The epigenetic study of Fgfr2 as a potential biomarker in the differential diagnosis of intestinal metaplasia and gastric cancer

Background Epigenetic alterations represent the potential role in the pathological process of cancer development. Epigenetic modulators affect some key gene methylation that can be applied as cancer biomarkers. Gastric cancer is faced with some limitations in diagnosis and differential diagnosis with intestinal metaplasia (IM). Intestinal metaplasia (IM) is generally considered a precancerous lesion in the carcinogenesis of gastric cancer cascade that is a major health burden worldwide. The broblast growth factor receptor 2 (FGFR2) gene is a receptor tyrosine kinase in which aberrant expression has a direct connection to GC. The purpose of the present study was to evaluate the prognostic relevance of FGFR2 methylation in the whole blood specimens obtained from patients with GC and IM and normal individual controls to examine the possible implication of epigenetic biomarker for differential diagnosis GC from IM. Material and Methods Appropriate epigenetic control regions in FGFR2 CpGs Island were specied by bioinformatic and differentially methylated regions (DMRs) enrichment analysis. The methylation aberration of FGFR2 selected CPG region was determined using MSRE-PCR and Real-time PCR on DNA extracted from blood samples of 125 participants, including 30 IM cases, 60 GC cases, and 35 normal controls individuals. Results A signicant FGFR2 hypomethylation has been obtained in IM (p = 0.01) and GC (p < 0.001) versus the normal control samples. ROC statistical analysis revealed sensitivity (96.67 %) and specicity (100 %) for FGFR2 as DNA epigenetic biomarker diagnostic of gastric cancer test with p < 0.001. These results suggest that the change in the methylation of FGFR2 (AUC = 0.97) is a promising epi-biomarker. Conclusions This study is the rst study to show that blood-based biomarkers FGFR2 gene may be a powerful epigenetic biomarker for diagnosing GC and IM and providing insights into gastric cancer pathogenesis and diagnosis. and GSE30601), including non-tumoral and gastric cancerous tissues, from the "NCBI GEO database" (Home - GEO - NCBI) and previous studies. A Venn diagram was plotted for hypomethylated (-1 < logFC < 0) genes using the online Venn diagram plotter tool (Draw Venn Diagram - Bioinformatics and Systems Biology) to identify the intersections between two datasets.


Background
Malignant Gastric adenocarcinoma (GAC), the most common type of Gastric cancer (GC), is one of the major causes of cancer-related deaths worldwide [1]. The GC at early stages is usually asymptomatic or manifests with vague symptoms. Advanced stages may be accompanied by persistent abdominal pain, anorexia, and irrational weight loss [2]. Although GC's morbidity and mortality rates have been reduced over the past decades, it remains the third most common cancer-related death worldwide [3]. In most cases, GC is developed from a multistep progression according to Correa's cascade [4] and is often initiated by H. pylori infection. Subsequent processes have been led to atrophic gastritis and intestinal metaplasia (IM). The IM is considered gastric cancers' precursors, and it is de ned as replacing the normal gastric mucosa with mucin-secreting intestinal mucosa [5].
Despite the advances in diagnosis and treatment of cancers but in regards to GC, the clinical outcome is still facing limitations because of its late diagnosis. Hence, a non-invasive effective method is required to be employed in the early diagnose of GC in patients. [6]. Since tumorigenesis is a stepwise accumulation of genetic and epigenetic alterations in oncogenes and tumor suppressor genes [7]. Recent advances in cancer epigenetic offer a better understanding of the underlying mechanism(s) of carcinogenesis and provide insight into the discovery of putative cancer biomarkers for early detection, disease monitoring, prognosis, and risk assessment [8]. Epigenetic modi cations are heritable changes that do not alter the DNA sequences. Aberrated epigenetic processes can lead to altered gene expression and malignant cellular transformation.
The broblast growth factor (FGF) receptor (FGFR) family is involved in various signi cantly affect tumorigenesis. Activation of these receptors can lead to activation of the RAS-MAPK pathway and the PI3K-AKT pathway, among others. The mechanisms by which FGFR can be misregulated vary between cancers. The FGFR1-4 are a family of receptor tyrosine kinases regulating fundamental processes, including cell proliferation, differentiation, migration, and survival [9,10].
The isoform IIIb of FGFR2 (known as keratinocyte growth factor receptor) is recognized as the regulator of epidermal differentiation and homeostasis in normal human tissues. Also, it is considered as an important tumor-suppressor role in-Vitro and -Vivo. The FGFR-2 IIIc isoform is expressed in epithelial and mesenchymal cells. The altered FGFR2 splicing and the aberrant expression of the mesenchymal FGFR2 IIIc isoform induce changes in the speci city for FGFs, leading to impairment in cellular differentiation, epithelial-mesenchymal transition (EMT), and tumor creation characteristics [11].
In this study, methylation aberration of FGFR2 promoter region was assessed in DNA extracted from blood samples WBC of patients with IM and GC and normal controls as a potential epigenetic biomarker.

Data Acquisition
The DMRs of gastric tumors and non-tumoral tissue checked by downloaded microarray datasets (GSE25869 and GSE30601), including non-tumoral and gastric cancerous tissues, from the "NCBI GEO database" (Home -GEO -NCBI) and previous studies. A Venn diagram was plotted for hypomethylated (-1 < logFC < 0) genes using the online Venn diagram plotter tool (Draw Venn Diagram -Bioinformatics and Systems Biology) to identify the intersections between two datasets.

Protein-Protein Interaction (PPI) Network Construction
Obtained genes uploaded to the String database (STRING: functional protein association networks) to construct a PPI network. Afterward, the constructed PPI network was analyzed using Cytoscape (Cytoscape: An Open-Source Platform for Complex Network, version: 3.8.2). Topological parameters of the PPI network, such as betweenness centrality, closeness centrality, and degree, were analyzed with the Network Analyzer. Enrichment Analyses (Gene Ontology) The selected gene was explored by the gene ontology (GO) analysis website (GOnet -DICE Tools) to understand the genes' function. The GO analysis was also undertaken from the biological process and molecular function categories.

Differentially Methylated Regions (DMRs) of chosen genes and designing primers
We investigated the regulator regions from the "GeneCards database" (GeneCards -Human Genes | Gene Database | Gene Search) to identify the promoter regions. The "SMART APP" (SMART -bioinfo-zs.com) is used to recognize the signi cant distinguished methylated region's accurate location. The SMART App explored and interpreted the DNA methylation data across 33 cancer types from the TCGA database [12]. Afterward, the restriction endonuclease recognition sites are determined by the "NEBcutter analysis tool" This study has been con rmed by the National Institute of Genetic Engineering and Biotechnology's ethical committee (Code number: IR.NIGEB.EC1398.12.3.A). About 4 mL of whole blood samples were collected in EDTA tubes from participants. The participant's clinicopathological features are presented in Table 1, including age, sex, and disease type. PCR condition was performed in 35 cycles, including strings denaturation (95°C for 30"), primer annealing (60°C for 30"), extension (72°C for 30"). Also, initial denaturation set at 95°C for 5' and a nal extension set at 72°C for 7' were considered. PCR products run on the 1.5% agarose gel, stained with ethidium bromide, and bands visualized under UV radiation.
Each treated sample with RE was compared with its undigested one as 100 % methylated due to no RE were used in these tubes to identify the accurate samples methylation. The intensity of the treated samples ampli ed products has a direct relationship with the methylation level. The methylation intensity was calculated by gel analyzer software (GelAnalyzer 19.1).
Digested and undigested DNA samples were also ampli ed using real-time PCR. All PCRs were performed in a Rotor-Gene 6000 thermal cycler (Corbett Life Science, Australia). Real-time PCR performed with the following constituents: 1 µL of DNA solution was added to 9 µL of a PCR mixture made up of 5 µL of 2× SYBR Green PCR Master Mix (Takara, Japan), 0.2 µL of forward and reverse primers, and 4.6 µL of water. PCR condition was performed in 40 cycles, including strings denaturation (95°C for 20"), primer annealing (61°C for 40"), extension (72°C for 60"). Also, initial denaturation and nal extension were set at 95°C for 2' and 72°C for 7'. ΔCt values determined as the difference between the obtained Ct values of the nontreated DNA and treated DNA (∆∆C T =∆C T (UN) -∆C T (D) ).

Statistical analysis
Results expressed as means ± standard deviation (SD), and all statistical analyses were performed using IBM SPSS version 26 (SPSS, Inc., Chicago, IL, USA) and Prism (GraphPad Software, San Diego, CA) version 9.0.0. The statistical signi cance of the difference between groups was determined using oneway ANOVA to determine the three groups' differences. P < 0.05 were considered statistically signi cant.

Results
Identi cation of DMRs in the GEO datasets Raw data were downloaded from two independent GEO datasets (GSE25869 and GSE30601); DMRs were identi ed using GEO2R. The top 250 hypomethylated regions have been selected from each dataset. Volcano plots displayed the differently methylated regions between tumoral samples and normal controls. Venn diagram indicated the intersections containing 68 DMRs were commonly hypomethylated in two datasets. As shown in Fig. 1, a regulatory network of obtained genes was constructed. The top 10 hub genes were EGFR, ERBB2, TJP1, YAP1, FOXA1, APP, GATA6, FGFR2, KLF4, and CXCL1. (Fig. 1) GO enrichment analyses The DMRs analysis was conducted using GO. The GO enrichment analyses indicated that DMRs were signi cantly related to the molecular function of "protein kinase binding" (p = 1.53E-06) and "transmembrane receptor protein tyrosine kinase activity" (p = 1.53E-06) and "transmembrane receptor protein tyrosine kinase activity" (p = 2.51E-06). Biological pathway results also showed that DMRs were mainly enriched in "positive regulation of protein kinase B signaling" (p = 8.00E-10) and "cellular response to growth factor stimulus" (p = 1.60E-09). (Fig. 2) Validation of FGFR2 with TCGA database The FGFR2 was chosen for validation according to the differentially methylated regions and the proteinprotein interaction result. The data of TCGA and GTEx (SMART -bioinfo-zs.com and New tab (cancerpku.cn)) revealed the methylation status of FGFR2 was signi cantly (p = 0.032) lower at the region of probe cg12835048 (chr10: 121598092, 121598093), and the expression was upregulated in GC compared to the normal gastric tissue (p < 0.05). (Fig. 3)

Clinicopathological parameters
The present study analyzed the association between the promoter methylation status of FGFR2 and clinicopathological characteristics. Participants consisted of 46.4% female and 53.6% male. The cohorts' age range consisted of 39-82 years old, 37-65 years old, and 34-83 for normal, IM, and GC cohorts, respectively. (Table 1).
Methylation analysis by MSRE-PCR method Implication of MSRE-PCR assessments of FGFR2 promoter methylation indicated a hypermethylation in normal samples (mean = 94.8 %) compared with that in IM (mean = 72.5 %) and in GC intestinal type (mean = 25.2 %) and diffuse (mean = 28.3 %) patients. Intriguingly, there is no signi cant difference between patients' methylation status with Intestinal type of GC versus diffuse type (Table 1, Fig. 5, Fig. 6-A).
A signi cant difference among FGFR2 promoter hypomethylation were detected among normal controls versus metaplasia (p < 0.01) and gastric cancer (p < 0.001) cases. The Receiver operating characteristic (ROC) curve assay of FGFR2 methylation levels in GC patients and normal controls (p < 0.001) revealed high sensitivity, speci city 96.67, 100%, respectively, with a cut-off < 77.50% and area under the ROC curve (AUC) of 0.9700. (Fig. 6-B) Conclusion This is the rst study based on bioinformatics and experimental assays to examine the methylations alteration of the FGFR2 promoter region in GC and IM compared to normal controls. FGFR2 has been studied as a potential epigenetic biomarker in peripheral blood WBCs as a non-invasive, reliable, costeffective tool to be employed in GC and IM diagnosis.
Signi cant methylation changes were observed among studied GC and IM cases compared to normal controls. The level of signi cances was p < 0.001 for GC, and p = 0.02 for IM. Also, signi cant differences were observed between the methylation status of GC and IM cases themselves (p < 0.001). Our results conveyed that the mean of methylation changes in the FGFR2 promoter was equal to 26.75 % in total GC cases, whereas it was 72.5 % in IM cases and 94.8 % in normal individuals in DNA samples extracted from WBCs.
This study's bioinformatics analysis is based on the TCGA database and represented methylation aberrations and mRNA expression pro ling data of GC samples. The obtained data were used as the training dataset to screen for signi cant DMRs. Interestingly, based on the TCGA data and NCBI GEO database, the FGFR2 in patients with GC has signi cant demethylation of the chosen FGFR2 promoter region (p = 0.032). As a result, the expression becomes upregulated consequently.
The high expressed levels of FGFR2 in colorectal [13], lung [14], gastric [15][16][17] cancer, coding mutations [18], and more recently, FGFR fusions that lead to pathway activation [19] have been demonstrated that the FGFR2 has a signi cant oncogenic potential across multiple cancer types. ROC analysis was employed to compare the predictive accuracy of the methylation status of patients with gastric cancer. Our ndings demonstrated the methylation status of FGFR2 (with a cut-off at < 77.50 %, p < 0.001) could re ect the presence of malignant gastric lesions.
The MSRE-PCR has some advantages in assessing the methylation of genes promoter over another common method that used bisul te DNA treatments. For instance, Methylation-speci c PCR (MS-PCR) uses bisul te conversion to detect methylated DNA sequences. In the ideal state, this conversion deaminates unmethylated cytosines to uracil but does not affect methylated cytosines. In the experimental condition, DNA manipulation and treatments may result in incomplete conversion, falsepositive extensive, and DNA degradation (of up to 90%) [20]. Because the WBCs extracted DNA's amount and quality are high, studies are shifting to assay WBCs DNA as a potential biomarker [21]. One study demonstrated a positive relationship between estrogen receptor alpha methylation in leukocytes and colonic tissue of patients with colorectal cancer (CRC) [22].
Gastric cancer is accompanying a high mortality rate and remaining an important cause of cancer death worldwide. The prognosis of GC is poor because cases are diagnosed at an advanced stage; hence, the treatment options are limited. Our ndings indicated that FGFR2 hypomethylation status in WBC DNA of patients with IM and GC signi cantly relates to lesions' malignancy. Also, the cornerstone of the study's approach is the non-invasiveness and preciseness of the utilized method. Hence, MSRE-PCR is used to evaluate the methylation status. In conclusion, we can deduce the methylation status of FGFR2 as an important prognostic biomarker of gastric cancer.

Declarations
Ethics approval and consent to participation The Ethical Committee of the National Institute of Genetic Engineering and Biotechnology (NIGEB) approved the study (Ethical code #: IR. NIGEB.EC1398.12.3.A). The informed consent form has been prepared and signed by all participants and the minors' parents under eighteen years old to use their clinical samples and personal data under their physician's supervision. All methods were carried out following relevant guidelines and regulations.

Availability of data and materials
The data generated during and/or analyzed during the current study are not publicly available, but are available from the corresponding author who was an organizer of the study.

Consent for publication
Not applicable.

Competing interests
All authors have read the manuscript and declared that they have no competing interest.
Authors' contributions SAA designed the concepts and methodology. AM, NA, and MA carried out the sampling and experimental laboratory work. AM performed the data acquisition, analyzing and interpretation. AM and SAA performed the manuscript writing and revisions. Administrative, technical, and material were supported by SAA. SAA supervised this study. All authors read and approved the nal manuscript.

Funding
Our study was supported by grants number 633 and 667 from the National Institute of Genetic Engineering and Biotechnology (NIGEB).