Introduction

Gastric cancer (GC) is the fourth most common cancer and the third most frequent cause of cancer-related death worldwide (723,000 deaths in 2012)1. Despite improvements in the diagnosis of GC at an early stage and the availability of new anticancer agents, the 5-year survival rate of patients with advanced GC is only 25–30%2. Increased efforts to eradicate Helicobacter pylori in Asian countries are expected to reduce the incidence of GC in the middle or lower gastric tract3,4. In contrast, the incidence of the intestinal type of differentiated GC located in the upper stomach or esophagogastric junction will likely increase worldwide5. Such tumors have relatively high incidence of hematogenous metastasis than GC in the middle or lower gastric tract. Therefore, effective management of hematogenous metastasis of GC is raised as an important clinical issue that must be resolved urgently. For the first important step to achieve this goal, discovery of sensitive and specific biomarkers for hematogenous metastasis is required to identify patients at high risk.

GC metastasizes through three dominant processes; hematogenous, lymphatic and direct dissemination from the serosal surface. Among them, hematogenous metastasis requires a distinctive multistep process involving vascular invasion, detachment from a primary site, survival of tumor cells in hypoxic portal blood, tissue engraftment, evasion of the hepatic immune system, and colonization6,7. The application of next-generation sequencing technology reveals that an underlying molecular signature of a tumor cell’s ability to metastasize varies according to the metastatic process and target organs8,9,10,11. We hypothesized that metastasis from primary GC via hematogenous route employs a specific mechanism that can be exploited to identify specific biomarkers for hematogenous metastasis of GC.

Therefore, in the present study, we conducted global expression profiling according to the metastatic route to identify molecules specific for hematogenous metastasis and show that the G protein-coupled receptor 155 (GPR155) may serve as a candidate biomarker.

Results

Identification of Candidate Marker

We sought candidate markers according to the following sieving criteria (Fig. S1). Firstly, this time we targeted molecules with decreased expression levels in GC tissues compared with the corresponding noncancerous mucosal tissues. Secondly, we focused on the ability which had been already got in the primary lesion to metastasize via hematogenous route. Then, we imposed the condition that there were no differences in gene expression in the primary lesion and hepatic metastases. As a result, the transcriptome analysis identified 21 candidate genes related to hematogenous metastasis (Table 1). Among them, the largest decrease in expression in GC tissues compared with noncancerous mucosae was that of GPR155. G protein-coupled receptor (GPCR) family have been reported to contribute to tumor progression through interactions with cancer-related signaling pathways. However, there are no reports to our knowledge of an association of GPR155 with GC. Therefore, we focused on GPR155 in this study.

Table 1 Candidate tumor suppressor genes downregulated in gastric cancer tissues of patients with hepatic metastasis.
Full size table

Expression of GPR155 and Potentially Interacting Molecules

Reduced expression of GPR155 mRNA was detected in all GC cell lines compared with FHs74Int (Fig. 1A). MKN1, MKN45 and N87 are cell lines which were established from hepatic metastasis. GPR155 mRNA expression was strongly suppressed in MKN45 and N87 cells. Little is known about cancer-related molecules that interact with GPR15512. To address this gap in our knowledge, we conducted PCR array analysis and found that the expression level of GPR155 mRNA correlated inversely with those of TWIST1 and WNT5B mRNAs (Fig. 1B and Fig. S2A). The expression levels of GPR155, TWIST1, and WNT5B were not related to the state of differentiation of GC cell lines (Fig. S2A). Pathway analysis of AGS cells indicated that inhibition of GPR155 mRNA expression increased the levels of p-ERK1/2 and p-STAT1 by >50% (Fig. 1C and Fig. S3).

Figure 1: Analysis of GPR155 expression and identification of candidate molecules that interact with GPR155: assays used indicate downstream effects.
figure 1

(A) Reduced expression of GPR155 mRNA was detected in all GC cell lines compared with that of FHs74Int. Copy number loss was detected in AGS and SC-6-JCK cells. Methylation of the promoter region of GPR155 was not detected. (B) PCR array analysis showing that the expression level of GPR155 mRNA inversely correlated with those of TWIST1 and WNT5B mRNAs. (C) Cell signaling pathway analysis of AGS cells indicated that inhibition of GPR155 mRNA increased the levels of p-ERK1/2 and p-STAT1.

Full size image

Analysis of Mechanisms that Inhibit GPR155 Expression

We conducted bisulfite sequencing to investigate whether DNA methylation inhibits GPR155 transcription. Methylation of cytosine residues within the promoter region of GPR155 was not detected in the GC cell lines or FHs74Int, (Fig. 1A and Fig. S2B). TaqMan Copy Number Assays for identification of possible regulatory mechanisms of GPR155 expression other than DNA methylation detected copy number loss in AGS and SC-6-JCK cells but not in FHs74Int cells (Fig. 1A).

Effects of GPR155 Knockdown on GC Cell Activities

We inhibited GPR155 mRNA expression by transfecting AGS and MKN1 cells with siGPR155 (Fig. S2C) to determine the influence of GPR155 on cell phenotype and the contribution of GPR155 to untransfected AGS and MKN1 those expressed relatively high levels of GPR155. We evaluated the proliferation, migration, and invasion of AGS and MKN1 cells. GPR155 knockdown significantly increased cell proliferation on day 7 compared with untransfected and siControl cells both in AGS and MKN1 cells (Fig. 2A and Fig. S4A). Moreover, the number GC cells invading the matrigel were increased in cells transfected with siGPR155 both in AGS and MKN1 cells (Fig. 2B and Fig. S4B). There was no significant change in migration of AGS cells when GPR155 expression was inhibited (Fig. 2C), whereas the migration ability of MKN1 cells was increased by knockdown of GPR155 expression (Fig. S4C).

Figure 2: Phenotypes of AGS cells transfected with siGPR155.
figure 2

(A,B) Knockdown of GPR155 expression increased cell proliferation and invasiveness. (C) There was no significant change in cell migration.

Full size image

Diagnostic Performance of GPR155 Expression Levels

Primary GC tissues expressed significantly lower levels of GPR155 mRNA compared with the corresponding noncancerous mucosal tissues (Fig. 3A). Even in tissues of patients with Stage I GC, GPR155 mRNA expression was reduced by approximately 10-fold. GPR155 mRNA levels in GC tissues were significantly lower in stage IV patients with synchronous hematogenous metastasis compared with those without synchronous hematogenous metastasis. By looking into patients with stage II/III GC, GPR155 mRNA levels were lower in GC tissues from patients with metachronous hematogenous metastasis compared to those from patients without metachronous hematogenous metastasis, though the difference was not statistically significant (Fig. 3A).

Figure 3: Expression of GPR155 in clinical samples, and the cumulative incidence of hematogenous recurrence in patients with Stage II/II GC.
figure 3

(A) Significantly lower levels of GPR155 mRNA were detected in primary GC tissues compared with the corresponding noncancerous mucosal tissues. Patients with Stage IV GC with synchronous hematogenous metastasis had lower expression levels of GPR155 mRNA compared with those of patients without hematogenous metastasis. (B) The optimal cut-off value of GPR155 expression was determined at 0.0009. (C) The cumulative incidence of hematogenous recurrence was significantly higher in the low GPR155 group patients with Stage II/III GC. Abbreviations: Hem-rec; hematogenous recurrence, Hem; synchronous hematogenous metastasis.

Full size image

The optimal cut-off value of the GPR155 mRNA level in GC tissues was determined at 0.0009 having a modest correlation (AUC = 0.684) with synchronous and metachronous hematogenous metastasis by the receiver operating characteristic curve analysis in all 200 GC patients (Fig. 3B). Using this cut-off value, we stratified patients into low GPR155 (GPR155 < 0.0009) and high GPR155 (GPR155 ≥ 0.0009) groups. Analysis of the association between the GPR155 mRNA level and clinicopathological factors shows that low GPR155 mRNA levels in GC tissue significantly associated with age ≥ 65 years, male sex, macroscopic type (other than Borrmann type4/5), low T stage, differentiation, expansive growth type, and synchronous hematogenous metastasis (liver n = 11, lung n = 2, bone n = 1 and brain n = 1) (Table 2).

Table 2 Association between GPR155 mRNA expression levels and clinicopathological characteristics of 200 patients with gastric cancer.
Full size table

The cumulative incidence of hematogenous recurrence was significantly higher in the low GPR155 group (Fig. 3C), while the incidence of postoperative adjuvant chemotherapy (S-1 monotherapy) was not different between two groups (Table 2). Hematogenous recurrences were found at the liver (n = 9), bone (n = 1), brain (n = 1) and ovary (n = 1). Otherwise, there was no significant difference in overall survival between the two groups in all Stages (5-year survival rates, 53% and 59%, respectively). Similarly, in patients with Stage II/III GC with or without hematogenous recurrence with significant differences in GPR155 mRNA levels, there were no significant differences in overall survival (Fig. 4B) and disease-free survival (Fig. 4C) between the low and high GPR155 groups. Multivariate analysis identified low GPR155 mRNA expression levels in GC tissue as an independent predictive factor of synchronous hematogenous metastasis and metachronous hematogenous metastasis after curative gastrectomy (hazard ratio, 5.38; P = 0.001), together with CEA > 5 ng/ml, vessel invasion, and expansive growth (Table 3).

Figure 4: IHC analysis of GPR155 expression in tissues of representative patients with GC and the survival curve of patients with Stage II/III GC.
figure 4

(A) GPR155 staining intensity of GC tissues was generally low compared with the corresponding noncancerous mucosae. GPR155 staining intensity correlated significantly with synchronous and metachronous hematogenous metastasis. (B) There was not significant difference in overall survival between low GPR155 and high GPR155 groups with Stage II/III GC. (C) There was no significant difference in disease-free survival between the low GPR155 and high GPR155 groups with Stage II/III GC. Abbreviations: T; tumor tissue, N; non-cancerous component, GC; primary gastric cancer tissue, Hem/Hem-rec; hematogenous metastasis/recurrence.

Full size image
Table 3 Predictive factors for hematogenous metastasis/recurrence in 200 patients with gastric cancer.
Full size table

Association between Hematogenous Metastasis and in situ Expression of GPR155 Protein

We performed Immunohistochemical (IHC) to verify whether the expression of GPR155 was significant for diagnosing and predicting hematogenous metastasis. GPR155 staining intensity in GC tissues was lower compared with the corresponding noncancerous mucosae in most cases. Moreover, lower levels of GPR155 correlated significantly with synchronous hematogenous metastasis and metachronous hematogenous metastasis that occurred after curative gastrectomy (Fig. 4A). Strong (focal or diffuse) staining TWIST and WNT5B tended to be found in patients with suppressed GPR155 at GC components (Fig. 5).

Figure 5: Immunohistochemical staining for GPR155, TWIST and WNT5B proteins.
figure 5

(A) Results in a representative case using tissues from primary GC and hepatic metastasis. (B) Correlations in staining patterns between GPR155, TWIST and WNT5B. Abbreviations: T; tumor tissue, N; non-cancerous component, GC; primary gastric cancer tissue.

Full size image

Discussion

We performed transcriptome analysis to identify hematogenous metastasis-specific biomarkers of GC. Among candidate molecules identified by our transcriptome analysis, the largest difference in expression levels between primary GC tissue and the corresponding noncancerous mucosal was exhibited by GPR155 mRNA. In contrast, GPR155 mRNA was expressed at equivalent levels in primary GC and hepatic metastatic lesions. These findings indicate that decreased expression of GPR155 may reflect the inherent metastatic potential of GC cells in the primary tumor that was not acquired during metastasis via hematogenous route.

GPR155, which resides on chromosome 2q31.1, comprises eighteen exons and encodes a 97 kDa protein. GPR155 is a member of the seven-transmembrane domain of the GPCR family13. Ligand binding activates the guanine nucleotide exchange factor activity of GPCRs that exchange GDP for GTP on its associated G protein. The Gα subunit bound to GTP dissociates from the Gβ and Gγ subunits to activate intracellular signaling proteins or target proteins directly. Limited information is available on the ligands for GPCRs at present. Huang XP, et al. detected some ligands for other GPCR families, GPR68 and GPR6514. GPCRs mediate diverse physiological processes such as the visual sensing, immune function, cell proliferation, and tumor metastasis15,16. It is therefore not surprising GPCRs represent 30–50% of the targets of currently marketed therapeutic drugs17,18,19,20. GPR155 is a unique member of GPCRs and there have been only a few reports on involvement in malignancies, such as follicular type papillary thyroid carcinoma and colorectal cancer21,22. In those earlier studies, GPR155 was listed in the results of microarray or proteomic analysis, and no data on the function and clinical significance of GPR155 was presented.

In the present study, qRT-PCR revealed varying levels of GPR155 mRNA expression level in GC cell lines independent of their differentiation phenotypes. In all GC cell lines tested here, GPR155 mRNA expression was reduced compared with that of the human intestinal epithelial cell line FHs74Int. To seek for the regulatory mechanisms of GPR155 transcription, DNA methylation of the CpG island of a promoter region prevents transcription23, and the promoter region of GPR155 harbors a CpG island; however, we were unable to detect promoter methylation in the genomes of any of the GC cell lines studied here. Therefore, we investigated copy number loss as a possible secondary suppression mechanism. Copy number loss was detected in 20% of the GC cell lines, which suggests the possibility that copy number loss may contribute to the suppression of GPR155 mRNA expression. However, histone methylation, micro-RNAs, transcription suppressors and RNA editing may contribute to the suppression of GRP155 expression, and therefore, further investigations are mandatory to elucidate the regulatory mechanisms of GPR155 expression24,25.

As little evidence is available on the oncological roles of GPR155, we performed PCR array and cell signaling pathway analyses to identify cancer-related molecules that potentially interact with GPR155. Although statistical power was not strong due to number of analyzed cell lines, we detected an inverse correlation between GPR155 mRNA expression and those of TWIST1 and WNT5B mRNA. TWIST1 promotes metastasis through its effects on the epithelial-mesenchymal transition and promotes the formation of distant metastasis in GC26,27,28,29. WNT5B is associated with tumor formation and malignant transformation in GC, breast cancer, and squamous cell carcinoma of the head and neck30,31,32,33. The correlation between the expression of GPR155 and these molecules suggests that the suppression of GPR155 expression interferes with oncogenic signaling pathways. To test this hypothesis, we investigated the downstream effect of inhibiting GPR155 expression on signaling transduction pathways that mediate cell proliferation. We detected increased p-ERK1/2 and p-STAT1 compared with controls. Gα subunit are divided into several subtypes, in which there are two major subtypes, Gsα activating adenylyl cyclase (AC) and Giα suppressing AC34. To date, the type of Gα subunit coupling to GPR155 has not been specified. Our result suggests that the Gα subunit coupling to GPR155 may be Giα subunit which suppresses AC.

We selected two GC cell lines expressing relatively high GPR155 mRNA from differentiated type for assays of cell phenotype, AGS having copy number loss and MKN1 established from hepatic metastasis. In both GC cell lines, GPR155 knockdown led to significant increases in cell proliferation and invasion, indicating that GPR155 has downstream effect to cancer-related signaling pathways and therefore acts as a tumor suppressor. Studies on the effect of GPR155 knockdown on apoptosis and the cell cycle may reveal how GPR155 affects cell proliferation. Using cells cultured under hypoxic conditions or in suspension may reveal how GPR155 participates in the mechanism of hematogenous metastasis. Further, mouse xenograft models might provide information on tissue engraftment, invasion, and colonization via hematogenous route.

The most important finding of the present study was that GPR155 expression demonstrated high diagnostic and predictive performance for hematogenous metastasis of GC. GPR155 mRNA levels in GC tissues were significantly decreased in all stages, including Stage I, compared with the corresponding normal mucosa. While GPR155 mRNA expression significantly decreased in GC tissue, it decreased further in patients with synchronous hematogenous metastasis compared to those without. These findings indicate that GPR155 represents a hematogenous metastasis-specific biomarker for GC. Low differentiation, serosal invasion, diffuse type, young age and infiltrating growth are well-known risk factors for peritoneal dissemination35,36,37,38, and vascular invasion, advanced age, differentiation, Borrmann type 1 or 2, expansive growth have been reported to be risk factors for hepatic metastasis, most frequent hematogenous metastasis of GC35,36,39,40. The clinicopathological factors associated with reduced GPR155 expression reported here conflict with known risk factors for peritoneal dissemination and are, however, associated with known risk factors for hepatic metastasis. Moreover, multivariate analysis revealed that suppression of GPR155 mRNA expression was an independent prognostic factor for synchronous hematogenous metastasis and metachronous hematogenous metastasis that occurred after curative gastrectomy. These clinicopathological analyses support the hypothesis that GPR155 is specifically associated with hematogenous metastasis of GC and that detecting reduced levels of GPR155 mRNA in primary GC tissue is useful for the diagnosis of synchronous hematogenous metastasis as well as for determining a patient’s risk of hematogenous recurrence after curative gastrectomy. On the other hand, lower expression of GPR155 mRNA in GC tissues was not associated with overall or disease-free survival. Nevertheless, the cumulative incidence of hematogenous recurrence in patients with Stage II/III GC was significantly higher in the low GPR155 group. This is rational result, because GPR155 expression level is specific for hematogenous metastasis and survivals are significantly influenced by other metastatic patterns, peritoneal dissemination or distant lymph node metastasis.

Translating our results into clinical practice, physicians can stratify GC patients according to the risk for hematogenous metastasis by performing qRT-PCR or IHC analysis of biopsy or surgical specimens of primary tumor. For patients at high risk of hematogenous metastasis, appropriate management may be provided according to careful preoperative examinations, and postoperative surveillance focusing on hematogenous metastasis will facilitate early detection and therapeutic intervention. These measures will likely contribute to improve the outcomes of patients with GC.

This study has certain limitations. First, this study was limited by the relatively small sample size. The clinical significance and the cut-off value of GPR155 expression level as a hematogenous metastasis-specific marker of GC should be evaluated in a larger patient cohort. Second, further analyses of putative interacting molecules indicated here are required to identify the molecular mechanisms underlying the biological activities of GPR155 in patients with GC. Third, enforced expression of GPR155 is required for further evaluation of the function of GPR155 in GC. Finally, the mechanism of GPR155 expression in GC remains to be identified.

Our results indicate that GPR155 is a biomarker that is useful for the diagnosis and prediction of hematogenous metastasis of patients with GC.

Methods

Transcriptome Analysis

Surgically resected specimens of four patients with GC with synchronous hepatic metastasis without metastasis to the peritoneal cavity or distant lymph nodes were subjected to transcriptome analysis. Global expression profiling was conducted using the HiSeq System (Illumina, San Diego, CA, USA) to compare the expression levels of 57,751 genes in primary GC tissues, the corresponding noncancerous adjacent gastric mucosae, and hepatic metastases.

Sample Collection

We used GC cell lines, which were obtained from the Japanese Collection of Research Bioresources Cell Bank (Osaka, Japan) as follows: MKN1, MKN45, MKN74, NUGC2, NUGC3, NUGC4, and SC-6-JCK. The GC cell lines AGS, KATOIII, and N87 were purchased from the American Type Culture Collection (ATCC) (Manassas, VA, USA). The human intestinal epithelial cell line FHs74Int (ATCC) served as a nontumorigenic control. Primary GC tissues and corresponding noncancerous mucosal tissues were collected from 200 patients who underwent gastrectomy at Nagoya University Hospital between 2001 and 2014. None of the patients underwent preoperative chemotherapy.

The methods were carried out in accordance with relevant guidelines. The study protocol was approved by the Medical Ethics Committee of the Nagoya University Hospital, protocol No. 2014–0043. Informed consent was obtained from all patients. Written informed consent for the use of clinical samples and data, as required by the Institutional Review Board, was obtained from all patients.

Quantitative Real-Time Reverse-Transcription Polymerase Chain Reaction (qRT-PCR) and PCR Array Analysis

GPR155 mRNA levels were determined using qRT-PCR. Total RNAs (10 μg) were extracted from GC cell lines, FHs74Int, and 200 pairs of clinical samples and were amplified using GPR155-specific primers (Table 4). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA (TaqMan, GAPDH control reagents, Applied Biosystems, Foster City, CA, USA) was quantified in each sample for standardization (Table 4). GPR155 mRNA expression level was determined as the value of GPR155 divided by that of GAPDH. The qRT-PCR protocol was performed as previously described41. To identify genes encoding putative GPR155-interacting proteins, we used the Human Epithelial to Mesenchymal Transition (EMT) RT2 Profiler PCR Array (Qiagen, Chatsworth, CA, USA)42.

Table 4 Sequences of Primers and siRNAs.
Full size table

DNA Methylation Analysis and Copy Number Analysis

We used CpG Island Searcher software (http://cpgislands.usc.edu/)43,44 to detect predicted CpG islands in the GPR155 promoter region (chr2:174486637–174487426). Genomic DNAs of the cell lines were treated with bisulfite, and bisulfite sequence analysis was performed as previously described (Table 4)45. Using purified genomic DNA isolated from GC cell lines, DNA copy numbers were determined using TaqMan Copy Number Assays (Applied Biosystems). The assays were as follows: upstream (assay ID: Hs01092594_cn, 175351658 within exon 1), midstream (assay ID: Hs01971174_cn, 175335170 within exon 6), and downstream (assay ID: Mn00059996_cn, 73351855 overlaps intron 14 and exon 14). Copy number alteration in the GPR155 locus were determined using CopyCaller™ Software (Life Technologies, Carlsbad, CA, USA)46.

Inhibition of GPR155 Expression

Small interfering RNAs (siRNAs) specific for GPR155 mRNA (siGPR155) (Table 4) (Hokkaido System Science, Sapporo, Japan) were used to transfect AGS cells. AccuTargetTM Negative Control siRNA Fluorescein-labeled (Cosmo Bio Co. Ltd., Tokyo, Japan) served as a control nontargeting siRNA (siControl). AGS and MKN1 cells were seeded to grow to 60–80% confluence 24 h later and transfected with siRNAs using LipoTrust EX Oligo (Hokkaido System Science) as previously described47. After 72 h incubation following siRNA transfection, total RNAs and proteins were extracted. For assays of cell phenotype, the transfected cells were treated with EDTA-trypsin.

Cell Signaling Pathway Analysis

We used the PathScan® Intracellular Signaling Array Kit (Cell Signaling Technology, Beverly, MA, USA), according to the manufacturer’s protocol, to investigate the downstream effects of inhibiting GPR155 expression on cell signaling pathways in AGS cells.

Assays of Cell Phenotype

The proliferation of AGS and MKN1 cells transfected with siGPR155 was evaluated using a Premix WST-1 Cell Proliferation Assay System (Takara Bio Inc., Kusatsu, Japan). Cells (5 × 103 cells per well) were seeded into 96-well plates in DMEM supplemented with 2% FBS. The optical density of each well was measured in six replicates consist of six technical and two biological replicates on days 0, 1, 3, 5, and 7 after seeding. The ability of cells to invade Matrigel was determined using BioCoat Matrigel Invasion Chambers (BD Biosciences, Bedford, MA, USA) according to the manufacturer’s protocol. Cells (2.5 × 104 cells per well) were seeded into the upper well of the chamber in serum-free DMEM, and DMEM supplemented with 10% FBS was supplied into the lower well. After an appropriate incubation time (AGS; 48 h, MKN1; 36 h), we used a light microscope to count the cells present on the surface of the membrane in eight randomly selected fields. Migration was determined using a published wound-healing assay48. Cells (2 × 104 cells per well) were seeded into 12-well plates in serum-free DMEM using the ibidi Culture insert method (ibidi, Martinsried, Germany) to establish wound gaps of a defined width. After 24 h, the insert was removed, and the width of the wound was measured at 100-μm intervals (20 per well, 40× magnification).

Immunohistochemical Analysis

IHC was performed to determine the localization of GPR155, TWIST1 and WNT5B in 32 representative sections of well-preserved GC tissue and 4 representative sections of well-preserved hepatic metastasis. Sections were incubated for 1 h at room temperature with a rabbit polyclonal antibody raised against GPR155 (sc-137511, Santa Cruz Biotechnology Inc., Dallas, TX, USA) diluted 1:100, a mouse monoclonal antibody raised against TWIST (ab175430, Abcam, Cambridge, UK) diluted 1:500 or a rabbit polyclonal antibody raised against WNT5B (ab115563, Abcam, Cambridge, UK), diluted 1:200 in Antibody Diluent (Dako, Carpenteria, CA, USA). Antigen-antibody complexes were visualized using liquid 3, 3’-diaminobenzidine (Nichirei, Tokyo, Japan) after a 2 min incubation. To avoid bias, specimens were randomized, coded, and then analyzed by two independent observers who were uninformed of the identities of the samples. Each observer evaluated all specimens at least twice within a given time interval to minimize intraobserver variation. Staining intensity of GPR155 was categorized to three groups; increased (GC > non-cancerous component), equivalent and decreased (GC > non-cancerous component). Staining intensities of TWIST and WNT5B were scored by proportions of stained cells in GC component as follows; no-staining (0%), minimal (less than 30%), focal (30–60%) and diffuse (more than 60%)49.

Statistical Analysis

Differences between the data of two groups were evaluated using the Mann–Whitney test. To divide patients into two groups, a cut-off value was determined according to the receiver operating characteristic curve analysis drawn by data from all 200 GC patients in stage I–IV GC including those with hematogenous metastasis. The χ2 test was used to analyze the significance of the association between GPR155 mRNA expression levels and patients’ clinicopathological characteristics, the significance of the association between IHC intensity and the incidences of hematogenous metastasis, and correlations in IHC intensities. Survival rates were calculated using the Kaplan–Meier method, and the differences in survival curves were evaluated using the log-rank test. Predictive factors were evaluated using multivariate analysis. All statistical analyses were performed using JMP 10 software (SAS Institute Inc., Cary, NC). A p value < 0.05 was considered statistically significant.

Additional Information

How to cite this article: Shimizu, D. et al. GPR155 Serves as a Predictive Biomarker for Hematogenous Metastasis in Patients with Gastric Cancer. Sci. Rep. 7, 42089; doi: 10.1038/srep42089 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.