CME
Physicians: Maximum of 1.00 AMA PRA Category 1 Credit™
Released: April 25, 2023
Expiration: April 24, 2024
In this module, Manish A. Shah, MD, discusses emerging data on agents targeting claudin18.2 (CLDN18.2) for the treatment of gastric (G) and gastroesophageal junction (GEJ) cancers.
Please note that the key points discussed in this module are illustrated with thumbnails from the accompanying downloadable PowerPoint slideset, which can be downloaded here or by clicking any of the slide thumbnails in the module alongside the expert commentary.
Clinical Care Options plans to measure the educational impact of this activity. A few questions will be asked twice: once at the beginning of the activity and then again after the discussion that informs the best choice. Your responses will be aggregated for analysis, and your specific responses will not be shared.
Before continuing with this educational activity, please take a moment to answer the following questions.
Gastric cancer is a very prevalent disease worldwide, with more than 1 million patients diagnosed and almost 800,000 deaths yearly.1 The standard of care for patients with unresectable or metastatic disease remains chemotherapy.2 Targeted therapies may provide benefit for some patients based on tumor genetics, such as trastuzumab for HER2-positive disease and nivolumab for those with PD-L1 combined positive score ≥5.2 However, many patients fall outside of those genetically defined subgroups and receive chemotherapy alone as the standard of care, with an expected survival of less than 1 year. There is a clear unmet need for new targeted therapies for the treatment of gastric cancer.
CLDN18.2 is a potential new target that could help fill that unmet need.3 CLDN18.2 is a gap protein that comprises intercellular tight junctions in the gastric epithelia, with minimal surface exposure under normal conditions. The malignant transformation causes the cells to lose polarity and the transmembrane protein to be exposed on either the basal or luminal side of the epithelia. The novelty of CLDN18.2 as a target is that, despite its constitutive expression in normal gastric tissue, it only becomes available for targeting after malignant transformation and structural alterations in epithelial cells.
The level of CLDN18.2 expression is typically scored with immunohistochemistry.4 The categories of immunohistochemical staining can be seen in photomicrographs on the right, from strong reactivity in the top row to negative in the bottom row. Overexpression can occur in all histologic types of gastric cancer: diffuse, intestinal, and mixed. High expression for CLDN18.2 is characterized as ≥75% of cancer cells having moderate to strong immunohistochemistry staining, intermediate as ≥40% of tumor cells, and low as <40% of tumor cells. Approximately 45% of gastric cancers have high or intermediate and 30% have high CLDN18.2 expression levels across the globe, with very similar rates in North America, Asia, and Europe.3 CLDN18.2 expression is also very similar between intestinal and diffuse histologies. To reiterate, these expression levels are not assessments of the protein levels present in the tissue but of protein exposed because of the loss of polarity of the cancer cells. Of interest, there is relatively little overlap between CLDN18.2 and PD-L1 expression. Only 15% to 20% of tumors that highly express CLDN18.2 also have a PD-L1 combined positive score ≥5.
CLDN18.2 expression does not seem to affect response to treatment or clinical outcomes of disease, as seen in this single-center chart review. Patients with high levels of CLDN18.2 expression had similar survival to those with lower or no expression. In addition, outcomes of different therapies, including first-line or second-line chemotherapy and anti–PD-L1 therapy, were equivalent regardless of CLDN18.2 expression level.
Numerous strategies for targeting CLDN18.2 are currently being evaluated.5 The agents farthest along in development are engineered monoclonal antibodies, a modality most widely known from the example of trastuzumab, which targets HER2. Bispecific antibodies target both CLDN18.2 and another antigen such as PD-L1 to bring those 2 molecules close in space. ADCs engage CLDN18.2 to target the delivery of a toxic drug payload. Finally, CLDN18.2-targeted CAR T-cells can bring manufactured lymphocytes into the proximity of cells expressing that protein. It is very exciting to have multiple ways to address this up-and-coming target.
The first-in-class CLDN18.2-targeting agent is the IgG1 monoclonal antibody zolbetuximab.6 The randomized phase II FAST trial examined zolbetuximab in combination with epirubicin, oxaliplatin, and capecitabine (EOX) compared with EOX alone in 161 patients with G/GEJ, or esophageal adenocarcinoma, with a primary endpoint of PFS. Of note, patients enrolled on this trial were required to have CLDN18.2 expression in ≥40% of tumor cells, which corresponds to intermediate or high expression level.
Primary results from the FAST trial were quite encouraging.6 The intention-to-treat population with intermediate or high CLDN18.2 expression level showed that the addition of zolbetuximab to chemotherapy improved median PFS from 5.3 months to 7.5 months, a 56% improvement, which was highly significant. In an analysis of the patient subgroup with high CLDN18.2 expression, comprising approximately two-thirds of the entire trial population, the improvement in median PFS was even more striking: 9 months with zolbetuximab plus EOX vs 5.7 months with EOX alone, for 62% improvement. Based on these data, 2 definitive phase III studies evaluating zolbetuximab in G/GEJ cancer were performed.
The first of 2 global, randomized, double-blind, placebo-controlled phase III studies was SPOTLIGHT, examining zolbetuximab in combination with mFOLFOX6 chemotherapy vs mFOLFOX6 and placebo.7 Zolbetuximab is administered every 3 weeks, whereas FOLFOX is administered every 2 weeks, so this mismatch was accommodated with 6-week treatment cycles.
Zolbetuximab dosing was identical to the protocol used in the phase II study: a single loading dose of 800 mg/m² followed by a continuing maintenance dose of 600 mg/m². The primary endpoint of this study was PFS, but it was powered for OS comparisons as well. Of note, patients were required to have high CLDN18.2 expression (≥75% of tumor cells) to enroll on phase III trials, whereas the phase II trial enrolled patients with intermediate CLDN18.2 expression levels as well (≥40% of tumor cells).
Topline results from SPOTLIGHT presented at ASCO GI in January 2023 showed that the addition of zolbetuximab to FOLFOX significantly improved PFS (median PFS: 10.61 vs 8.67 months; HR: 0.75; P = .0066).7 The curves split after 4-5 months and then continue to remain separated throughout follow-up for a 14% improvement in PFS at 1 year. At 2 years, almost one quarter of patients on the zolbetuximab arm has not progressed, which is an unprecedented and compelling outcome for advanced gastric cancer.
A significant OS benefit also was seen with the addition of zolbetuximab to mFOLFOX6.7 These curves began to split at 9 or 10 months and remained separated, reaching a median OS of 18.23 months with zolbetuximab plus mFOLFOX6 vs 15.54 months with placebo plus mFOLFOX6. Of note, this is the longest median OS reported from a global phase III trial in advanced gastric cancer (HR: 0.75), very similar to that seen for PFS.
Response rates were similar between treatment arms (61% vs 62%). Duration of response was also very similar, approximately 8 months in each arm.7
The most frequent treatment-emergent adverse events seen with zolbetuximab were gastrointestinal (GI): nausea, vomiting, and decreased appetite.7 Grade ≥3 nausea and vomiting were each approximately 16% in the zolbetuximab arm, approximately 10% higher than seen in the control arm. Because CLDN18.2 is expressed in normal gastric mucosa, this adverse event (AE) profile is to be expected and can almost be considered an on-target effect. Very few treatment-related adverse events (TRAEs) led to death (<2% in each arm), and 14% of patients discontinued zolbetuximab because of TRAEs vs 2% with placebo. Overall, zolbetuximab was shown to be associated with some level of toxicity, but these AEs are manageable considering the potential for PFS and OS benefit with this drug.
The second global, randomized, double-blind phase III study of zolbetuximab was GLOW. It also evaluated the addition of zolbetuximab to chemotherapy in patients with previously untreated advanced G/GEJ cancer.8 Here the chemotherapy regimen was CAPOX with zolbetuximab or placebo. CAPOX is administered on 3-week cycles, which aligns with zolbetuximab dosing, so the eight 3-week cycles in this trial resulted in the same number and frequency of zolbetuximab treatments used in the SPOTLIGHT trial. The dose level was also the same as in SPOTLIGHT: 800 mg/m² as the initial loading dose and 600 mg/m² thereafter.
Five hundred, seven patients were treated in GLOW, for a total of more than1000 patients treated across both phase III trials. Of note, in a subset of nearly 300 enrolled patients, 78% had PD-L1 combined positive score <5.
The primary endpoint of centrally assessed PFS was met, with a median of 8.21 months for zolbetuximab added to CAPOX vs 6.80 months for CAPOX and placebo (HR: 0.687; P = .0007).8 This HR is similar to that seen with the addition of zolbetuximab to a different chemotherapy regimen, mFOLFOX6, in SPOTLIGHT (HR: 0.751).
A significant survival benefit also was seen with the addition of zolbetuximab to CAPOX (median: 14.39 vs 12.16 months; HR: 0.771; P = .0118).8 Again, this HR is very close to that seen with the addition of zolbetuximab to mFOLFOX6 in SPOTLIGHT (HR: 0.750). Taken in conjunction, the results from these 2 phase III trials have validated CLDN18.2 as a clinically meaningful target in advanced G/GEJ adenocarcinoma.
Response rates were similar between treatment arms (53.8% with the addition of zolbetuximab vs 48.8% with placebo), as was the duration of response (median: 8.51 vs 8.11 months).8
The most frequent TRAEs were identical to those reported for SPOTLIGHT: nausea, vomiting, and decreased appetite.8 TRAEs leading to death were from 2% to 3% in each treatment arm, and 7.1% of patients discontinued zolbetuximab because of TRAEs vs 2.8% with placebo. Overall, the safety profiles were very similar between the 2 phase III studies.
Nausea and vomiting typically occurred with the first administration of zolbetuximab and could often be managed with antiemetics so that the infusion could be resumed and completed.
It also should be noted that the different chemotherapy dosing schedules used for mFOLOFX6 and CAPOX may have affected the occurrence of AEs. In SPOTLIGHT, one half of the zolbetuximab doses were given alone on days when the patients did not receive chemotherapy and the associated steroid or antiemetic treatments, whereas all zolbetuximab doses in GLOW aligned with chemotherapy and steroids. This may explain why there appears to be less GI toxicity with the CAPOX chemotherapy backbone.
Osemitamab (formerly TST001) is another monoclonal antibody targeting CLDN18.2.9 A phase I study determined 6 mg/kg to be the optimal dose of this agent in combination with CAPOX, and 36 patients have been treated with that dose in the dose-expansion phase. Patients in the dose escalation phase were not required to have CLDN18.2 overexpression, but those recruited for the dose expansion were required to have intermediate or high expression levels (≥40% of tumor cells).
The first reported efficacy results for osemitamab showed a response rate of 73% (11/15 evaluable patients).9 Of note, responses were observed in patients across CLDN18.2 expression levels, including 1 patient with low expression from the dose escalation phase. Safety was considered manageable with primarily GI AEs, which were mostly grade 1/2. Ongoing trials are continuing to explore osemitamab plus CAPOX as well as other chemotherapy combinations.
A third monoclonal antibody, ASKB589, is also in early-phase trials in combination with CAPOX.10 Forty-five patients, again including those with intermediate or high CLDN18.2 expression levels, have been treated with this agent in the dose-expansion phase.
The first reported results for ASKB589 are similar to those seen for osemitamab: a very high response rate (75%; 18/24 patients), including responses in patients with low or moderate CLDN18.2 expression and even at lower doses of the antibody.10 The safety profile was also similar. The most frequent AEs were grade 1/2 GI events. This also seems to be a very potent treatment, and the maturing data are anticipated.
Several ongoing trials of monoclonal antibodies are currently enrolling patients with gastric or gastroesophageal cancer. Zolbetuximab in combination with pembrolizumab is being explored in cohort 3 of the ILUSTRO trial, and osemitamab is under study in several combinations.
Bispecific antibodies are an interesting variation on monoclonal antibody therapy with the potential to bring 2 types of cells together.11 In this case, immune cells expressing PD-1 are brought into close contact with tumor cells expressing CLDN18.2.
The bispecific antibody Q-1802 is currently in phase I development in a patient population including multiple solid tumor types, regardless of CLDN18.2 expression.12 Interim data as of September 2022 are promising, showing no dose-limiting toxicities up to the maximum dose explored. The most common AEs were nausea and vomiting, which can be seen as an on-target effect of CLDN18.2-targeting agents because of their expression in the gastric mucosa. Q-1802 shows promising early activity, with 5/7 patients with CLDN18.2-expressing GI tumors and 3/4 with high CLDN18.2 expression experiencing some benefit.
ADCs use monoclonal antibodies to deliver cytotoxic drugs to cells of interest.13 The drug is typically added to the antibody through a linker that can dissociate once the drug has reached its target.14 One key characteristic of any particular ADC is its drug to antibody ratio: the number of drug molecules linked to each antibody.15 A widely known ADC in GI cancer, trastuzumab deruxtecan, which is approved for the treatment of HER2-positive gastric cancer in patients previously treated with trastuzumab, has a drug to antibody ratio of approximately 8.16
CMG901 is a CLDN18.2-targeted ADC carrying monomethyl auristatin E, which has antitumor activity, including bystander killing and antibody-dependent and complement-dependent cytotoxicity.17 The FDA granted fast-track designation to this agent in 2022. In a phase Ia dose escalation trial, CMG901 was found to be tolerable with 1 dose-limiting toxicity at 2.2 mg/kg. The most common AEs were vomiting and decreased appetite. The appreciable activity was seen even with the small number of patients treated, with a response rate of 75% in CLDN18.2-positive gastric tumors. CMG901 is another very exciting drug in this field and will be continued into the dose-expansion phase for G/GEJ and pancreatic cancers with CLDN18.2 expression.
Two additional CLDN18.2 ADCs, TPX-4589 (also known as LM-302) and EO-3021 (also known as SYSA1801), with monomethyl auristatin E as the payload, are also in early phase development and currently enrolling patients.
The final CLDN18.2-targeting construct under investigation is CAR T-cell therapy.18-20 Essentially, the patient’s T-cells are isolated and modified with an engineered CAR recognizing CLDN18.2, and these modified expanded CAR T-cells are readministered into the patient. The modified T-cells are thus directed to more efficiently attack CLDN18.2-expressing tumor cells.
This diagram more clearly shows the components of the CAR T-cell construct and the role each plays in its function.21 The CAR comprises an intracellular domain that supports T-cell activation and expansion, a transmembrane domain that participates in cytokine release, a single-chain variable fragment domain that specifies the target and determines immunogenicity, and a hinge domain that positions the T-cell to optimally engage the targeted tumor cell.
One CLDN18.2-targeted CAR T-cell therapy has reported results thus far: a very exciting study from China that has been followed up with a parallel US cohort.22,23 CT041 is constructed with an anti-CLDN18.2 single-chain variable fragment, CD8α hinge, CD28 costimulatory domain, and CD3ζ signaling domain. The phase I dose escalation and dose-expansion phases enrolled patients with intermediate or high CLDN18.2 expression level.
The initial trial results for CT041 have shown really remarkable activity.22,23 Response rates for gastric cancer were 57% and 60% in each cohort, and many patients had evidence of tumor shrinkage despite not reaching the threshold for partial response. However, the duration of response was rather short, with only 53% of responses maintained to 6 months, so there is room to improve for this therapy to reach its full therapeutic potential. Of note, the serious toxicities that have been sometimes seen with CAR T-therapy in hematologic malignancies were absent in this trial: no grade ≥3 cytokine release syndromes and immune effector cell–associated neurotoxicity syndromes were observed in either the US or China cohorts.
Trials for CT041 are still open for enrollment, as is a Chinese phase I trial for another CLDN18.2-targeting CAR T-cell therapy, LCAR-C18S.
In summary, CLDN18.2 has been well validated as a promising target for the treatment of upper GI cancers. High expression is seen in 30% to 40% of patients with G/GEJ cancers, with minimal overlap with PD-L1 combined positive score ≥5 or HER2-positive tumors, meaning that it may address an unmet need for those patients who are not currently able to benefit from targeted therapy. The first-in-class agent zolbetuximab may soon change practice for patients with CLDN18.2-positive, HER2-negative gastric cancers. Moreover, numerous other agents targeting CLDN18.2, including other monoclonal antibodies, bispecific antibodies, ADCs, and CAR T-cell therapies, are also under development and have shown encouraging activity.
Back
