Humanized anti-CLDN1 mAb derived from a rat anti-human CLDN1 mAb (OM-7D3-B3) exhibits preclinical antitumor activity through the suppression of cancer stemness and tumor invasion and reprogramming of the immunosuppressive tumor microenvironment in HCC xenograft models, especially those with WNT/β-catenin signaling activation and an epithelial-mesenchymal transition phenotype (30,37).

The humanized anti-CLDN1 mAb ALE.C04 has also been shown to have preclinical antitumor activity as a single agent and in combination with immune checkpoint inhibitors in HNSCC xenograft models via perturbation of interactions between tumor and stromal cells to reverse extracellular matrix remodeling, tissue fibrosis and T cell immune evasion (40,41).

ALE.C04 has received fast track designation by the US Food and Drug Administration (FDA) for treatment of recurrent or metastatic CLDN1-positive HNSCC (42). A phase I/II clinical trial of ALE.C04, as a monotherapy and in combination with pembrolizumab, is ongoing, with an estimated completion date of February 2028 [trial no. National Clinical Trial (NCT)06054477; Table II].

The bsAb ASP1002, which targets CLDN4 and 4-1BB, has been shown to induce antitumor activity in cancer with CLDN4 upregulation through costimulatory T cell signaling activation and subsequent T cell proliferation and cytokine production (43); however, to the best of our knowledge, single-chain variable fragment (scFv) components targeting CLDN4 and CD137 and their epitopes, in vitro functions and in vivo antitumor effects of ASP1002 are unknown. ASP1002 is in a phase I clinical trial for the treatment of CLDN4-positive solid tumors, such as colorectal, ovarian and prostate cancer, NSCLC, TNBC and urothelial carcinoma (trial no. NCT05719558; Table II).

ASP1650 is a mouse/human chimeric anti-CLDN6 mAb that exhibited antitumor effects on ovarian cancer and testicular tumor cells via antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity in a preclinical study (50) and has entered a phase I clinical trial for ovarian cancer (trial no. NCT02054351) and phase II clinical trial for male patients with germ cell tumors (trial no. NCT03760081). The objective response rates (ORRs) of ASP1650 were 2 (1/41) and 0% (0/13) in the NCT02054351 and NCT03760081 clinical trials, respectively (51,52).

TORL-1-23 (CLDN6-23-ADC) is an anti-CLDN6 ADC that consists of humanized CLDN6-23-mAb that recognizes the second extracellular loop of CLDN6, a protease cleavable linker and a monomethyl auristatin E (MMAE) cytotoxic payload with a drug/antibody ratio (DAR) of 4.1 (53). In a preclinical study, TORL-1-23 was >10-fold more potent than CLDN6-23-mAb and exerted in vivo antitumor effects on bladder, endometrial and ovarian cancer (53). Notably, in a phase I clinical trial for the treatment of advanced cancer, including endometrial, ovarian and testicular cancer (trial no. NCT05103683), the ORR of TORL-1-23 in CLDN6-positive cancer was 32% (7/22) (54).

AMG 794, BNT142 and XmAb541 are bsAbs that simultaneously bind CLDN6 on tumor cells and CD3 on T cells for the elimination of tumor cells through recruitment, activation and proliferation of CLDN6-targeting cytotoxic T cells (55-57). Mouse model experiments have revealed that AMG 794 exerted antitumor effects on epithelial ovarian cancer and NSCLC (55) and that BNT142 exerted antitumor effects on serous ovarian cancer and ovarian teratocarcinoma (56). A phase I clinical trial of AMG 794 for patients with NSCLC, epithelial ovarian cancer and other types of solid tumor (trial no. NCT05317078), a phase I/II clinical trial of BNT142 for CLDN6-positive solid tumors (trial no. NCT05262530) and a phase I clinical trial of XmAb541 for solid tumors (trial no. NCT06276491) are ongoing.

SAIL66 is an anti-CLDN6 tsAb that binds CLDN6 on tumor cells and CD3, as well as 4-1BB on T cells to induce more robust immune reactions in CLDN6-positive tumor cells compared with conventional anti-CLDN6 bsAbs (58). In a preclinical study, SAIL66 successfully enhanced T cell infiltration and mitigated T cell exhaustion via potentiation of 4-1BB-mediated costimulatory signaling in syngeneic mouse model experiments (58). In a phase I clinical trial of SAIL66 for CLDN6-positive solid tumors, the anti-interleukin-6 receptor mAb tocilizumab was added to ameliorate cytokine release syndrome (CRS) caused by enhanced antitumor immunity (trial no. NCT05735366).

BNT211 consists of CLDN6-targeting and 4-1BB-stimulating CAR T cells and a CAR T cell-amplifying RNA vaccine (CARVac) (59). CARVac is a liposomal CLDN6-expressing RNA that induces ectopic CLDN6 expression on antigen-presenting cells, such as dendritic cells and macrophages, for stimulation and expansion of CLDN6-targeting CAR T cells (59). BNT211 exerted antitumor effects on ovarian and lung tumors in a preclinical study (45) and was assessed in a phase I/II clinical trial for the treatment of patients with CLDN6-positive solid tumors (trial no. NCT04503278). Although signs of CRS appeared in ~50% of patients due to enhanced immunity, the ORR of BNT211 was 33% (7/21) (59).

AB011 (65), ASKB589 (66,67), FG-M108 (68), MIL93 (69), osemitamab (70,71), ZL-1211 (72) and zolbetuximab (73-75) are representative anti-CLDN18.2 mAbs with ADCC activity that have entered clinical trials (Table II). ASKB589, FG-M108 and zolbetuximab have been investigated in phase III clinical trials, and a phase III clinical trial of osemitamab is underway (Fig. 3).

ASKB589, FG-M108 and osemitamab are humanized/human mAbs that had manageable safety profiles and were well-tolerated in early-phase clinical trials (trial nos. NCT04632108/NCT05632939, NCT04894825 and NCT04396821/NCT04495296, respectively). Notably, the single-agent ORR of ASKB589 in solid tumors was 22% (2/9) (66) and that of osemitamab was 58% (23/40) (70). The ORRs of ASKB589 + capecitabine and oxaliplatin (CAPOX) without or with anti-programmed cell death protein 1 (PD-1) mAb sintilimab in GEA were 75 (9/12) and 80% (12/15), respectively (66,67). The ORR of FG-M108 + nab-paclitaxel + gemcitabine in PDAC was 50% (7/14) (68) and the ORR of osemitamab + CAPOX + anti-PD-1 mAb nivolumab was 57% (45/79) in GEA (71). Currently, randomized, double-blind phase III clinical trials of ASKB589 + CAPOX + sintilimab vs. placebo + CAPOX + sintilimab (trial no. NCT06206733) and FG-M108 + CAPOX vs. placebo + CAPOX (trial no. NCT06177041) in first-line settings for CLDN18.2-positive GEA are ongoing. Osemitamab was granted orphan drug designation by the FDA and combination therapy of osemitamab + chemotherapy + nivolumab is advancing toward a phase III clinical trial for the treatment of GEA (trial no. NCT06093425).

Zolbetuximab is a mouse/human chimeric mAb with a single-agent ORR of 9% (4/43) in CLDN18.2-positive GEA (73). The combination of zolbetuximab + modified 5-fluorouracil, leucovorin and oxaliplatin chemotherapy (mFOLFOX6) resulted in greater clinical activity than placebo + mFOLFOX6 in the SPOTLIGHT study (trial no. NCT03504397) for GEA with CLDN18.2 upregulation. The median progression-free survival (mPFS) time was 10.6 vs. 8.7 months [hazard ratio (HR), 0.75; 95% CI, 0.60-0.94] and the median overall survival (mOS) time was 18.2 vs. 15.5 months (HR, 0.75; 95% CI, 0.60-0.94) (74). The combination of zolbetuximab + CAPOX improved clinical activity in comparison with placebo + CAPOX in the GLOW study (trial no. NCT03653507) (75). Zolbetuximab was granted priority review designation by the FDA in July 2023 but rejected due to unspecified deficiencies in a third-party manufacturing facility in January 2024 (76). The data were resubmitted to the FDA in May 2024 and decision is expected in November 2024 under the Prescription Drug User Fee Act (77).

ATG-022 (78), AZD0901 (CMG901) (79,80), EO-3021 (CPO102 or SYSA1801) (81), IBI343 (82), LM-302 (83,84) and SOT102 (85) are representative anti-CLDN18.2 ADCs in clinical trials (Table II). ATG-022, AZD0901, EO-3021 and LM-302 are human/humanized anti-CLDN18.2 mAbs conjugated with MMAE via cleavable linkers, whereas IBI343 and SOT102 are anti-CLDN18.2 ADCs connected to the topoisomerase I inhibitor exatecan and a derivative of PNU-159682 anthracycline, respectively (78-85). AZD0901 has been assessed in a phase III clinical trial, and IBI343 and LM-302 are entering phase III clinical trials (Fig. 3).

AZD0901, with a DAR of 4, has been shown to exert direct cytotoxic effects on CLDN18.2-overexpressing tumor cells and bystander killing effects on surrounding tumor cells in preclinical studies, and had a single-agent ORR of 44% (39/89) in GEA with CLDN18.2 upregulation in the KYM901 phase I clinical trial (trial no. NCT04805307) (79,80). AZD0901 also exhibited manageable safety profiles in the clinic. Anemia (62.8%), vomiting (57.5%) and hypoalbuminemia (57.5%) were common treatment-emergent adverse events (TEAEs), and decreased neutrophil count (18.6%) and anemia (13.3%) were the most frequent grade ≥3 TEAEs (79). AZD0901 monotherapy for treatment of CLDN18.2-overexpressing GEA was granted fast track designation by the FDA in April 2022 (86) and is currently in a phase II clinical trial with an expected completion date of May 2025 (trial no. NCT06219941); another phase III randomized clinical trial of AZD0901 monotherapy vs. apatinib, docetaxel, irinotecan, paclitaxel or TAS-102, has an expected completion date of April 2026 (trial no. NCT06346392).

IBI343 had single-agent ORRs of 28 [7/25 patients with PDAC and biliary tract cancer (BTC)] and 40% (4/10 patients with CLDN18.2-positive PDAC) in a phase I clinical trial (trial no. NCT05458219) and a manageable safety profile with any-grade [anemia (37%), nausea (26%), vomiting (26%) and decreased white blood cell count (20%)] and grade ≥3 treatment-related adverse events (TRAEs) [anemia (6%) and decreased white blood cell counts (3%)] (82). IBI343 monotherapy was granted fast track designation by the FDA for the treatment of PDAC (87). By contrast, for the treatment of GEA, a phase II clinical trial of combination therapy of IBI343 + sintilimab (trial no. NCT06321913) and a phase III randomized clinical trial of IBI343 vs. irinotecan or paclitaxel, (trial no. NCT06238843) are planned but not yet recruiting.

LM-302 had superior antitumor efficacy to zolbetuximab in a preclinical study using gastric cancer model (83), and a single-agent ORR of 31% (11/36) in CLDN18.2-positive GEA and a manageable safety profile in a phase I/II clinical trial (trial no. NCT05161390) (84). LM-302 is proceeding to a phase III randomized clinical trial of LM-302 vs. the investigator's choice of therapy, apatinib or irinotecan, for treatment of CLDN18.2-positive GEA (trial no. NCT06351020).

AZD5863 (88), givastomig (89), gresonitamab (90), IBI389 (91), PM1032 (92,93), PT886 (94), Q-1802 (95) and QLS31905 (96) are representative anti-CLDN18.2 bsAbs (Table II).

AZD5863, gresonitamab, IBI389 and QLS31905 simultaneously binds CLDN18.2 on tumor cells and CD3 on T cells (88,90,91,96), which activates CD3 signaling in T cells and redirects cytotoxic T cells toward tumor killing through lytic synapse formation and release of granzymes and perforins (97,98).

Givastomig, PM1032, PT886 and Q-1802 simultaneously bind CLDN18.2 and non-CD3 immune antigens; givastomig and PM1032 enhance antitumor immunity via 4-1BB-induced activation of costimulatory signaling (89,92); PT886 reactivates phagocytosis via CD47-dependent inhibition of 'do not eat me' signaling (94); and Q-1802 inhibits immune checkpoints via programmed death-ligand 1 (PD-L1)/PD-1 signaling blockade (95). Givastomig and PT886 have been granted orphan drug designation by the FDA (94,99).

Compared with anti-CLDN18.2 mAbs and ADCs, anti-CLDN18.2 bsAbs are in relatively early phases of clinical trials (88-96): Givastomig, IBI389 and Q-1802 are in phase I clinical trials; AZD5863, PM1032 and PT886 are in phase I/II clinical trials; and QLS31905 is proceeding to phase I/II clinical trials (Fig. 3).

In a phase I clinical trial, IBI389 caused any-grade TRAEs in 97.4% (111/114) and grade ≥3 TRAEs in 55.3%, including any-grade CRS (57.0%) and grade 3 CRS (0.9%), of patients with solid tumors, such as GEA (n=37) and PDAC (n=66) (trial no. NCT05164458). The single-agent ORR of IBI389 was 31% (8/26) in GEA with CLDN18.2 overexpression (91).

In a phase I/II clinical trial, PM1032 caused any-grade TRAEs in 73% (22/30) and grade ≥3 TRAEs in 10% of gastrointestinal cancer cases (trial no. NCT05839106). The single-agent ORR of PM1032 was 20% (2/10) in CLDN18.2-positive GEA (93).

In a phase I clinical trial (trial no. NCT04856150) of Q-1802, any-grade TRAEs, including nausea (62%; 18/29), vomiting (62%) and abdominal pain (28%), immune-related adverse events (AEs; such as abnormal thyroid function, rash and arthritis) (24%), and grade 3 TRAEs such as nausea and vomiting (24%) and grade 4 TRAE hyponatremia (3%) were observed. The single-agent ORR of Q-1802 was 22% (2/9) in gastrointestinal cancer (95).

QLS31905 treatment was associated with any-grade TRAEs in 98% (51/52) and grade ≥3 TRAEs in 40% of patients treated with 0.5-500.0 mg/kg once/week (qW) or once every 2 weeks (q2W) QLS31905, including grade 3 CRS in 2 patients in the 350 mg/kg qW QLS31905 cohort, in a phase I clinical trial (trial no. NCT05278832). The single-agent ORR of QLS31905 was 11% (3/27) in the phase I study (96), and a phase I/II clinical trial of QLS31905 plus chemotherapy is planned (trial no. NCT06041035).

Satricabtagene autoleucel (Satri-cel) refers to autologous CAR T cells targeting CLDN18.2: CAR T cells were generated from peripheral blood mononuclear cells via lentiviral transduction of a second-generation CAR construct consisting of an extracellular humanized anti-CLDN18.2 scFv and a CD8α hinge region, CD28 transmembrane domain and cytoplasmic CD28 costimulatory and CD3ζ signaling domains. CAR T cells were formulated and infused back following preconditioning combination therapy with cyclophosphamide, fludarabine and nab-paclitaxel or gemcitabine (100,101).

Satri-cel was assessed in a phase I clinical trial (trial no. NCT03874897) for the treatment of solid tumors on the basis of the results of a preclinical study that revealed antitumor effects with persistent infiltration of CAR T cells into CLDN18.2-positive gastric cancer patient-derived xenograft models (100). A total of ~75% of patients in the Satri-cel clinical trial received bridging therapy, such as folinic acid, fluorouracil and irinotecan, nab-paclitaxel or irinotecan, during autologous CAR T cell production (median, 27 days; range, 22-187 days) (101). Satri-cel exhibited tolerability and safety profiles with manageable AEs, such as preconditioning-associated transient hematological toxicity (grade ≥3, 100%), CRS (any grade, 97%; grade ≥3, 0%), nausea (any grade, 67%; grade ≥3, 1%), vomiting (any grade, 53%; grade ≥3, 3%) and gastric mucosal injury (grade 1/2, 7%; grade 3, 1%) and without immune effector cell-associated neurotoxicity syndrome, and clinical activity, as indicated by an ORR of 39% (38/98), mPFS time of 4.4 months (95% CI, 3.7-6.6) and mOS time of 8.8 months (95% CI, 7.1-10.2) (101). Currently, a randomized phase I/II clinical trial of Satri-cel vs. apatinib, irinotecan or paclitaxel, for the treatment of CLDN18.2-positive GEA and PDAC is ongoing (trial no. NCT04581473).

CLDN18.2-targeting CAR T cells, such as LB1908 expressing a CAR with a 4-1BB costimulatory domain (102), IMC002 harboring a CAR with anti-CLDN18.2 variable heavy domain of heavy chain antibody instead of a scFv (103), AZD6422-armed CAR T cells with a dominant-negative TGF-β type II receptor (dnTGFBR2) to overcome the TGF-β-induced immunosuppressive tumor microenvironment (104) and KD-496 bispecific CAR T cells that simultaneously recognize CLDN18.2 and NK group 2 member D (NKG2D) ligands on tumor cells (105), are also in phase I clinical trials for the treatment of patients with CLDN18.2-positive tumors (Table II).

Phase III SPOTLIGHT and GLOW clinical trials for CLDN18.2-positive GEA demonstrated greater clinical activity of zolbetuximab + chemotherapy than placebo + chemotherapy (74,75). Other anti-CLDN18.2 mAbs have been assessed, and ASKB589, FG-M108 and osemitamab are being assessed in randomized phase III clinical trials with expected completion dates of December 2026, January 2027 and October 2025, respectively (Fig. 3). The single-agent ORR of chimeric mAb zolbetuximab is 9% (73), whereas that of humanized mAbs ASKB589 and osemitamab, which have increased affinity and enhanced ADCC effects, is 22 and 58%, respectively (66,70). Previous-generation mAb drugs were chimeric antibodies that elicit host immune responses to remaining variable regions derived from mouse antibody, while current-generation mAb drugs are human/humanized antibodies without mouse-derived variable regions (106). Comparison of the clinical activity (mPFS and mOS) of zolbetuximab and other humanized/human anti-CLDN18.2 mAbs is warranted.

Anti-CLDN ADCs (78-85) are classified into two types. Most ADCs (CLDN6, TORL-1-23; CLDN18.2, ATG-022, AZD0901, EO-3021 and LM-302) include a human/humanized mAb, cleavable linker and an MMAE microtubule inhibitor. Others have cytotoxic DNA-damaging payloads (CLDN18.2, IBI343 and SOT102; Table II). In general, the clinical activity of these anti-CLDN ADCs is affected by mAb epitope affinity and modification, linker chemistry, and payload pharmacology (107,108).

CLDN-targeting ADCs have exhibited improved clinical outcomes over CLDN-targeting chimeric mAbs in early-phase clinical trials; the single-agent ORRs of ATG-022 (78), AZD0901 (79), EO-3021 (81), IBI343 (82), LM-302 (84) and TORL-1-23 (54) were 20, 44, 38, 28, 31 and 32%, respectively, compared with 0-9% for the chimeric anti-CLDN mAbs ASP1650 (51,52) and zolbetuximab (73).

ORRs of the aforementioned anti-CLDN ADCs are similar to those of FDA-approved ADCs targeting other tumor-associated antigens, such as the ORR of 31.5% of anti-trophoblast cell surface antigen 2 ADC sacituzumab govitecan (109) and the ORR of 42.9% of the anti-nectin cell adhesion molecule 4 ADC enfortumab vedotin (110). AZD0901, IBI343 and LM-302 have entered or are entering phase III clinical trials, which will yield mPFS and mOS data in the future.

Anti-CLDN bsAbs in phase I or I/II clinical trials (Table II) are classified into CD3-directed (CLDN6-targeting AMG 794, BNT142 and XmAb541; and CLDN18.2-targeting AZD5863, gresonitamab, IBI389 and QLS31905) and non-CD3-directed (CLDN4-targeting ASP1002; and CLDN18.2-targeting givastomig, PM1032, PT886 and Q-1802) (43,55-57,88-96). IBI389, PM1032, Q-1802 and QLS31905 have single-agent ORRs of 31, 20, 22 and 11%, respectively (91,93,95,96). A study on gresonitamab was terminated due to business decisions of Amgen (trial no. NCT04260191).

In clinical oncology, bsAbs such as amivantamab (targeting EGFR and MET) (111), petosemtamab (targeting EGFR and leucine-rich repeat-containing G-protein coupled receptor 5) (112) and zenocutuzumab (targeting human EGFR 2 and 3) (113) dually target tumor-specific antigens (97,98). Since upregulation of CLDN18.2 (56%) (27), EGFR (27%) (114), HER2 (21%) (115), fibroblast growth factor receptor 2 isoform b (4-5%) (116,117) and MET (24%) (118) in gastric cancer has been detected and mAbs targeting these receptor tyrosine kinases (RTKs) have been developed (111,112,119), bsAbs dually targeting CLDN18.2 and RTKs may serve as antibody-based drugs for the treatment of gastric cancer in the future.

Satri-cel, an autologous CAR T cell therapy with a second-generation CLDN18.2-directed CAR, has an ORR of 39% (101), which is similar to the ORRs of BNT211 (CLDN6-targeting CAR-T cell + RNA vaccine; 33%) (59) and other solid tumor-targeting CAR T cell therapies directed at carcinoembryonic antigen (15%; 6/40) (120), EGFR (18%; 2/11) (121), glypican 3 (50%; 11/22) (122) and guanylyl cyclase 2C (26%; 5/19) (123), but less effective than FDA-approved CAR T cell therapies targeting hematological malignancies, such as the CD19-directed CAR T cells axicabtagene ciloleucel (82%; 83/101) (124) and B cell maturation antigen-directed CAR T cells ciltacabtagene autoleucel (98%; 95/97) (125).

Numerous challenges hinder the activities of CAR T cells, especially those in solid tumors. These include epitope-losing subclonal replacement on the basis of pretreatment heterogeneity and/or posttreatment evolution (126-128), decreased infiltration of CAR T cells into an immunosuppressive tumor microenvironment harboring angiogenic endothelial cells, extracellular matrix-remodeling cancer-associated fibroblasts, regulatory T cells, M2-type macrophages and monocytic myeloid-derived suppressor cells (129-131), and decreased persistence of CAR T cells due to mechanisms similar to the exhaustion of effector T cells with diminished anti-tumor activity dependent on TGF-β signaling as well as PD-L1/PD-1 and other coinhibitory signaling (132-134).

bsCAR T cells dually targeting CLDN18.2 and NKG2D ligands have been developed to address intratumor heterogeneity (105). CLDN6-directed CAR-NK cells, which incorporate the NKG2D transmembrane domain, CD244 costimulatory domain and the DNAX-activation protein 10 signaling domain have been developed to enhance antitumor immunity (60). CLDN18.2-directed CAR T cells armed with dnTGFBR2 have been developed to eliminate immunosuppressive effects of TGF-β (104) and preclinical CAR T cells armed with forkhead-box O1 transcription factor have been developed to maintain the stem/memory cell subpopulation, retain effector functions and prevent the exhaustion phenotype (133,134). These strategies might further improve the clinical activity of CLDN-targeted CAR therapy.