Claudin-18.2 as a therapeutic target in cancers: cumulative findings from basic research and clinical trials

ABSTRACT

Claudins are major components of tight junctions that maintain cell polarity and intercellular adhesion. The dynamics of claudins in cancer cells have attracted attention as a therapeutic target. During carcinogenesis, claudin expression is generally downregulated; however, overexpression of claudin-18.2 has been observed in several types of cancers. Upregulated and mislocalized claudin-18.2 expression in cancer cells has been suggested as a therapeutic target. Research on claudin-18.2 has revealed its involvement in carcinogenesis. Clinical trials using zolbetuximab, a monoclonal antibody targeting claudin-18.2, for patients with advanced cancer yielded positive results with few high-grade adverse events; thus, it is expected to be a novel and effective therapeutic. Here, we review current insights into the role that claudin-18.2 plays in basic cancer research and clinical applications. A better understanding of these roles will facilitate the development of new treatment strategies for cancer patients with poor prognoses.

KEYWORDS:

• Tight junction
• claudin-18.2
• cancer
• zolbetuximab

Introduction

Tight junction molecules have intercellular adhesion functions in all cell types. Recently, the dynamics of these proteins in cancer cells have attracted attention. Epithelial cells are held together by several major classes of intracellular junctions, and adherens and tight junctions play essential roles in the development and maintenance of epithelial cells. Tight junctions are typically located on the apical side of epithelial cells, and their components interact with those of neighboring cells. Tight junction proteins are often dysfunctional or altered in various types of cancer cells, and their controlled paracellular permeation and polarity are lost in cancer cells¹. Altered tight junction proteins also modulate cytoskeletal elements and signaling molecules bound to these proteins, resulting in the loss of regulated cell migration and proliferation.²

The two major components of tight junctions are occludin and claudin proteins, which bind to the PDZ domains of zonula occludens (ZO) proteins to anchor the actin cytoskeleton.^3,4 Claudins are 20–27-kDa transmembrane proteins that form extremely tight associations with their counterparts on adjacent cells.⁴ These proteins have four transmembrane domains and two extracellular loops to maintain cell-cell integrity and regulate paracellular ion transport.³ Currently, there are 27 known members of the claudin family and cancer cells have a specific claudin expression pattern according to the tumor cell origin.⁵ In 2008, Sahin et al. demonstrated that the claudin-18 splice variant 2, claudin-18.2, is highly expressed in several tumors.⁶ Research on claudin-18.2 has gradually elucidated its expression patterns and functions in cancer cells. Moreover, clinical trials using anti-claudin-18.2 antibodies as a cancer-specific molecular targeted therapy for advanced gastric adenocarcinoma are ongoing.^7–9 This accumulating body of evidence indicates that claudin-18.2 is one of the most clinically relevant tight junction proteins.

For successful antibody therapy targeting claudin-18.2, it is important to understand the dynamics, localization, and function of claudin-18.2 in cancer cells. In addition, the efficacy, eligible patients, and adverse events of monoclonal antibodies to claudin-18.2 in cancer patients need to be analyzed together with in vitro and in vivo results. While most reviews describe either the results of basic research or the results of clinical trials, we have summarized both findings to provide a better understanding of the dynamics of claudin-18.2 and cancer therapy targeting claudin-18.2. Here, we review current insights into the role that claudin-18.2 plays in basic cancer research and clinical applications. A better understanding of these roles will facilitate the development of new treatment strategies for cancer.

Claudins in cancer, and therapeutic approaches targeting claudins

Claudins have three main functions: barrier, fence, and intracellular signaling.^1,10 The barrier function is the ability to selectively regulate paracellular permeation of water, ions, macromolecules, and immune cells, whereas the fence function separates the apical and basolateral domains and regulates the movement of substances within the plasma membrane.^10,11 Tight junctions, including claudins, also act as signaling hubs by binding to multiple signaling molecules. Claudins in cancer cells are known to activate signaling pathways associated with tumor progression and metastasis. For example, claudin-1 activates the c-Abl-Ras-Raf-1-ERK1/2 signaling pathway in hepatocellular carcinoma¹² and the Notch/phosphoinositide-3-kinase (PI3K)/Akt signaling pathway in colitis-associated cancer.¹³ Claudin-2 in colon cancer is substantially upregulated via the epidermal growth factor receptor-ERK1/2 kinase axis and overexpressed claudin-2 increases cell proliferation and tumor growth in vivo.¹⁴ Claudin-3 has a suppressive role in epithelial-mesenchymal transition via activation of the Wnt/beta-catenin signaling pathway in lung squamous cell carcinoma¹⁵ and hepatocellular carcinoma.¹⁶ Thus, the impact of claudins on carcinogenesis has been investigated in various types of cancer cells.

In normal epithelial cells, claudins are mostly present in the apical tight junction fraction, which forms intercellular strands.¹⁷ Claudin expression is generally downregulated and mislocalized during carcinogenesis and epithelial-mesenchymal transition of cancer cells, and loss of claudins causes dysfunction of their epithelial polarity and barrier function.^2,18 However, overexpression of claudins has also been observed in several types of cancer. The expression of each cancer cell tight junction molecule is abnormally regulated.¹⁹ Claudin-18.2 expression of cancer cells was transcriptionally upregulated with the binding of cyclic AMP–responsive element binding protein to the methylated CLDN18.2 promoter region.⁶ The upregulation of claudins in cancer cells in comparison to normal epithelial cells could be a diagnostic tool and a therapeutic target.²⁰ A recent study has shown that a human IgG1 monoclonal antibody against the second extracellular loop of claudin-3 (h4G3) recognizes claudin-3-expressing tumors rather than normal organs in mouse xenograft models.²¹ Claudin-3 is overexpressed in various carcinomas, including breast, colorectal, gastric, pancreatic, prostate, and ovarian cancer,²² while it is also expressed in normal tissues such as the colon, rectum, thyroid, salivary glands, pancreas, prostate, liver, and kidney.²³ The antibody h4G3 had a lower affinity to normal tissues than to tumor sites in mice bearing xenograft tumors despite its ability to recognize mouse claudin-3, and it had no antitumor efficacy in vivo.²¹ Thus, although claudin-3, which is present in various organs of the body, has potential value for cancer diagnosis, its specific antibody can cause severe adverse events involving cell death. As claudins are generally expressed in normal epithelial and endothelial cells at various locations,²⁴ the application of anti-claudin antibodies to humans includes the potential for off-target effects and fatal adverse events.

Antibodies against claudins specific to cancer cells are being investigated as novel therapeutic tools for cancer patients. An anti-claudin-6 monoclonal antibody (ASP1650, also called IMAB027) was generated and used in a phase I clinical trial (OVAR; NCT02054351) for patients with advanced and recurrent ovarian cancer²⁵ and in a phase II clinical trial (NCT03760081) for patients with advanced germ cell tumors of testicular cancer.²⁶ Claudin-6 expression is activated in various cancers, including gastric adenocarcinoma and embryonic carcinoma, while it is limited to embryonic development in normal tissues.²⁷ It was demonstrated that ASP1650 binds specifically to claudin-6 without cross-reactivity with claudin-3, −4, and −9 and induces antibody-dependent cell-mediated cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC) in claudin-6-expressing cells in vitro.²⁸ However, clinical efficacy and safety have not been reported in published articles. Monoclonal antibodies against cancer-specific claudins are appropriate for the treatment of cancer; however, cancer-specific claudins have not been found, and the development of antibodies is difficult because each claudin is highly homologous with others. Thus far, only antibodies targeting claudin-6 and −18.2 have been used in clinical trials for cancer treatment.

Roles of claudin-18.2 in individual cancer types

In this section, we describe the function, localization, and expression level of claudin-18.2, and its correlation with intracellular signaling pathways, clinicopathological factors, and prognosis in several cancers. Claudin-18.2 has been actively studied owing to its potential value in the diagnosis of malignant tumors and the application of zolbetuximab, a monoclonal antibody against claudin-18.2.

Gastric cancer

Claudin-18 has two alternatively spliced variants: claudin-18.1 in the lung and claudin-18.2 in the stomach (Figures 1 and Figures 2).²⁹ The expression of claudin-18.2 in normal tissues is limited to differentiated epithelial cells of the gastric mucosa. Claudin-18.2 is retained in primary gastric cancers and their metastases and can also be found in pancreatic, esophageal, ovarian, lung, and colitis-associated colorectal tumors.^6,30 Claudin-18.2 of the epithelium in the stomach regulates cell lineage differentiation and blocks paracellular gastric acid leakage from the gastric lumen into the submucosal space.^31,32 Paracellular H⁺ leakage contributes to the progression of gastritis.³¹ Helicobacter pylori (H. pylori) infection attenuates claudin-18 expression in mice by 6 months post-infection, and its expression decreases over time.³² The loss of claudin-18.2 contributes to paracellular H⁺ leakage and promotes gastric tumor formation with and without H. pylori infection in mice.^32,33 Furthermore, the expression levels of C-X-C motif chemokine 5 (CXCL5), Toll-like receptor 2 (TLR2), and CD44 splice variants increased during gastric tumorigenesis in claudin-18 knockout mice.³³ Another analysis of claudin-18 knockout mice demonstrated that claudin-18 activates cellular signaling pathways, including Wnt/beta-catenin downstream effectors (CD44, EFNB1, EFNB2, and EPHB2) and the Yes-associated protein (YAP)/Hippo signaling pathway in gastric tumorigenesis.³² Notably, these studies revealed that claudin-18.2 knockout mice expressed claudin-18.1, which localized to tight junctions in the gastric mucosa.^32,33 This suggests that a different claudin-18 variant to claudin-18.2 is expressed and plays a role in protecting the stomach tissue.

Figure 1. Amino acid sequences of claudin-18.1 and claudin-18.2.

These sequences were obtained from the protein knowledgebase (UniProtKB: https://www.uniprot.org/). Red characters show the difference in sequence between claudin-18.1 and −18.2; boxes indicate transmembrane domains.

Figure 2. Claudin-18.2 expression in gastric cancer.

Hematoxylin and eosin staining (A, C) and claudin-18.2 immunohistochemical staining (B, D) of intestinal-type gastric cancer (A, B) and diffuse-type gastric cancer (C, D). Bar, 100 μm.

Claudin-18.2 of gastric cancer cells is regulated by the methylation status of CLDN18.2 promoter region⁶ and by the protein kinase C (PKC)/mitogen-activated protein kinase/activator protein 1-dependent pathway in vitro.³⁴ Furthermore, claudin-18.2 expression positively correlated with other adherens junction molecules, such as E-cadherin and Rho GTPase-activating protein (RhoGAP), which contribute to the organization of actin and microtubule cytoskeletons.³⁵ Gene fusion between CLDN18 and ARHGAP (the gene encoding RhoGAP) was observed in 15.1% of young adult patients with gastric cancer and 13.8% of “genomically stable” gastric cancer patients in The Cancer Genome Atlas dataset.³⁶ “Genomically stable” was defined as the absence of microsatellite and chromosomal instabilities. Fusion between CLDN18 and ARHGAP26 in gastric cancer cells resulted in the loss of their epithelial phenotype and enhanced migration capacity,³⁷ increased invasiveness, impaired barrier properties, and reduced cell-cell and cell-extracellular matrix adhesion.³⁸ In addition, CLDN18-ARHGAP fusion predicted distant organ metastasis³⁹ and poor prognosis in patients.^36,37

For the development of the cancer treatment using the anti-claudin-18.2 antibody, it is important to identify which gastric cancer types highly retain claudin-18.2. It has been found that claudin-18.2 is not expressed in all gastric cancers; however, its expression is dependent on cancer cell types and individuals. Most reports investigating the expression rate of claudin-18.2 in gastric cancer are based on immunostaining results using antibodies with an epitope on the C-terminal side common to claudin-18.1 and claudin-18.2 (Figure 1). According to the results, the expression rate of claudin-18 (claudin-18.2) in gastric cancer varies in the range of 53.0–87.0%.^6,40–43 Depending on the selected antibody, the delineation rate varies in the range of 55.6–77.3% using polyclonal antibodies,^6,43 53.0–87.0% using monoclonal antibodies clone 43–14A and 34H14L15 targeting the C-terminus of Claudin-18,^40–42 and 6.0–42.2% using clone EPR19202, a monoclonal antibody targeting amino acids 1–100, specific to claudin-18.2.^40,44 These expression rates include everything from weak to strong expression. The definition of strong expression has not been defined generally, so the criteria vary among publications. In the MONO study, patients with moderate to strong membrane expression in ≥50% of tumor cells (inclusion criteria) were 26.7% and in ≥70% were 14.4% of the 261 enrolled patients.⁴⁵

The expression rate of claudin-18.2 varies depending on which region the patient is from. In a study using the same monoclonal antibody (clone 43–14A), the rate was 53.0% in Germany,⁴⁰ compared to 87.0% in Japan.⁴¹ Nevertheless, only a few reports regarding the expression rate using the same antibody on regional differences exist, and future investigations are needed.

There have been various reports on the relationship between claudin-18.2 expression status and clinicopathological factors or the prognosis of gastric cancer patients. Several studies have shown that claudin-18.2 is retained in gastric cancer of diffuse-type rather than intestinal-type^6,41,42,46 Low expression of claudin-18 in an immunohistochemistry (IHC) staining was closely associated with nerve invasion⁴³and high levels of proliferation and invasion.^32,47 and indicated a poor prognosis of gastric cancer patients.^43,46,48 In metastatic diffuse-type gastric cancer, claudin-18.2 expression was reduced in patients with peritoneal metastasis, however, it increased in patients with bone metastasis.³⁵ Alternatively, increased expression of claudin-18.2 detected with clone 43–14A was not associated with histomorphological subtype, tumor localization, TNM stage, or the overall survival rate in a large Caucasian cohort.⁴⁰ In a recently published meta-analysis, no significant correlations were found between the claudin-18.2 expression level and the overall survival of patients or their clinicopathological features such as TNM stage; Lauren classification; HER2 expression; and gastric cancer grade.⁴⁹

In summary, it seems that approximately half of the patients develop claudin-18.2-positive gastric cancer; however, we should consider the impact of antibody clones such as 43–14A, EPR19202, and 34H14L15, which have a large impact on the claudin-18.2 detection rate in gastric tissue. Furthermore, patient race may affect the positive rate of claudin-18.2. Ongoing clinical trials, involving patients from a wide range of regions globally, may reveal racial differences in claudin-18.2 expression rates (Table 1).^7–9 In addition, they may clarify correlations between prognostic and pathological factors and claudin-18.2. These studies will finally identify appropriate candidates for zolbetuximab. Clinical trials with zolbetuximab will be discussed in a later section.

Table 1. Phase II and III clinical trials with results using zolbetuximab and CAR T cells (accessed on 29 March, 2021)

Name	Date	Phase	Patients	Status of claudin-18.2^a	Line	Clone	Experimental arm	Control arm	Regions	Main result	Identification
MONO^Citation45	2010–2015	II	Claudin-18.2 + G/GEJ/E cancer, metastatic or refractory or recurrent	≥50%	≥2nd	an anti-claudin-18.2 rabbit antiserum	Zolbetuximab	-	Europe^c	ORR: 9% mDOS: 24.6 week	NCT01197885
FAST^Citation50	2012–2019	II	Claudin-18.2+/HER2–G/GEJ/E cancer, unresectable or R2 after resection or recurrent or metastatic	≥40%	1st	CLAUDETECT. ™18.2	EOX + Zolbetuximab	EOX	Europe^d	mPFS^b: 7.5 vs 5.3 mo; OS^b: 13.2 vs 8.4 mo; ORR^b: 39 vs 25%	NCT01630083
ILUSTRO^Citation7	2018-	II	Claudin-18.2 + G/GEJ cancer, unresectable or metastatic (Cohort 3B: PD-L1+)	≥75% (Cohort 3A) and ≥50% (Cohorts 1A, 2 and 3B)	≥3rd (Cohorts 1A and 3A), 1st(Cohorts 2), 3rd (Cohort 3B)	Not listed	Cohort1A:Zolbetuximab, Cohort2: mFOLFOX6 + Zolbetuximab, Cohort 3A/3B: Pembrolizumab + Zolbetuximab	-	Europe, Asia, and North America^e	-	NCT03505320
SPOTLIGHT^Citation8	2018-	III	Claudin-18.2+/HER2 – G/GEJ cancer, unresectable or metastatic	≥75%	1st	Not listed	Zolbetuximab + mFOLFOX6	placebo +mFOLFOX6	North America, Latin America, Europe, Asia, Middle East^f	-	NCT03504397
GLOW^Citation9	2018-	III	Claudin-18.2+/HER2 – G/GEJ cancer, unresectable or metastatic	≥75%	1st	Not listed	Zolbetuximab + CAPOX	Placebo +CAPOX	North America, Latin America, Europe, Asia, Middle East^g	-	NCT03653507
A Study to Assess the Antitumor Activity and Safety of IMAB362 in Combination With Nab-Paclitaxel and Gemcitabine (Nab-P + GEM) as First Line Treatment in Subjects With Claudin 18.2 (CLDN18.2) Positive, Metastatic Pancreatic Adenocarcinoma^Citation51	2019-	II	Claudin-18.2+ Metastatic Pancreatic Adenocarcinoma	≥75%	1st	Not listed	Zolbetuximab + nab-paclitaxel + gemcitabine	nab-paclitaxel +gemcitabine	North America, Europe, and Asia^h	-	NCT03816163
Study to Evaluate the Efficacy, Safety and Pharmacokinetics of CT041 Autologous CAR T-cell Injection	2020-	I/II	Claudin-18.2+ advanced G/GEJ and pancreatic adenocarcinoma	positive on IHC staining	G/GEJ ca.: ≥3rd, pancreatic ca. ≥2nd	Not listed	CT041 autologous CAR T-cell injection	-	Asia^i	-	NCT04581473

EOX: Epirubicin + Oxaliplatin + Capecitabine, CAPOX: Capecitabine + Oxaliplatin

OS, overall survival; mPFS, median progression-free survival; ORR, objective response rate, mDOS: median duration of response, mo: months

^awith moderate-to-strong membranous staining

^bstatistically significant

^cBulgaria, Germany, Latvia, Lithuania, and Switzerland

^dBulgaria, Czechia, Germany, Latvia, Russian Federation, Ukraine

^eUnited States, France, Italy, Japan, Republic of Korea, and Taiwan

^fAustralia, Belgium, Brazil, Canada, Chile, China, Colombia, France, Germany, Israel, Italy, Japan, Republic of Korea, Mexico, Peru, Poland, Spain, Taiwan, United Kingdom, United States

^gArgentina, Austria, Canada, China, Croatia, Greece, Ireland, Japan, Republic of Korea, Malaysia, Netherlands, Portugal, Romania, Spain, Taiwan, Thailand, Turkey, United Kingdom, United States

^hUnited States, Australia, France, Ireland, Italy, Spain, Japan, Republic of Korea, and Taiwan

ⁱChina

Another therapeutic approach for claudin-18.2 is chimeric antigen receptor (CAR) T immunotherapy. CAR T cells can recognize tumor-associated antigens and promote T cell proliferation, cytotoxicity, and the transcription of genes encoding cytokines, resulting in antitumor activity.⁵² The CAR T cells developed by Jiang et al. achieved partial or complete tumor elimination in a claudin-18.2-positive patient-derived tumor xenograft model of gastric cancer.⁵³ The CAR T cells had no obvious off-target effects, including that on gastric tissue, even though the expression level of murine claudin-18.2 in the gastric mucosa was similar to that of human claudin-18.2 in gastric cancer cell lines. The authors noted two possible reasons for the absence of adverse events in the normal gastric mucosa: first, claudin-18.2 of cancer cells is located in the tight junction fraction and on the basolateral membrane, and second, the microenvironment of the normal tissue protects it from cytotoxic effects of CAR T cells.⁵³ The ongoing clinical phase I trial (NCT03159819) using CAR T cells against claudin-18.2 in advanced gastric or pancreatic adenocarcinoma also reported no severe gastric toxicity or cytokine release syndromes.⁵⁴ Furthermore, the total objective response rate (ORR) was 33.3%, and as of 2018, one of 11 study participants with gastric cancer achieved complete remission. Currently, there is one clinical phase I/II trial with CAR T cells (NCT04581473) (Table 1). It is expected to elucidate the mechanism of CAR T cell-specific adverse events through in vivo experiments, and the benefits and toxicity risks for enrolled patients.

Pancreatic cancer

Claudin-18.2 is highly expressed in pancreatic adenocarcinoma and pancreatic cancer metastases. IHC analysis showed that 59.2% of primary pancreatic adenocarcinoma, 69.4% of metastatic lymph nodes, and 65.7% of liver metastases are claudin-18.2 positive. In contrast, only 20% of neuroendocrine neoplasias are claudin-18.2-positive.⁵⁵ In addition, claudin-18.2 is expressed in atypical and cystic lesions such as pancreatic intraepithelial neoplasia, intraductal papillary mucinous neoplasm, and mucinous cystic neoplasm.^56–59 As claudin-18.2 expression is pronounced in well-differentiated pancreatic cancers, patients with high expression levels of claudin-18 survive longer than those with low claudin-18 expression.⁵⁶

Claudin-18.2 is regulated at the transcriptional level via PKC signaling pathways in human gastric cancer,³⁴ pancreatic cancer, and normal pancreatic duct epithelial cells (HPDEs).⁵⁹ Claudin-18.2 mRNA and claudin-18 protein are markedly induced by the PKC activator 12-O-tetradecanoylphorbol-13-acetate (TPA) in well- and moderately differentiated human pancreatic cancer cell lines and HPDEs transfected with the human telomerase reverse transcriptase (hTERT) gene. In pancreatic cancer cell lines, TPA-induced claudin-18 expression is localized on apical and basolateral cell surfaces. The PKC family contains at least 12 different isozymes. Activation of PKCα, PKCδ, and PKCε has been reported to be associated with the upregulation of claudin-18 in human pancreatic cancer cell lines, whereas PKCα, PKCδ, and PKCθ activation has been associated with the upregulation of claudin-18 in hTERT-transfected HPDEs.⁵⁹ Furthermore, the increase in TPA-induced claudin-18 expression is enhanced by DNA demethylation. Consequently, the regulation of claudin-18.2 is correlated with genomic hypomethylation of promoter CpG islands.⁶

Türeci et al. revealed that zolbetuximab induced ADCC and CDC against human pancreatic cancer cells in ex vivo models using human peripheral blood mononuclear cells and serum as effectors.⁴⁵ They also showed that zolbetuximab suppressed tumor development and lung metastasis formation in human pancreatic cancer cell lines transduced with lentiviral claudin-18.2 in mouse xenograft models.⁶⁰ Notably, claudin-18.2 expression on the cell surface was increased by gemcitabine or 5-fluorouracil (chemotherapy agents) administration in vitro. This phenomenon means that even if pancreatic cancer cells are not killed by chemotherapy, the patients can be newly eligible for zolbetuximab therapy, owing to the increased expression of claudin-18.2. It is unclear why this phenomenon occurred, and there is no mention of it in previous literature; however, it could be useful information for zolbetuximab use in the treatment of pancreatic cancer patients.

Cholangiocarcinoma

The upregulation of claudin-18.2 in cholangiocarcinoma is thought to have potential clinical use in the detection of cancer cells because it is difficult to diagnose benign or malignant types based on small preoperative bile duct specimens. The expression level of claudin-18 increased in intra- and extrahepatic cholangiocarcinoma and biliary intraepithelial neoplasia.^61,62 Claudin-18 IHC assays were useful to distinguish the neoplastic region from the non-neoplastic region in surgical specimens and preoperative biopsies.⁶² The assay also improved the sensitivity of the cytological diagnosis of pancreaticobiliary adenocarcinoma.⁶³

As noted in the previous section, claudin-18.2 in pancreatic cancer is an indicator of well-differentiated cancer types and good prognosis,⁵⁶ however, it was shown to correlate with poor overall survival and lymph node metastasis in intrahepatic cholangiocarcinoma.⁶¹ Claudin-18 induced growth and invasiveness of bile duct cancer cell lines.⁶⁴ Notably, wound edges of confluent tumor cells showed strong claudin-18 expression in the wound healing assay. The authors demonstrated that claudin-18 in cancer cells activated ERK1/2, suggesting that claudin-18 plays a role in biliary carcinogenesis.⁶⁴ Claudin-18 was absent in the normal epithelium of the gallbladder; however, high expression of claudin-18 was detected in almost all metaplastic cells and half of the cancer cells.⁶⁵ Its expression was associated with metaplastic change, which was indicated by SOX2 positivity; however, it was not associated with pathological T stage and histological differentiation.

Lung cancer

The role of claudin-18.2 in lung cancer is still not well defined. Therefore, in this section, we will focus on the lung-specific isoform claudin-18.1. Claudin-18.1 is highly expressed in the lung alveolar epithelium and is not detectable in normal airways and lung endothelium.^29,66 Claudin-18 was shown to play a role in the epithelial barrier function of lung alveolar epithelium in claudin-18 knockout mice.^67,68 Claudin-18.1 was expressed at lower levels in asthmatic patients than in healthy controls.⁶⁹ In mice, claudin-18 knockout caused enlargement of the lung and other organs, increased the proliferation of alveolar epithelial type II (AT2) cells, which are lung progenitor cells, and tumorigenesis via YAP activation. Thus, claudin-18.1 has been suggested to regulate the proliferation of AT2 cells in normal lung epithelial cells and promote regeneration following lung injury by disrupting tight junctions.⁷⁰

Claudin-18.1 expression in human lung adenocarcinoma was also downregulated with cancer progression,⁷⁰ and a minority of non-small cell lung cancers showed claudin-18.2 expression (30 of 73 cancer specimens).⁶ Claudin-18 expression was lower in lung cancer than in normal lung tissue, regardless of whether it was small cell or non-small cell carcinoma.⁷¹ Claudin-18 suppressed the proliferation and motility of lung adenocarcinoma cells by inhibiting the PI3K/PDK1/Akt signaling pathway⁷¹ and regulating ZO-2 and matrix metalloproteinase 2.⁷² Notably, the latter two investigations did not differentiate between claudin-18.1 and −18.2.

Clinical advances using monoclonal antibodies targeting claudin-18.2

Clinical trials have shown the efficacy of monoclonal antibodies against claudin-18.2, such as zolbetuximab (also known as IMAB362 and claudiximab).⁵⁰ Zolbetuximab is a structurally chimeric IgG1 monoclonal antibody that specifically binds to the first extracellular loop of claudin-18.2 on the tumor cell surface. This antibody causes ADCC and CDC^45,73 and these cytotoxic effects induce apoptosis and inhibit proliferation of tumor cells, resulting in beneficial effects for cancer patients.⁷⁴ This section will focus on the results of recent clinical trials using zolbetuximab, viz. the MONO and the FAST studies, and summarize their main findings.

In the MONO study, a phase II investigation using zolbetuximab as a single agent in patients with metastatic or advanced gastric/gastroesophageal junction/esophageal adenocarcinoma, patients with moderate or strong claudin-18.2 membrane staining intensity in ≥50% of tumor cells were included.⁴⁵ The treatment showed a 9% ORR (the proportion of patients with complete and partial response in terms of tumor size) and 23% clinical benefit rate (partial response and stable disease).⁴⁵ Moreover, in a subgroup of patients with claudin-18.2 expression in ≥70% of tumor cells, the ORR increased up to 14%.

Recently published results of the FAST study, a phase II investigation with zolbetuximab plus epirubicin, oxaliplatin, and capecitabine (EOX) vs. EOX alone, demonstrated that zolbetuximab had antitumor activity in patients with claudin-18.2-positive advanced gastric/gastroesophageal junction/esophageal adenocarcinoma.⁵⁰ Claudin-18.2 positivity was defined as ≥40% of tumor cells with moderate or strong staining intensity in the CLAUDETECT™18.2 IHC assay. The

CLAUDETECT 18.2 Kit (developed by Ganymed and now acquired by Astellas) was introduced for in vitro diagnosis of a semi-quantitative IHC assay for claudin-18 protein expression.^45,49,73,75 The authors defined three arms: arm 1, EOX alone; arm 2, zolbetuximab + EOX (loading dose 800 mg/m², then 600 mg/m²); arm 3, zolbetuximab + EOX (1000 mg/m²). Of the 686 patients, 334 (49%) had claudin-18.2-positive tumor cells. In arms 1 and 2, 70.2% and 74.0% of the tumors had ≥70% cells with claudin-18.2 staining, respectively. Between arms 1 and 2, both progression-free survival (hazard ratio [HR], 0.44; 95% confidence interval [CI], 0.29–0.67; P < .0005) and overall survival (HR, 0.55; 95% CI, 0.39–0.77; P < .0005) were significantly improved with zolbetuximab + EOX compared to EOX alone. This significant benefit of progression-free survival was retained in patients with claudin-18.2 expression in ≥70% of tumor cells (HR, 0.38; 95% CI, 0.23–0.62; P < .0005). However, in patients with 40–69% claudin-18.2 expression in tumor cells, progression-free survival and overall survival were not significantly different between the arms (HR, 0.78; 95% CI, 0.40–1.49; P = .401). Therefore, we speculate that the inclusion criteria for current clinical trials examining the effects of zolbetuximab are “the expression of claudin-18.2 in ≥75% of tumor cells”.^7–9,50 Moreover, high-dose zolbetuximab (arm 3) showed no significant improvement in overall survival compared with arm 1 (HR, 0.68; 95% CI, 0.44–1.05; P = .1285). The addition of zolbetuximab to EOX was associated with significant increases in ORR (39.0% vs. 25.0%; P = .034), response duration (32.6 months vs. 21.7 months; P = .023), and time to progression (39.1 months vs. 31.3 months; P = .0017), compared to EOX alone in the overall population. The authors suggest that these outcomes may be related to tumor inhibition and cytotoxic activity in tumor cells through ADCC and CDC.⁵⁰

The most common adverse events related to zolbetuximab plus EOX were nausea (81.8%), vomiting (67.5%), anemia (45.5%), and neutropenia (44.2%). The zolbetuximab plus EOX group showed no increase in the frequency of gastric hemorrhages, and the addition of zolbetuximab did not increase adverse events of any grade. In the zolbetuximab plus EOX arm, the incidence of vomiting was lower in patients who previously underwent total or partial gastrectomy (38.1%) than in patients who did not undergo gastrectomy (78.6%), whereas such a difference was not observed in the EOX alone arm. The profile of adverse events in the FAST study was similar to that in the MONO study.⁴⁵

The results of the MONO and FAST studies indicate that zolbetuximab has an antitumor effect on gastric cancer and has little effect on the normal gastric mucosa, despite the expression of claudin-18.2 in normal gastric mucosal and cancer cells. Zolbetuximab resulted in tumor shrinkage, prolonged survival duration, and a low incidence of adverse events such as gastric ulcer and perforation. It is speculated that most anti-claudin-18.2 antibodies cannot bind to normal gastric mucosal cells because the expression of claudin-18.2 in cancer cells is abundant on the membrane including the basolateral region, while its expression in normal mucosal cells is limited to inside the tight junction complex.⁵⁰ Two possible mechanisms may be responsible for this. One is that targets on the apical membrane are further from the microvasculature and are more difficult to reach than targets on the basolateral membrane, and the other is that IgG antibodies have difficulty binding to the target which is surrounded by several tight junction components. However, these speculations have not been confirmed. To confirm this hypothesis, further molecular experiments are essential and will aid to further our understanding of anti-cancer treatment using zolbetuximab.

Based on currently available results of phase I/II studies including that of MONO and FAST, two phase II studies (NCT03816163 for metastatic pancreatic cancer⁵¹ and NCT03505320 for advanced G/GEJ cancer, ILUSTRO⁷) and two phase III studies (NCT03504397, SPOTLIGHT⁸ and NCT03653507, GLOW⁹ for advanced G/GEJ cancer) are ongoing to investigate the efficacy of zolbetuximab (Table 1). Table 1 shows ongoing phase II/III clinical trials with the main characteristics and results using zolbetuximab and CAR T cells. Claudin-18.2 staining in cancer cells has been increased to ≥75% in the inclusion criteria of recent studies, as opposed to 40–50% in the MONO and FAST trials. Thus far, the trials have only included patients in Europe; however, they will now include patients from all parts of the world. As mentioned in the “Gastric Cancer” section, the detection rate of the claudin 18.2 protein varies depending on the antibody clone and race. The clones of antibodies used to evaluate the expression of claudin-18.2 in ongoing studies are not currently listed. Evaluating the claudin-18.2 expression is also important in determining the patient’s indication for the future treatment.⁴⁹

Conclusion

Normally, claudin-18.2 is strictly expressed at the surface of differentiated epithelial cells of the gastric mucosa and prevents the development of gastritis,^31,32 whereas claudin-18.1 is expressed in lung alveolar epithelium and is thought to be responsible for regeneration during lung injury.⁷⁰ Claudin-18.2 in gastric cancer cells assists cancer progression, fusing with ARHGAP26.^37,38 Furthermore, the expression of claudin-18.2 was transcriptionally upregulated in several malignancies.⁶ Upregulated claudins in cancer cells are appropriate targets for anti-cancer treatment because of their dysregulated location and abundant expression. Carcinogenesis alters their localization from the apical side to the whole membrane, resulting in the loss of polarity. Antibodies or drugs targeting claudins in the bloodstream cannot reach the apical side or the tight junctions of normal cells. In cancer cells, claudins on the basolateral membrane are not bound to neighboring tight junction proteins; thus, they could be good targets for directly binding therapeutics. Unlike other claudins, claudin-18.2 is a promising target in several types of cancer because of its absence in normal tissues, except the non-neoplastic gastric mucosa, and its overexpression on the surface of cancer cells. By expanding the application of the anti-claudin-18.2 antibody to treat other cancers including pancreatic, esophageal, ovarian, and lung cancer,⁶ we will have more therapeutic options for incurable diseases. Anti-claudin antibodies will damage cancer cells through multiple mechanisms, including the induction of the complement system and immune cell aggregation, and the inhibition of intracellular signals required for cancer progression⁶ (Figure 3). To understand the dynamics of cancer cells following antibody binding, we need to carefully observe the results of basic molecular and clinical studies. Further investigation of claudins, which are highly and specifically expressed in cancer cells, will lead to the development of powerful new treatment options for advanced cancer patients with poor prognoses.

Figure 3. Action of zolbetuximab on claudin-18.2 of cancer cells.

In normal epithelial cells, tight junction molecules are complexed and expressed on the apical side, making it difficult for zolbetuximab to bind to claudin-18.2. In contrast, claudin-18.2 is overexpressed in cancer cells and is also expressed on the basolateral side, suggesting that zolbetuximab binds easily to claudin-18.2 in cancer cells. Zolbetuximab shows a cytotoxic effect through the action of antibody-dependent cell-mediatedcytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC).

Author contributions

D.K. wrote the manuscript. All authors reviewed and revised the manuscript.

Acknowledgments

The authors would like to thank Editage for their assistance with English language review.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by JSPS KAKENHI under the grant number [JP19K24018]