRCA, lung cancer, ovarian cancer ( Figure 1). These results indicated EGR1 and EGR3 might be tumor suppressors. However, the expression of EGR2 was not synchronous in different cancers ( Figure 1). Moreover, EGR1, EGR2 and EGR3 remarkably downregulated in BRCA tissues compared to those in normal tissues ( Figure 1). The mRNA expression level of EGR4 was absent in BRCA ( Figure 1). In total, the mRNA expression levels of EGR1/2/3 were negatively correlated with EGR4 and more studies should be devoted to explore the biological mechanism in various tumors. We further compared the transcriptional levels of EGRs in BRCA tissues with different ER/PR/HER2 status. We found EGR1 mRNA expression was increased in the ER + /PR + BRCA tissues, which was opposite to HER2 + tissues with decreased expression level of EGR1 (Figure 4a, e, i). The upregulated EGR3 was signi cantly associated to ER + /PR + status, but the downregulated EGR3 was signi cantly correlated to HER2 + status (Figure 4c, g, k). For EGR4, the relationship of mRNA level was signi cantly downregulated in BRCA tissues with ER + /PR + /HER2 + status (Figure 4d, h, l). However, the expression level of EGR2 was unrelated to ER/PR/HER2 status (Figure 4b, f, j). These results implied the transcriptional levels of EGRs were immensely related to clinical characteristics in BRCA and could be identi ed as potential biomarkers for the poor differentiation and metastasis status.  (Figure 5a-e). These results suggested EGR1-4 were associated with RFS, which could be considered as prospective biomarkers to predict survival times of BRCA patients.

PPI and enrichment analysis of EGR family
Under the knowledge of the potential values of EGRs for BRCA patients, a mutual PPI network of EGRs was constructed via GeneMANIA (Figure 6a). To acquire a functional understand, EGRs and their relevant genes were submitted for GO and KEGG analysis. The results showed that EGRs-related genes mainly participated in transcription from RNA polymerase II promoter, positive regulation of transcription from RNA polymerase II promoter, regulation of transcription and located in nucleus, nuclear chromatin, nucleoplasm. Also, they mediated transcriptional activator activity, DNA binding, transcription factor activity, and enriched in Hepatitis B, T cell receptor signaling pathway, B cell receptor signaling pathway and MAPK signaling pathway (Figure 6b). These data supplied the essential foundation for EGRs participating in the exploration of pathological mechanism and biological role of BRCA.

Prognostic values of EGR1-4 DNA methylation in MethSurv
MethSurv was employed to detect the DNA methylation levels of EGR1-4 and the prognostic value of each CpG in TCGA (Table 1). Eight CpGs of EGR1, seven CpGs of EGR2, three CpGs of EGR3, and two CpGs of EGR4 were relevant to meaningful prognostic impact. Cg19729803 of EGR1, cg12397802 of EGR2, cg13713148 of EGR3, and cg02287817 of EGR4 revealed the highest DNA methylation levels ( Figure 8a-d). These CpGs sites of EGRs were largely advantageous for the exploration of the biological mechanism of BRCA.

The low expression of EGR1 and its migration resistant role in BRCA
Based on numerous bioinformatics analyses of EGRs, we found that EGR1-4 showed distinct transcriptional expression level between BRCA and normal samples and presented signi cant prognostic value in RFS. Thus, we further examined the protein expression and role of EGR1 by a series of cytobiological experiments and immunostaining. IHC staining presented that the EGR1 was remarkably low expression in BRCA compared to that in normal tissues, which corresponded with the ndings from bioinformatics analysis (Figure 9a). Similarly, the expression level of EGR1 was notably decreased in MDA-MD231 and SUM1315 cells compared to mammary epithelial cell MCF10A, excepting MCF-7 ( Figure  9b).
Next, we measured cell migration of MCF10A and MDA-MD-231 cells after the transfection of siRNA-1, siRNA-2 and siRNA-3 targeting EGR1. The knockdown e ciency of siRNA-2 targeting EGR1 was the highest (Figure 9c). Boyden chamber assay exhibited that MCF10A had an increased migration capacity after EGR1 silence (Figure 9d, e). However, the knockdown of EGR1 had not alter the migration of MDA-MD-231 cells in wound healing assay (Figure 9f, g). To sum up, these ndings preliminarily suggested an anti-oncogene role of EGR1 in BRCA.

Discussion
Based on online databases, we discovered EGR1/2/3 expression levels were signi cantly downregulated, while EGR4 was upregulated in BRCA tissues. The prognostic values of EGR1-4 showed a positive relationship with better RFS of BRCA patients. Although accumulating evidences con rm EGRs regulate the initiation and/or development of multiple cancers, the expression pro le and prognostic value of EGR1-4 and the role of EGR1 in BRCA remain unclear [9,26]. According to experiment validations, our investigations found that EGR1 protein was highly expressed and resisted cell migration in BRCA. It is the rst time to systemically and comprehensively analyze the expression levels, potential prognosis, TIICs status and DNA methylation level of EGR1-4 in BRCA by bioinformatics methods. EGR1, considered as a tumor suppressor, is negatively associated with poor prognosis and early recurrence. Yang et al. reported overexpressed EGR1 repressed cell apoptosis and promoted cell proliferation by interacting with DNMT3L to inhibit the miR-195-AKT3 pathway in gastric cancer [27]. In this study, EGR1 was expressed at a remarkably lower level in BRCA tissues than that in normal breast tissues. Upregulated EGR1 mRNA expression was notably correlated with ER + /PR + status and the downregulation of EGR1 was associated with HER2 + status. The high expression of EGR1 exhibited a correlation with ne RFS. Crawford et al. found the expression of EGR1 was reduced in BRCA, which was in agreement with our results [28]. Besides, active EGR1 elevated PAC1 expression with excessive oxygen species, ultimately causing the chromatin remodeling mechanism of effector T cells [29]. Analogously, we found immune in ltrated cells were prominently related to the mRNA expression of EGR1 from the TIMER platform, such as B cell, CD8 + T cell, macrophage cell.
Owing to the signi cant difference of the transcriptional level, clinical characteristics, prognostic value, PPI, TIICs and DNA methylation of EGR1, we further explored the protein expression of EGR1 by Western blotting. Also, the role of EGR1 in cell migration was determined by wound-healing assay and Boyden chamber assay. Overexpressed miR-125b-2-3p notably increased lymphatic invasion and distant migration by targeting EGR1 in clear cell renal cell carcinoma [30]. Similarly, our result showed the cell migration of human mammary epithelial cell MCF10A was increased when EGR1 was silenced. However, the migration of MAD-MD-231 cell had no alteration after EGR1 downregulation.
EGR2 was strongly expressed in Ewing sarcoma compared with other solid tumors and the silence of EGR2 suppressed proliferation, clonogenicity and spheroidal growth of Ewing sarcoma in vitro [31]. In our study, EGR2 expression was decreased and high expression of EGR2 was related to favorable RFS, indicating its prognostic value in BRCA. However, EGR2 had no signi cant difference of PR -/+ /ER -/+ /HER2 -/+ status, which might need to further research. EGR3, frequently declined in hepatocellular carcinoma tissues, retarded cell proliferation and induced apoptosis in vitro [13]. The microarray data revealed decreased expression of EGR3 especially acted as a potential candidate gene for the diagnosis and prognosis of cutaneous squamous cell carcinoma [32]. Interestingly, our results displayed the upregulation of EGR3 was largely correlated with good RFS in BRCA.
EGR2 and EGR3 play important roles in adjusting the transition between proliferation and differentiation of effector CD4 + and CD8 + T cells [15,33]. In our report, EGR2 was strongly related to CD8 + T cell, CD4 + T cell, macrophage cell, neutrophil cell and dendritic cell. EGR3 presented a conspicuous association with immune in ltrates cells as well, like B cell, CD8 + T cell, CD4 + T cell and macrophage cell. Gong et al. found EGR4 facilitated tumor cell growth with high expression in cholangiocarcinoma [16]. Surprisingly, EGR4 was highly expressed and had a signi cantly negative association with ER + /PR + /HER2 + status. In BRCA, EGR4 expression presented a positive correlation with better RFS of BRCA patients. The biological function and molecular processes of EGR4 in cancers was still rarely discovered.
Up to now, the study of methylation of EGRs remains limited. In our analysis, the DNA methylation heat maps were clearly shown in all CpG islands. Moreover, DNA methylation levels in several EGRs CpG islands displayed signi cant association with prognosis of BRCA patients.

Conclusion
We systematically analyze the transcriptional levels and prognostic values of EGRs in BRCA via public databases. Our nding reveals that EGRs are possible to be novel prognostic biomarkers for BRCA patients. Besides, EGR2 and EGR3 are promising prognostic biomarkers for predicting RFS of BRCA patients. This study provides a comprehensive insight into the characteristic investigation of the EGR family and the role of EGR1 in the prognostic value and cell migration of BRCA.

Consent for publication
All authors approve of the manuscript.

Availability of data and materials
The datasets generated and analyzed during this study are available from the corresponding author on reasonable requirements.