Cancers of the gastrointestinal (GI) tract accounted for an estimated 4.8 million new cancer cases and 3.4 million cancer-related deaths worldwide in 2018; these numbers correspond to over one-quarter (26%) of global cancer cases and over one-third (35%) of all cancer-related deaths (1). Thus, cancers of the GI tract remain important contributors to the global cancer burden. Improving the efficacy of tumour treatment is the key to improving the survival and prognosis of patients with GI tumours.

In tumour treatment, surgery and chemotherapy are the common treatment approaches (2). For patients with early cancer, surgical resection remains the primary treatment option; for patients with advanced metastasis or patients who cannot undergo surgical resection, chemotherapy is the most indispensable treatment (3). However, chemotherapy resistance and tumour progression and recurrence in cancer patients have become the most common causes of treatment failure (4). In the past decade, remarkable advances made by studies of tumour biology, tumour immunotherapy and targeted therapy have revolutionised the treatment of a fair number of cancers (5). Small molecule targeted drugs have shown certain advantages in tumour treatment.

Recently, increasing attention has been given to a potential target, Claudin18.2 (CLDN18.2). Its restricted expression in normal tissues and specific expression in tumour tissues makes it an ideal target for researchers to develop targeted therapy and immunotherapies for solid tumours (6). CLDN18.2 was initially identified as a gastric tumour-specific target, and many clinical trials of monoclonal antibodies targeting CLDN18.2 have been conducted for the treatment of patients with gastric and gastroesophageal cancer. However, several other novel targeting methods, including chimeric antigen receptor T (CAR-T) therapy, bispecific antibodies and antibody–drug conjugates (ADCs), have rapidly emerged in recent years, and a variety of corresponding clinical trials that have broadened their treatment adaptation to populations, especially those with GI tract cancer, have been conducted. This review will discuss the expression pattern of CLDN18.2, its roles in tumours and its relevant molecular mechanisms, as well as the current progress in the development of drugs targeting CLDN18.2 in clinical research.

Claudins, a family of at least 27 transmembrane proteins, are important components and functional structures of tight cell junctions. Claudin-18 is a major component of tight junctions located on the cell membrane surface; it plays an important role in the maintenance of cell polarity and barrier function and promotes acid resistance (7–10). For example, in the stomach, claudin-18 normally forms a paracellular barrier against H(+), promotes acid resistance, and causes paracellular H(+) leakage, persistent upregulation of proinflammatory cytokines and atrophic gastritis in mice (11). The human CLDN18 gene locus on chromosome 3q22 has a molecular weight of approximately 35 kb and contains 6 exons and 5 introns. The first exon of CLDN18 can be alternatively spliced, forming two different splice mutants (CLDN18.1 and CLDN18.2) that have highly homologous amino acid sequences (12). The transmembrane protein CLDN18 consists of two extracellular loops (ECLs), four transmembrane domains and a cytoplasmic domain. Both the C-terminus and the N-terminus of CLDN 18 are located in the cytoplasm ( Figure 1 ) (6). Two CLDN18 protein isoforms are expressed in a tissue-specific manner—CLDN18.1 and CLDN18.2 are specifically expressed in normal stomach and lung tissues, respectively (13). CLDN18 is also expressed in cancer tissues and has altered functions that are linked to tumour formation, proliferation, invasion and migration (6, 14–18).

CLDN18.1 is highly expressed in lung alveolar epithelium and is involved in regulating solute and ion permeability, as well as cytoskeletal rearrangement (9). It has been found that CLDN18.1 knockout increased solute permeability and alveolar fluid clearance, and enhanced epithelial sodium channel and Na-K-ATPase activity (9). However, CLDN18.1 expression was significantly decreased in lung cancer and may play a role in inhibiting tumour progression (19, 20). Luo et al. found that when restored in lung adenocarcinoma (LUAD) cells that have lost expression, CLDN18.1 markedly attenuates tumour growth in vivo as well as cell proliferation, migration, invasion and colony formation in vitro (20). Zhou et al. also elucidated that knockout of Cldn18.1 results in increased lung size, progenitor cell proliferation, and tumourgenesis in mice (17).

CLDN18.2 is also involved in regulating ion permeability including H(+) and other cations (11). CLDN18.2 deficiency leads to neutrophils recruiment and inflammation development (11). As the focus of our review, the expression pattern and role of CLDN18.2 in tumours will be elaborated in detail later in the article.

ARHGAP26 is a GTPase-activating protein (GAP) that promotes conversion of RHO GTPases to the GDP state and has been indicated in enhancing cellular motility (21). CLDN18-ARHGAP26 fusion gene is one of the most frequent somatic genomic rearrangements in gastric cancer, especially in the genomically stable subtype (22, 23). Cell survival and migration studies performed by Ushiku et al. showed that CLDN18-ARHGAP fusions complemented cancer cell survival and induced cell migration under RHOA knockdown condition. RHOA mutations exist specifically in diffuse type gastric cancers (DGC) and are considered one of the DGC driver genes. Thus, CLDN18-ARHGAP fusion is possible a factor for RHOA mutation to maintain cancer cell survival and promoted cell migration (24).

For a long time, CLDN18.2 has been widely studied in patients with various solid tumours, especially in patients with cancers of the GI tract. Data for the expression pattern and clinical significance of CLDN18.2 in cancers of the GI tract are summarised in Table 1 .

Up to present, for CLDN18.2 detection modalities, the vast majority of clinical studies use tissue samples to detect its protein expression by immunohistochemistry, including all studies in Table 1 . Besides, Sanada et al. also tested CLDN18.2 mRNA expression by reverse transcription (RT)-polymerase chain reaction (PCR) (41). In addition, one study innovatively designed a CLDN18.2 molecular beacon (MB) with high resolution and selectivity to distinguish CLDN18.2 RNA expression in circulating tumour cells (CTCs). They compared immunohistochemistry to MB data in 10 GC patients and verified that the concordance in the expression of CLDN18.2 between CTCs and tissue biopsy is 100% (negative: 3 vs 3; positive: 7 vs 7) (42). Thus, CTCs detection in peripheral blood, a non-invasive and method, may be a promising approach for evaluating CLDN18.2 expression and subsequently contribute to timely treatment for patients diagnosed with CLDN18.2 highly expressed GC. Further investigations are needed for the other courses of tumours like prognosis and the tests for more patients.

The CLDN18.2 protein is a highly selective gastric lineage marker expressed in short-lived differentiated cells but not the stem cell zone of stomach mucosa (43). In addition to its orthotopic expression, frequent ectopic activation of CLDN18.2 was detected in pancreatic, oesophageal, biliary tract, ovarian, and lung tumours, whereas downregulation was found in gastric cancer (GC) and sarcoma (gastric stromal tumour) tissues compared to their corresponding normal tissues (38, 43–45). It is worth noting that CLDN18.2 expression is decreased but still stable during gastric malignant transformation. Moreover, CLDN18.2 activation in metastatic lesions (lymph node and liver metastases) is comparable to that in primary lesions (26, 30, 31). These findings are of great significance, as they indicate that not only those with locally advanced disease but also those with metastatic cancer may be candidate patient populations for clinical trial targeting CLDN18.2.

CLDN18 expression is associated with histological subtype in GC. In GCs grouped by the Lauren subtype, CLDN18.2 expression was significantly higher in diffuse-type GCs (26, 30). In terms of the mucin phenotype, samples with the intestinal phenotype predominantly showed downregulation of CLDN18.2, suggesting that the loss of CLDN18 is involved in the pathogenesis of tumours of the intestinal phenotype and indicating that its expression can be used as a marker for the gastric phenotype (28, 41, 46).

In the relationship between CLDN18.2 expression and clinicopathological factors, Jun et al. showed a significant inverse correlation between claudin-18 expression and perineural invasion (27). In addition, a study of Japanese patients with gastric adenocarcinoma revealed that moderate-to-strong CLDN18.2 expression positively correlated with a higher grade of dedifferentiation (59.0% vs. 37.7% for G3 and G1/2 tumours, respectively; P = 0.005) (26). However, Soini et al. identified an inconsistent result with the previous study that positive CLDN18 expression is associated with a higher differentiation stage (35). In the pancreas, a correlation analysis revealed that lymph node-positive tumours had significantly higher CLDN18.2 expression indicating by the positive cell fractions and membrane staining intensities (33).

Claudin-18 has also been reported to be correlated with survival in GC patients. Jun et al. suggested the loss of claudin-18 as an independent indicator of a poor prognosis in patients with GC, as patients with claudin-18 expression had longer overall survival (OS) than those without claudin-18 expression (27). Sanada et al. also indicated that downregulation of claudin-18 was correlated with poor survival in patients with advanced GCs (41). However, in pancreatic cancer (PC), although there was an association between CLDN18.2 expression and lymph node metastasis (LNM) and dedifferentiation stage, as mentioned above, no correlation between CLDN18.2 expression and patient survival was found (33–36). In endocervical adenocarcinoma, gastric immunophenotypes (CLDN18.2) were identified as an independent predictor of poorer progression-free survival (PFS) (47).

Moreover, CLDN18-ARHGAP26/6 fusions have been identified in GCs (48). Tanaka et al. analysed 254 cases of GC (172 diffuse-type and 82 intestinal-type) using RT-PCR and situ hybridisation (FISH) and identified 26 fusion-positive cases (26/254, 10.24%), 22 of which were diffuse GCs (22/172, 12.79%), indicating that the CLDN18-ARHGAP26/6 fusion was specific to diffuse-type GCs (49). However, Hashimoto et al. found that intestinal-type adenocarcinoma with anastomosing glands, a genetically distinct group of intestinal-type adenocarcinoma, also frequently had CLDN18-ARHGAP fusion (50). Shu et al. found that the frequency of CLDN18-ARHGAP fusion is 18.25% in signet-ring cell carcinoma, and the CLDN18-ARHGAP26/6 fusion is associated with signet-ring cell content, age at diagnosis, female/male ratio, and TNM stage. The survival rate of patients with CLDN18-ARHGAP26/6 fusion is worse, and patients with this fusion obtain no benefit from oxaliplatin/fluoropyrimidine-based chemotherapy (51). Further studies have identified significant drug resistance acquired in CLDN18-ARHGAP26-introduced cell lines (51). Matsusaka et al. found that the combination of CDH17 and CLDN18 demonstrates homogeneous and robust expression in more than 90% of GC cases, and their coupling is helpful to detect and localise GC metastasis in vivo, which provides a possibility to overcome the challenge of intratumoural heterogeneity (52). Nakayama et al. found that the CLDN18‐ARHGAP fusion rate was 15.1% in 146 patients with GC under the age of 40. Among them, 18 cases (12.3%) were CLDN18-ARHGAP26 fusions, and 2 cases (1.4%) were CLDN18-ARHGAP6 fusions. Two new fusion types, CLDN18‐ARHGAP10 and CLDN18‐ARHGAP42, were identified. Moreover, patients with fusion were identified to more likely to have large tumours, LNM, and advanced stage and to have exhibited significantly lower OS by Kaplan‐Meier analysis. These observations indicate that the CLDN18‐ARHGAP fusion is enriched in younger age‐onset GCs, and its presence could contribute to their aggressive characteristics (53). Yao et al. demonstrated that the CLDN18-ARHGAP26 fusion mediates epithelial disintegration and contributes to GC by loss of CLDN18 and gain of ARHGAP26 functions (54). The CLDN18-ARHGAP fusion is one of the molecular characteristics of diffuse GC and is also an independent prognostic risk factor for GC. In addition, it is related to multiple clinicopathological characteristics, including age, sex, chemotherapy resistance, LNM and tumour stage. However, further exploration is required to elucidate the mechanism of the CLDN18-ARHGAP fusion gene and identify potential targeted therapeutic strategies.

It is universally acknowledged that a comprehensive and systematic understanding of target molecules’ functions in tumours is rather essential, partly because it can help to correctly apply therapeutic strategies to treat tumour patients. Studies for molecular regulatory mechanism include upstream mechanism study aiming to explore the pathways and molecules that regulate target molecules, and downstream mechanism study aiming to explore the pathways and molecules that are regulated by target molecules. The focus is on the latter in most cases. However, understanding the upstream mechanism is also very important. A comprehensive understanding of upstream regulatory factors can better clarify the source of target molecules and the causes of their changes, providing more insights for antitumour therapies that target molecular markers. In this section, we review the factors that regulate CLDN18 expression, as well as the roles of CLDN18 in cancers of the GI tract.

Researchers have successively developed more specific or effective antibodies to detect CLDN18.2 expression (76, 77) or target it as a therapeutic candidate for tumours (78–82). In the field of therapies targeting Claudin18.2, there are many drugs in the clinical development stage, including monoclonal antibodies, CAR-T cells, bispecific antibodies and ADCs. Among them, several clinical studies of monoclonal antibody therapy and CAR-T therapy have disclosed clinical trial data to date. The latter two emerging drug types, ADCs and bispecific antibodies, have been approved for clinical testing in recent years, and studies are still ongoing. In contrast to previous clinical trials that only assessed the safety and efficacy of CLDN18.2-targeted therapies in populations with advanced gastric and oesophagogastric junction cancers, recently, approved clinical trials have also expanded to include patients with other advanced solid tumours, mostly tumours of the digestive system, including PC, colorectal cancer, intrahepatic cholangiocarcinoma, and extrahepatic bile duct cancer. The CLDN18.2 clinical trials with disclosed results are shown in Table 2 ; ongoing clinical trials of CLDN18.2 are shown in Table 3 .

CLDN18.2 is restricted to normal gastric tissue and is downregulated in GC tissues but is highly activated in other tumour tissues, including pancreatic, oesophageal, biliary tract, ovarian, and lung tumours. In addition, CLDN18.2 still has high expression during gastric tumour formation and retained expression in metastatic lesions. Therefore, CLDN18.2 is suggested as an ideal target for tumour treatment, especially in cancers of the GI tract. Moreover, CLDN18.2 can act as an oncogene or a tumour suppressor gene in different tumour types, and its expression is modulated by the methylation status of CpG islands, PKC, ERK/MAPK, HER2/HER3 signalling pathways and microRNA.

From the disclosed data, monoclonal antibody and CAR-T therapies have shown good safety in CLDN18.2-positive gastric and pancreatic cancer populations and have achieved encouraging preliminary results in terms of clinical efficacy. In addition, we are very much looking forward to encouraging the results of ongoing clinical trials of long-researched or emerging drugs targeting CLDN18.2 in various tumours, which may provide several new potential methods and benefits for millions of people with CLDN18.2-positive tumours, including GC, PC, hepatocellular carcinoma, and bile duct cancer. Further attention should be given to the definition of populations with high CLDN18.2 expression that may benefit from treatment, the prediction of marker efficacy, the prevention and control measures for adverse reactions, the selection of treatment regimens for multitarget coexpression populations, and the combination of immunotherapy, which are all directions for future exploration. Subsequent clinical trials of CAR-T cells, ADCs, and bsAb drugs will undoubtedly provide more options for patients with high CLDN18.2 expression and poor efficacy with targeted drugs or advanced multiline treatment.

XC and LY conceptualised the manuscript. JC, ZX, CH, SZ, and MZ collected the literature, JC and ZX collected the literature, wrote the manuscript and made the figures. XC and LY edited and made significant revisions to the manuscript. All authors contributed to the article and approved the submitted version.