Retrospective analysis of Claudin 18.2 expression in 64 patients with advanced gastrointestinal peritoneal disease: a potential therapeutic target for high-grade appendiceal mucinous carcinoma

Abstract

Aims

Specific to normal gastric epithelium, Claudin 18.2 (CLDN18.2) is a tight junction protein. CLDN18.2 immunohistochemistry is a companion diagnostic used to identify oesophageal and gastric adenocarcinoma patients eligible for targeted therapy with zolbetuximab. It has demonstrated survival benefits in advanced upper gastrointestinal (UGI) cancers, with many studies now exploring its potential use beyond the UGI tract. This study aimed to examine CLDN18.2 expression in a peritoneal malignancy patient cohort.

Methods

CLDN18.2 immunohistochemistry was performed on archived tissue from patients who underwent cytoreductive surgery at a national centre for peritoneal malignancy. CLDN18.2 immunoprofiles were compared in two key tumour cohorts: peritoneal metastases from colorectal adenocarcinoma and appendiceal mucinous neoplasms. Positivity was defined as moderate-to-strong membranous staining in >75% of tumour cells. In addition, any CLDN18.2 positivity in tumour cells (>0% of cells showing moderate-to-strong membranous staining) was recorded in all cases.

Results

64 cases were examined; primary tumour sites of origin were colorectal (n=34, 53%) and appendiceal (n=30, 47%). CLDN18.2 positivity in the overall appendiceal group was 17% compared with 3% in colorectal adenocarcinomas (p=0.090). In the appendiceal group, CLDN18.2 positivity was only demonstrated in high-grade appendiceal mucinous carcinomas (22%, 5 of 23). CLDN18.2 positive staining (>0% of tumour cells) was significantly increased in the appendiceal group compared with colorectal adenocarcinoma (53% vs 3%, respectively) (p<0.0001).

Conclusions

These results demonstrate CLDN18.2 expression in a subset of appendiceal peritoneal metastases. This is one of the first focused studies showing CLDN18.2 expression in this tumour group, perhaps highlighting a future role for targeted therapies in those with limited curative oncological options.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Claudin 18.2 (CLDN18.2) is a tight junction protein under normal circumstances specific to gastric mucosa. It is expressed in more than one-third of advanced gastric and oesophageal adenocarcinomas. In recent years, CLDN18.2 has gained popularity in its role as a companion diagnostic to zolbetuximab, a monoclonal antibody with proven survival benefits in advanced upper gastrointestinal (UGI) adenocarcinomas.

WHAT THIS STUDY ADDS

• CLDN18.2 expression beyond the UGI tract is heavily under investigation. However, little is known about CLDN18.2 expression in appendiceal neoplasms. Using an advanced peritoneal cancer cohort, this study demonstrates CLDN18.2 positivity in 22% of high-grade appendiceal mucinous carcinomas.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• CLDN18.2 expression in appendiceal neoplasms may play a future role in targeted therapies for this rare tumour group.

Introduction

Zolbetuximab is an anti-Claudin18.2 monoclonal antibody, with a companion diagnostic immunohistochemical biomarker for advanced and inoperable oesophagogastric cancers. It has shown meaningful survival benefits in those with upper gastrointestinal (UGI) adenocarcinomas. Initial phase 2 clinical trials demonstrated almost a doubling of overall survival in those with inoperable disease.¹ The SPOTLIGHT and GLOW phase 3 trials demonstrated more modest yet significant survival advantages when given in combination with standardised chemotherapy.^{2 3}

Under normal circumstances, Claudin 18.2 (CLDN18.2) expression is unique to gastric epithelial cells.^{4 5} Retained CLDN18.2 expression is seen in >70% of metastatic UGI adenocarcinomas. However, when applying the zolbetuximab therapeutic cut-off of >75% of tumour cells with 2+ (moderate) and/or 3+ (strong) membranous CLDN18.2 positive staining,^{2 3} only approximately one-third of patients fulfil the criteria for treatment.^{1–4 6 7} Studies have shown that the signet-ring cell carcinoma subtype demonstrates higher rates of CLDN18.2 positivity,^{4 7} possibly benefiting those with the more aggressive histologic variant where other targeted and immunotherapy options have been less effective.^{8 9}

While zolbetuximab is not yet approved for any purpose other than in advanced oesophageal and gastric adenocarcinomas, in recent years, studies have begun to examine CLDN18.2 expression beyond these UGI sites.^{10 11} To date, there has not been a specific focus on a patient cohort undergoing cytoreductive surgery (CRS) for advanced peritoneal disease. Furthermore, there is an absence of published data on CLDN18.2 expression in appendiceal malignancies, a rare tumour type, often treated with colorectal cancer chemotherapy regimens given a lack of high-level data.^12–16 This is despite proven differences in the molecular profiles between these different tumour groups.¹² It therefore seems justified, given that zolbetuximab is indicated for those with advanced and inoperable disease, although gastric in type, to explore whether it should be considered as part of the treatment armamentarium in other gastrointestinal (GI) peritoneal cancers.

The National Centre for Peritoneal Malignancy at the Mater Misericordiae University Hospital, Dublin, Ireland, has extensive clinical and published experience of multimodal treatment for appropriately selected patients with advanced peritoneal malignancy of colorectal, appendiceal and ovarian origin.^17–20 Using this unique patient cohort, the aim of this study was to examine the expression of CLDN18.2 in patients with peritoneal disease where the primary tumour site of origin was beyond the UGI tract.

Materials and methods

Patient selection

A retrospective review of a prospectively maintained peritoneal malignancy database identified 605 patients (387 GI and 218 gynaecological malignancies) who had undergone CRS at the Mater Misericordiae University Hospital between April 2016 and August 2024. A subset of CRS GI cases of colorectal and appendiceal origin (the predominant subtypes) were analysed, selected based on sufficient adequate tumour cellularity in archived formalin-fixed, paraffin-embedded (FFPE) tissue for biomarker assessment. For appendiceal mucinous neoplasms, the Peritoneal Surface Oncology Group International (PSOGI) classification system was applied.^{20 21} Histologic confirmation of the primary site of tumour origin was determined by GI pathologists, using site-specific immunohistochemistry (CDX2, SATB2 and PAX8 immunostains) as necessary, and with multidisciplinary review of relevant clinical details. In many cases, the primary diagnostic slides were reviewed in tandem to facilitate the final diagnosis. Approval for this study was granted by the Mater Misericordiae University Hospital Research and Ethics Institutional Review Board.

Cytoreductive surgery (CRS)

CRS was performed with the aim of achieving a complete surgical clearance of the primary tumour and all sites of peritoneal metastases through a combination of peritonectomy and visceral resections. The surgical cytoreduction was followed in selected cases by the delivery of heated intraperitoneal chemotherapy, with the aim of eradicating presumed residual microscopic disease. Most patients with metastatic colorectal cancer and high-grade appendiceal cancer received systemic chemotherapy as part of a multimodal perioperative approach.

CLDN18.2 immunohistochemistry and scoring

Tissue was sectioned at a thickness of 4 µm. CLDN18.2 immunohistochemistry was performed using the Ventana CLDN18 (43–14A) RxDx assay as per manufacturer’s (Roche, USA) guidelines. A CLDN18.2 percentage and intensity score were recorded for all examined cases. Specifically, CLDN18.2 tumour status was assessed, applying the published cut-off for CLDN18.2 positivity (≥75% of tumour cells with moderate (2+)-to-strong (3+) CLDN18.2 membranous staining). Scoring was performed by a specialist trained GI pathologist (NJOF). Double scoring was performed by a second registered pathologist (AT), blinded to the original CLDN18.2 percentage scoring, for all cases with >75% positive staining and borderline positivity and selected representative negative and low-percentage positive immunostaining cases.

A further subanalysis was performed, grouping moderate (2+)-to-strong (3+) CLDN18.2 membranous staining into null expression (0% of tumour cells showing CLDN18.2 staining), low expression (>0%–<25% of tumour cells showing CLDN18.2 staining), moderate expression (25%–<75% of tumour cells showing CLDN18.2 staining) and high-expression/CLDN18.2 positive (≥75% of tumour cells showing CLDN18.2 staining).

Statistical analysis

Data were analysed with GraphPad Prism V.10.6.1 (GraphPad Software, San Diego, California, USA). Differences between continuous variables were calculated using Mann Whitney U and Kruskal Wallis tests, where appropriate. Differences between categorical variables were calculated using Fisher’s exact test. Statistical significance was defined as p≤0.05.

Results

Patient clinicopathologic features

64 cases were examined. Patients were predominantly female (n=43, 67%), with a median age of 60 years (range: 28–78) (table 1). The primary sites of tumour origin were colorectal adenocarcinoma (n=34, 53%), followed by appendiceal neoplasms (n=30, 47%). The appendiceal tumour group consisted predominantly of mucinous carcinomas with associated pseudomyxoma peritonei (n=28, 93%) and two cases (7%) of goblet cell adenocarcinoma with associated peritoneal carcinomatosis.

Table 1. Clinicopathologic differences between primary colorectal and primary appendiceal tumour groups (n=64).

	Primary colorectal (n, %)	Primary appendiceal (n, %)	P value
Number of patients (%)	34 (53)	30 (47)
Sex			1.0
Female	23 (68)	20 (67)
Male	11 (32)	10 (33)
Age			0.3931
(Median, range)	60 (28–78)	58 (28–77)
Claudin 18.2 status			0.0901
CLDN18.2 positive	1 (3)	5 (17)
CLDN18.2 negative	33 (97)	25 (83)
Percentage of Claudin 18.2 positive tumour cells			<0.0001
Median % (range)	0 (0–100)	5 (0–95)
Claudin 18.2 categorisation			<0.0001
CLDN18.2 null expression	33 (97)	14 (47)
CLDN18.2 low expression (>0–<25)	0	6 (20)
CLDN18.2 moderate expression (25-<75)	0	5 (17)
CLDN18.2 positive/high expression (>75)	1 (3)	5 (17)

The appendiceal tumour group demonstrated significantly increased CLDN18.2 moderate to strong membranous staining in any tumour cells compared to the colorectal group (53% vs 3%, resepctively).

CLDN18.2, Claudin 18.2.

CLDN18.2-positive (≥75% moderate-to-strong membranous staining) group

Figure 1 shows representative images of CLDN18.2 immunostaining in all positive cases. Of the entire cohort, 9% (6 of 64) of cases showed >75% moderate-to-strong positive membranous CLDN18.2 staining, 17% (5 of 30) of appendiceal tumours and 3% (one of 34) of colorectal cancers.

Figure 1. Representative images of all Claudin 18.2 (CLDN18.2)-positive appendiceal (A–E) and caecal (F) adenocarcinomas. (A) High-grade mucinous carcinoma peritonei (with signet-ring cells), appendiceal in origin (i, H&E ×4), with CLDN18.2 immunohistochemistry (IHC) showing strong (3+) membranous staining in 95% of tumour cells (ii, ×4 and iii, ×20). (B) High-grade appendiceal mucinous carcinoma peritonei (with signet-ring cells) (i, H&E x4), with CLDN18.2 IHC showing moderate (2+) to strong (3+) membranous staining in 90% of tumour cells (ii, ×4 and iii, ×20). (C) High-grade appendiceal mucinous carcinoma peritonei (no signet-ring cells identified) (i, H&E ×4), with CLDN18.2 IHC showing strong (3+) membranous staining in 80% of tumour cells (ii, ×4 and iii, ×15). (D) High-grade mucinous carcinoma peritonei, appendiceal in origin (no signet-ring cells identified) (i, H&E x4), with CLDN18.2 IHC showing moderate (2+) to strong (3+) membranous staining in 75% of tumour cells (ii, ×4 and iii, ×20). (E) High-grade mucinous carcinoma peritonei, appendiceal in origin (no signet-ring cells identified) (i, H&E ×4), with CLDN18.2 IHC showing moderate (2+) to strong (3+) membranous staining in 75% of tumour cells (ii, ×4 and iii, ×20). (F) Metastatic caecal adenocarcinoma (signet-ring cell type) (i, H&E ×4), with CLDN18.2 IHC showing strong (3+) membranous staining in 100% of tumour cells (ii, ×4 and iii, ×20). CLDN18.2, Claudin 18.2.

Comparatively, there was a trend towards increased CLDN18.2 positivity in the appendiceal cancers (p=0.090) (table 1), in which all were confirmed invasive mucinous adenocarcinoma (PSOGI high-grade mucinous carcinoma peritonei) (table 2) - 40% (2 of 5) with signet-ring cells and 60% (3 of 5) without an identifiable invasive signet-ring cell component.

Table 2. Clinicopathologic features of patients with primary appendiceal tumour site of origin, stratified based on Claudin 18 (clone 43.14A) expression.

	Total number, N (%)	CLDN18.2 negative (n, %)	CLDN18.2 positive (n, %)	P value
Number of patients (%)	30 (100)	25 (83)	5 (17)
Sex				0.3
Female	20 (67)	18 (90)	2 (10)
Male	10 (33)	7 (70)	3 (30)
Age				0.9239
(Median, range)	58 (28–77)	60 (28–77)	56 (34–66)
Histologic grade				0.5562
Low-grade mucinous carcinoma peritonei	5 (17)	5 (100)	0 (0)
High-grade mucinous carcinoma peritonei	23 (77)	18 (78)	5 (22)
Goblet cell adenocarcinoma	2 (7)	2 (100)	0 (0)

CLDN 18.2, Claudin 18.2.

By consensus agreement (NJOF and JA), the one CLDN18.2-positive colorectal cancer was an MSI-high carcinoma, with signet-ring cell morphology (figure 1F). This tumour demonstrated MLH1 and PMS2 nuclear loss, secondary to MLH1 promoter methylation. Histologically it was determined as caecal in origin, with direct invasion into and extensively involving the appendix.

CLDN18.2 expression (any percentage of tumour cells demonstrating moderate-to-strong membranous staining) in the appendiceal tumour group

When the established therapeutic cut-off of >75% of tumour cells with moderate-to-strong membranous staining was excluded, 53% of patients with appendiceal-related peritoneal disease demonstrated some CLDN18.2 expression (tables 1 and 3). Expression ranged from 5% to 95% of tumour cells with moderate-to-strong membranous staining. This was significantly higher than CLDN18.2 expression seen in 3% of colorectal-associated peritoneal malignancy cases (p<0.0001) (figure 2). Aside from the single previously described primary caecal adenocarcinoma invading the appendix, no other advanced colorectal cancer showed any moderate-to-strong CLDN18.2 expression.

Table 3. 53% (16 of 30) of patients with advanced GI peritoneal disease secondary to a primary appendiceal tumour showed some expression of CLDN18.2.

	Null expression (0%) N, %	Low expression (>0%–<25%) n, %	Moderate expression (25%–<75%) n, %	High expression (≥75%) n, %	P value
Number of patients (%)	14 (47)	6 (20)	5 (17)	5 (17)
Sex					0.6127
Female	10	4	4	2
Male	4	2	1	3
Age					0.2518
(Median, range)	50 (31–74)	63 (28–66)	68 (39–77)	56 (34–68)
Histologic grade					0.6192
Low-grade mucinous carcinoma peritonei	4	1	0	0
High-grade mucinous carcinoma peritonei	9	4	5	5
Goblet cell adenocarcinoma	1	1	0	0

This table demonstrates Claudin 18 (clone 43.14A) expression levels, divided as null expression, low expression, moderate expression and high-expression/CLDN18.2 positive.

CLDN 18.2, Claudin 18.2.

Figure 2. Percentage of Claudin 18.2-positive tumour cells in GI peritoneal cancer groups: Mann Whitney U test demonstrating significant differences in percentage of tumour cells expressing Claudin 18.2 in the appendiceal (median 5%, range 0%–95%) and colorectal (median 0%, range 0%–100%) cancer groups (p<0.0001).

Examination of CLDN18.2 expression at a second site of disease

In the six positive CLDN18.2 cases, an adequate second site of disease for repeat CLDN18.2 staining was considered available for four patients. Two cases were excluded from this sub-analysis due to a mucin-rich disease burden, with insufficient tumour at a second site to facilitate repeat testing.

Of the four with adequate tumour cellularity at a second peritoneal disease site, 75% (3 of 4), inclusive of the one primary caecal adenocarcinoma, showed comparable CLDN18.2 expression, with >75% of tumour cells demonstrating moderate-to-strong membranous immunostaining. One positive appendiceal high-grade carcinoma, with signet ring cells, showed unequivocal negative staining, with only 15% of tumour cells CLDN18.2 positive in the second analysed site, compared with 90% of positive tumour cells in the first examined site.

Discussion

This is one of the first studies to demonstrate CLDN18.2 expression in an advanced appendiceal cancer group. Notably, Wang et al examined in a smaller patient cohort of 10 metastatic appendiceal tumours the expression of CLDN18.2 in the setting of ovarian mucinous carcinomas.²² Their study showed all appendiceal tumours were negative for CLDN18.2 compared with 84% of primary ovarian mucinous tumours using a non-conventional scoring system, as the study predated established cut-offs set out by the SPOTLIGHT and GLOW phase 3 trials.^{2 3} Other than confirming a lower GI tract/appendiceal immunoprofile, there was no specific detail about the morphology of their appendiceal tumour cohort.

In our study, the established cut-off of >75% was used. All CLDN18.2-positive appendiceal tumours occurred in the setting of high-grade appendiceal mucinous and/or signet-ring cell carcinoma (figure 1). However, lower expression rates of >5% but <75% were also seen in a mix of low- and high-grade appendiceal mucinous neoplasms (table 3). Interestingly, this is significantly different from our colorectal cancer cohort. Possible explanations are speculative and do warrant dedicated future investigation. However, one reasonable hypothesis may be a consequence of increased GNAS (guanine nucleotide-binding protein, alpha stimulating) gene mutations seen in appendiceal mucinous neoplasms.^{12 23–26} Studies have demonstrated such mutations are associated with MUC5AC and MUC6 expression, two mucin types which correlate with gastric differentiation. In fact, supporting this theory is a recent study by Iwaya et al on colitis-associated colorectal adenocarcinomas.²⁷ Working under the hypothesis that colitis-associated cancers tend to lose intestinal markers and show aberrant gastric mucin expression, they demonstrated an association between CLDN18.2 and MUC5AC expression in their colitis-associated colorectal adenocarcinomas, concluding that this patient group may be candidates for zolbetuximab therapy.

While it is well established that CLDN18.2 is prevalent in gastric and oesophageal malignancies, it is also increased in pancreatic cancers,²⁸ with initial hope surrounding the therapeutic benefits of zolbetuximab in this cancer type. Disappointingly, however, the results from the registered Phase 2 GLEAM Trial (NCT03816163) did not show any survival advantage when zolbetuximab was given in combination with standard chemotherapy in pancreatic cancer. Relevant to our findings, genomic profiling of appendiceal malignancies has demonstrated a more appreciable molecular overlap between appendiceal tumours and pancreatic ductal adenocarcinomas compared with colorectal adenocarcinomas.¹² In fact, appendiceal mucinous neoplasms share many similarities, such as increased GNAS mutations, with intraductal papillary mucinous neoplasms, a precursor lesion for pancreatic ductal adenocarcinoma.^{29 30} Furthermore, a recent study by Trinh et al showed that these GNAS mutations in pancreatic tumours drive pyloric metaplasia,³¹ supporting our hypothesis that appendiceal malignancies show gastric lineage, perhaps providing a possible explanation for our unique finding of increased CLDN18.2 expression.

We showed CLDN18.2 positivity in 3% of colorectal adenocarcinomas, a finding consistent with other publications. In the largest analysis, comprising 1175 colorectal cancers, it is notable that CLDN18.2 positivity was increased in right-sided colonic tumours, primarily in the caecum and ascending colon.¹⁰ This is consistent with our results, given that the one positive peritoneal colon cancer case originated in the caecum. It is interesting, and perhaps not a coincidence, given these findings, that this outlier also involved the appendix.

Our results showed heterogeneous CLDN18.2 expression within a proportion of the examined positive cases. In pseudomyxoma peritonei, it is important to recognise that this is a challenging entity, with the mucin-rich environment certainly impacting the quality of immunohistochemical staining. It is, therefore, our recommendation, as is standard with all molecular testing, to select the most cellular and representative tumour block. However, despite this, in one positive case (25%) of high-grade appendiceal mucinous carcinoma, a discordance in expression was seen between two different metastatic sites of disease. This phenomenon of CLDN18.2 heterogeneity is recognised in gastric cancer; however, acceptable levels of more than 80% concordance in expression have been reported between primary and metastatic tumour sites.³² CLDN18.2 heterogeneity is a pitfall, which does have the potential to impact response to treatment.³³ Kim et al³³ in a study of 400 advanced gastric carcinoma cases saw a difference of 61% in CLDN18.2 expression between four different sites on tissue microarray cores. The group concluded that this intratumoral heterogeneity needed further research due to its potential prognostic implications. Our findings of CLDN18.2 heterogeneity in peritoneal disease seem consistent with those seen in UGI malignancies.

We recognise that our study size of 64 samples is small. In truth, this reflects the rarity of this cancer type, and the challenges accessing adequate tumour for analysis, particularly in the setting of pools of mucin, typically encountered in pseudomyxoma peritonei. Despite this, our study is one of the largest to date demonstrating CLDN18.2 positivity in patients with appendiceal-associated peritoneal metastases. Given only one-third of patients with advanced UGI adenocarcinoma are considered CLDN18.2 positive,^{1–4 6 7} a finding of 17% positive CLDN18.2 expression in appendiceal cancers seems a meaningful result, perhaps highlighting a future role for zolbetuximab in this underexplored cancer type.

Conclusion

17% of peritoneal cancer secondary to appendiceal neoplasms showed CLDN18.2 positivity, with up to 53% demonstrating some CLDN18.2 membranous expression. This molecular phenotype differed from CLDN18.2 expression in the advanced colorectal adenocarcinoma cohort, reinforcing the molecular differences between these tumour groups. These novel findings raise the possibility that in certain appendiceal malignancies, in particular high-grade appendiceal mucinous carcinomas, targeted therapy with zolbetuximab may play a future role.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.