Decreased Expression of Cell Adhesion Molecule 4 in Gastric Adenocarcinoma and Its Prognostic Implications

Cell adhesion molecule 4 (CADM4) is a novel tumor suppressor candidate. The prognostic implications of CADM4 in gastric cancer have not been conclusively elucidated. Therefore, we evaluated the clinicopathological significance and prognostic value of CADM4 expression in a large series of patients with gastric adenocarcinoma. Immunohistochemical staining for CADM4 was performed on 534 gastric adenocarcinomas. We evaluated the associations between CADM4 expression and the clinicopathological and molecular characteristics of the adenocarcinomas. The prognostic effect of CADM4 expression was evaluated by survival analyses. Low CADM4 expression was significantly associated with young age (p = 0.046), aggressive histological type (p < 0.001), high pT category (p < 0.001), nodal metastasis (p < 0.001), high stage (p = 0.002), lymphovascular invasion (p = 0.001), and perineural invasion (p = 0.001). Low CADM4 expression was more frequently observed in tumors without human epidermal growth factor receptor 2 (HER2) amplification (p = 0.002). Low CADM4 expression was associated with worse overall survival (p = 0.007) and recurrence-free survival (p = 0.005) in the survival analyses. Low CADM4 expression was associated with aggressive clinicopathological features and poor clinical outcomes. CADM4 can act as a tumor suppressor in gastric adenocarcinoma and can be considered a prognostic biomarker.


Introduction
Gastric cancer is a leading cause of cancer-related deaths in the world. In 2020, gastric cancer was the fifth most common type of newly diagnosed cancer cases and the fourth most common cause of cancer-related death [1]. Although the incidence and mortality rates of gastric cancer are gradually decreasing worldwide, there are many new cases in East Asia including Korea and Japan [2]. Gastric cancer is a heterogeneous group that can be divided based on distribution and histological patterns. In addition, gastric cancer can be classified according to molecular drivers, and the molecular classification published by The Cancer Genome Atlas (TCGA) research network is representative [3]. Each molecular subtype is associated with clinical characteristics such as the prognosis following surgery, benefit from chemotherapy, and the effect of anti-programmed cell death-1 (PD-1) antibodies [4]. Clinical research on a novel therapy targeting various biomarkers is underway, such as anti-PD-1 antibodies (for tumors with microsatellite instability high phenotype or expressing programmed cell death-ligand 1) and trastuzumab (for tumors with HER2 positivity) used in routine clinical practice [4]. 2 of 11 Cell adhesion molecule 4 (CADM4/TSLL2/SynCAM4/Necl-4/IGSF4C) is one of the immunoglobulin (IG) superfamily molecules and has structural similarity to tumor suppressor in lung cancer 1 (TSLC1) [5]. TSLC1 is a novel tumor suppressor gene for non-small cell lung cancer [6]. CADM4 shows significant homology with TSLC1, including a short cytoplasmic domain that is considered to play a critical role in tumor suppressor activity [7]. Therefore, there have been several studies on CADM4 expression in cancer cells and its role in the development or progression of cancer. In vitro studies on several types of cancer, including prostate cancer, glioma, colon cancer, renal cell carcinoma, and lung cancer, showed that CADM4 was overall down-regulated [7][8][9][10][11]. In addition, overexpression of CADM4 was efficient in suppressing the tumorigenicity of cancer cells [8,9,11]. Similar to these results, a decrease in or loss of expression of CADM4 in tumor cells was associated with aggressive clinicopathological phenotypes and poor prognosis, suggesting the potential of CADM4 as a tumor suppressor [10,[12][13][14][15][16][17][18].
For gastric cancer, Song et al. reported that a lower expression level of CADM4 mRNA was associated with aggressive features [19]. However, Sayar et al. could not reveal a significant relationship between CADM4 expression and clinicopathological characteristics in gastric cancer [16]. Therefore, in this study, we investigated the associations between CADM4 expression and clinicopathological features of gastric cancer by performing immunohistochemical (IHC) staining on a large series of resected samples. In addition, the associations between CADM4 expression and molecular characteristics including HER2 status were evaluated.

Patient Selection and Clinicopathological Data Collection
A total of 544 patients with gastric adenocarcinoma were enrolled retrospectively in this study. All patients underwent surgical resection (gastrectomy or submucosal dissection) between February 2005 and August 2010 at Hanyang University Hospital in the Republic of Korea. We excluded 10 cases due to incomplete follow-up data or no available tumor tissue and conducted the study with the remaining 534 cases. All tissue slides used at the time of diagnosis and pathologic reports were reviewed by two pathologists (Bang, S. and Shin, S.-J.). We assessed pathologic features of tumor location, gross type, histological type, Lauren classification, lymphatic or vascular invasion, perineural invasion, pT stage, and pN stage. The histological type was assessed using the World Health Organization classification for gastric adenocarcinoma, and the pathologic stage was assessed using the 8th edition of the American Joint Committee on Cancer (AJCC). Medical records were reviewed to obtain clinical information, including patient age, sex, and follow-up data.

Tissue Microarray Construction and Immunohistochemical Staining
Tissue microarray (TMA) construction was performed using a Tissue Microarray Set (Labro, Seoul, Korea) and formalin-fixed paraffin-embedded tissues. All gastric cancer tissues used in the study were resection specimens, and a representative portion of the tumor was selected by light microscopy. Then, we obtained a 3.0 mm tissue core from the donor block and transferred it to the recipient block (Labro, Seoul, Korea). Each TMA consisted of 6 × 5 samples.
We cut 4-µm-thick sections from each TMA block. IHC staining for CADM4 was performed with a Benchmark XT automated staining system (Ventana Medical Systems, Tucson, AZ, USA). Primary antibody against CADM4 (1:100, SAB4500746, Sigma-Aldrich, St. Louis, MO, USA) was used according to the manufacturer's instructions.

Interpretation of Immunohistochemical Staining
IHC staining assessment was performed by two pathologists (Bang, S. and Paik, S.) without access to the clinical data. For semi-quantitative assessment, cytoplasmic staining of the tumor cells was assessed using the immunoreactive score (IRS) as previously described [12]. The intensity of staining was categorized as 0 to 3 (0: negative, 1: weak, 2: moderate, and 3: strong), and the proportion of staining was graded as 0 (0%), 1 (1-25%), 2 (26-50%), 3 (51-75%), and 4 (>75%). IRS was calculated as the product of the intensity and proportion of staining and ranged from 0 to 12. A receiver operating characteristics (ROC) curve analysis was used to determine optimal cutoff value. We divided the cases into low expression (IRS ≤ 4) and high expression (IRS > 4) groups, and these were used for all statistical analyses.

Assessment of Molecular Characteristics
The TCGA Research Network proposed dividing the molecular classification of gastric cancer into four molecular subtypes: Epstein-Barr virus (EBV), microsatellite instability (MSI), genomic stability (GS), and chromosomal instability (CIN) [20]. To identify the MSI subtype, we performed IHC staining on representative tumor sections. IHC staining and silver DNA in situ hybridization (SISH) were used to determine HER2 status. Automatic staining was performed with anti-HER2 antibody on the whole tumor sections according to the manufacturer's instructions (4B5, Roche, Basel, Switzerland) using the INFORM HER2 Dual ISH DNA probe cocktail (Roche, Basel, Switzerland). We considered HER2-amplified cases as those with strong membranous reactivity (3+) in IHC staining or with a HER2-chromosome 17 ratio ≥ 2.0 in SISH [21].

Statistical Analyses
All data were analyzed using SPSS software version 25.0 (IBM, Armonk, NY, USA). Pearson's chi-square test was performed to evaluate the correlations between CADM4 expression and clinicopathological characteristics and between CADM4 expression and molecular characteristics. The Kaplan-Meier method with the log-rank test was used to identify the influence of CADM4 expression on overall survival (OS) and recurrence-free survival (RFS). The Cox proportional hazard model was used to determine the significant prognostic factors. A two-tailed p-value < 0.05 was considered statistically significant.

Correlations between CADM4 Expression and Clinicopathological Characteristics
The cytoplasmic expression of CADM4 was detected, and representative photomicrographs are presented in Figure 1. CADM4 expression was reduced or absent in 303 of 534 gastric adenocarcinomas. Low CADM4 expression was significantly associated with young age (p = 0.046), undifferentiated and other histologic types (p < 0.001), diffuse and mixed types of Lauren subtype (p < 0.001), high pT category (p < 0.001), nodal metastasis (p < 0.001), high stage (p = 0.002), lymphovascular invasion (p = 0.001), and perineural invasion (p = 0.001). The patient's sex and tumor location were not significantly associated with CADM4 expression. The correlations between CADM4 expression and clinicopathological characteristics are summarized in Table 2.

Correlations between CADM4 Expression and Molecular Characteristics
There was no significant correlation between CADM4 expression and the 35 (6.6%) cases of EBV subtype or the 43 cases (8.1%) of MSI subtype. We identified 26 HER2 amplification cases (4.9%) that were not included in the EBV and MSI subtypes. Low CADM4 expression was more frequently observed in the tumors without HER2 amplification and was statistically significant (p = 0.002, Table 3).

CADM4 Expression and Prognostic Implication
We performed Cox regression analyses to investigate the significant clinicopathological factors and to reveal the prognostic significance of CADM4 expression (Table 4). In univariate analyses, low CADM4 expression (p = 0.007), old age (p < 0.001), diffuse and mixed types of Lauren classification (p = 0.019), high stage (p < 0.001), lymphovascular invasion (p < 0.001), perineural invasion (p < 0.001), and HER2 amplification (p = 0.049) were associated with short OS. Low CADM4 expression (p = 0.006), undifferentiated and other histological types (p < 0.001), diffuse and mixed type of Lauren classification (p < 0.001), high stage (p < 0.001), lymphovascular invasion (p < 0.001), and perineural invasion (p < 0.001) were associated with short RFS. There was no prognostic difference according to MSI subtype and EBV subtype (data not shown). In multivariate analyses, low CADM4 expression (p = 0.002), old age (p < 0.001), high stage (p < 0.001), perineural invasion (p = 0.030), and HER2 amplification (p = 0.010) were associated with short OS, and high stage (p < 0.001) was associated with short RFS. The Kaplan-Meier method was used to reveal the prognostic significance of CADM4 expression. Patients with low CADM4 expression showed significantly shorter OS and RFS (log-rank test, p = 0.007 and p = 0.005, respectively; Figure 2) compared to those with high CADM4 expression. In the HER2 non-amplified cases, patients with low CADM4 expression had shorter OS and RFS (log-rank test, p = 0.003 and p = 0.004, respectively). However, CADM4 expression did not significantly affect OS and RFS in patients with HER2 amplification (Figure 3).
Diagnostics 2022, 12, 941 8 showed significantly shorter OS and RFS (log-rank test, p = 0.007 and p = 0.005, res tively; Figure 2) compared to those with high CADM4 expression. In the HER2 non plified cases, patients with low CADM4 expression had shorter OS and RFS (log-rank p = 0.003 and p = 0.004, respectively). However, CADM4 expression did not significa affect OS and RFS in patients with HER2 amplification (Figure 3).    showed significantly shorter OS and RFS (log-rank test, p = 0.007 and p = 0.005, respectively; Figure 2) compared to those with high CADM4 expression. In the HER2 non-amplified cases, patients with low CADM4 expression had shorter OS and RFS (log-rank test, p = 0.003 and p = 0.004, respectively). However, CADM4 expression did not significantly affect OS and RFS in patients with HER2 amplification (Figure 3).  In the HER2-positive group, there was no significant difference in overall survival or recurrencefree survival according to CADM4 expression. In the HER2-positive group, there was no significant difference in overall survival or recurrence-free survival according to CADM4 expression.

Discussion
In this study, we performed IHC staining with CADM4 antibody on 534 gastric adenocarcinomas. We evaluated the associations of CADM4 expression with clinicopathological features, prognostic significance, and molecular characteristics. Low expression of CADM4 was significantly associated with aggressive histological type, high pT category, nodal metastasis, high stage, lymphovascular invasion, and perineural invasion. The results suggest that CADM4 down-regulation is associated with aggressive clinicopathological features. Low CADM4 expression is an independent prognostic factor for OS by univariate and multivariate Cox regression analyses. Low CADM4 expression was associated with worse RFS in univariate analysis. However, it lost statistical significance in multivariate analysis. Using Kaplan-Meier survival analyses, we determined that low CADM4 expression was associated with poor prognosis. Low CADM4 expression was not associated with EBV or MSI subtype.
CADM4 is a potential prognostic biomarker and is thought to play a role as a tumor suppressor. Jang et al. showed that the loss of CADM4 expression is relatively frequent in colorectal adenocarcinomas and proposed that CADM4 plays an important role in cancer progression and patient survival [12]. Jang et al. demonstrated that a loss of or decrease in CADM4 expression likely plays an important role in breast cancer invasiveness and is associated with worse biological parameters [13]. Kawanishi et al. suggested that negative Necl-4 expression is associated with carcinogenesis and the aggressiveness of pancreatic ductal adenocarcinomas [14]. Kim et al. reported that the loss of CADM4 expression is associated with poor prognosis in patients with small intestinal adenocarcinoma [15]. Luo et al. showed that down-regulation of CADM4 promotes tumor growth and metastasis in non-small cell lung cancer [11], and Du et al. predicted CADM4 as a therapeutic target [22].
The mechanism regulating CADM4 expression in tumors is unknown. According to the TCGA's Pan-Cancer Atlas studies, genetic alterations of CADM4 have been reported in only a small number of cases (2.38%). Several studies have revealed that TSLC1, which has structural similarity to CADM4, is down-regulated by promoter methylation in some tumors (30-60%) [23]. CADM4 expression is also irreversibly down-regulated in cancer cells in several studies, and there is a possibility of a decrease in CADM4 expression due to DNA methylation [24,25], but further study is needed.
Two previous studies have reported the expression of CADM4 in gastric cancer. Song et al. demonstrated that the expression of CADM4 is down-regulated in gastric cancer tissues and cell lines, and that down-regulation of CADM4 in patients without lymph node metastasis is significantly associated with the degree of cell differentiation, depth of tumor invasion, lymph node metastasis, and TNM stage [19]. Sayar et al. performed IHC staining on 51 gastric cancer tissues and reported that expression of CADM4 at the protein level does not show a significant association with tumor differentiation, lymphovascular invasion, perineural invasion, depth of tumor invasion, vascular invasion, or metastasis [16]. In our study, we evaluate CADM4 expression in a large series of patients with gastric adenocarcinoma. Our study reveals that low CADM4 expression is significantly associated with aggressive clinicopathological features and poor prognosis. The relationships between CADM4 expression and molecular characteristics also are evaluated, and low CADM4 expression is more frequently observed in HER2-negative tumors.
In conclusion, we demonstrate the associations between low CADM4 expression and aggressive clinicopathological features and clinical outcomes for gastric adenocarcinoma. CADM4 can act as a tumor suppressor in gastric adenocarcinoma. Future studies are needed to define the molecular mechanisms regulating the expression of CADM4 and its usefulness as a therapeutic target.