68Ga-NC-BCH Whole-Body PET Imaging Rapidly Targets Claudin18.2 in Lesions in Gastrointestinal Cancer Patients

Visual Abstract

Gast rointestinal cancer is a global disease that seriously endangers human public health.In 2018, gastrointestinal cancers accounted for approximately 4.8 million new cases of cancer and 3.4 million deaths worldwide, or more than a quarter of the total cancer incidence and more than one third of the total cancerrelated mortality, respectively (1).Because of the insidiousness of early symptoms, most gastrointestinal cancers are diagnosed at an advanced stage, often leading to a poor prognosis and increased mortality.
Claudin-18 isoform 2 (CLDN18.2) is a subtype of CLDN18, a member of the tight junction protein family (2).CLDN18.2 is involved in the formation of intercellular adhesion structures, cell polarity control, paracellular transport, tissue permeability regulation, and signal transduction (3).Generally, its expression is strictly limited to the normal gastric mucosa; however, CLDN18.2 can also be abnormally activated during the appearance, metastasis, and invasion of gastrointestinal malignancies such as stomach, colon, pancreatic, esophageal, ovarian, and lung tumors (4)(5)(6).The randomized phase IIb study showed that zolbetuximab in combination with epirubicin plus oxaliplatin plus capecitabine improved overall survival and progression-free survival in patients with higher CLDN18.2expression (.70%) relative to those with lower CLDN18.2expression (40%-69%) (7).Therefore, the detection of CLDN18.2 expression levels is essential for identifying patients who can achieve greater clinical benefit.
To date, there is no standard test for CLDN18.2, and most detection methods involve immunohistochemistry (8).Immunohistochemistry is an invasive process that covers only a small amount of tissue and does not reflect the heterogeneity of CLDN18.2 expression within the tumor.Previously, we reported the first-to our knowledge-clinical results of the 124 I-labeled CLDN18.2 humanized monoclonal antibody 124 I-18B10(10L), which showed that CLDN18.2 can be detected in tumor lesions (9).However, the large molecular weight of the monoclonal antibodies results in a long imaging cycle that cannot be completed in a day.Singledomain antibodies are the smallest antibody units with complete functions, stability, and binding antigens; a molecular weight of approximately 15 kDa; and a short circulating half-life in the body with rapid blood clearance, allowing for the matching of radionuclides with short half-lives for same-day PET imaging (10,11).
In this study, we constructed the single-domain antibody molecular probe, 68 Ga-labeled nanobody ( 68 Ga-NC-BCH).With the advantages of whole-body PET, we initiated an open-label, single-center, single-arm, first-in-humans phase 0 trial to study the safety, systemic distribution and dosimetry, and CLDN18.2-targetingability of 68 Ga-NC-BCH in patients with gastrointestinal tumors.We focused on assessing whole-body physiologic CLDN18.2expression and tumor uptake and exploring its main relationship with patient prognosis before and after anti-CLDN18.2therapy using 68 Ga-NC-BCH.

Cell and Animal Models
The human gastric adenocarcinoma cell AGS was obtained at Peking University Cancer Hospital and Institute.The AGS CLDN18.2 cell line was generated by transfection with the full-length CLDN18.2.All animal experiments were performed in accordance with the guidelines of the Peking University Institutional Animal Care and Use Committee (approval number EAEC 2022-01).

Small-Animal PET/CT Protocol
Normal Kunming mice and AGS CLDN18.2 /AGS model nude mice were injected with 7.4 MBq of 68 Ga-NC-BCH via the tail vein (n 5 3).Then, 10-min static PET scans were acquired at each time point.Imaging was performed with a small-animal PET/CT scanner (Super Nova PET/CT; PINGSENG).

Patient Enrollment
The study was approved by the Medical Ethics Committee of the Peking University Cancer Hospital and registered at ClinicalTrials.gov(NCT02760225).All patients signed informed consent forms.All procedures performed in studies involving human participants complied with the ethical standards of institutional or national research councils and the 1964 Declaration of Helsinki and its subsequent amendments or similar ethical standards.
68 Ga-NC-BCH PET/CT Scanning All imaging was performed on a whole-body PET/CT uEXPLORER scanner (United Imaging).No specific preparation was required for patients.An intravenously administered dose of 68 Ga-NC-BCH (101.99 6 44.65 MBq; range, 56.61-177.97MBq) was used.Five patients underwent dynamic PET/CT imaging.A low-dose CT scan (120 kV, 50 mA, 5-mm slices) was first performed, and then the dynamic PET acquisition was started upon injection of the tracer.

Synthesis and Characterization of 68 Ga-NC-BCH
The molecular weight of the obtained primary single-domain antibody, CLDN18.2-targetingnanobody (ACN376), was defined as 16,075 atomic mass units.The molecular weight of ACN376-GGGGC was determined to be 15,380 atomic mass units (Supplemental Figs.1A, 1B, and 1D; supplemental materials are available at http://jnm.snmjournals.org).Maleimidomonoamide-NOTA was sitespecifically conjugated to ACN376-GGGGC via the maleimide-thiol reaction (Supplemental Fig. 2A).NOTA-ACN376 was obtained with an average NOTA-to-single-domain antibody ratio of approximately 1:1 (Supplemental Fig. 1C). 68Ga-NC-BCH was produced with a radiochemical yield of more than 95% and a radiochemical purity of more than 98%.The in vitro stability of 68 Ga-NC-BCH in 0.01 M phosphate-buffered saline and 5% human serum albumin was demonstrated by a radiochemical purity of more than 98% over 6 h (Supplemental Fig. 3).The quality control results are shown in Supplemental Table 1.

Small-Animal PET/CT Imaging and Immunohistochemistry Study
High uptake in the stomach was observed via static ex vivo imaging and a PET/CT image of Kunming mice spanning 1 h (Supplemental Fig. 5).
Small-animal PET/CT images of AGS CLDN18.2 tumor-bearing mice, AGS CLDN18.2 tumor-bearing mice pretreated with the antibody TST001 (1 mg) for 24 h, and AGS tumor-bearing mice were obtained at 30, 60, 120, and 240 min after injection of 68 Ga-NC-BCH (Fig. 1).The kidney SUV mean at 1 h after injection was 14.56 6 0.29 in the AGS CLDN18.2 group, 11.87 6 0.16 in the AGS CLDN18.2 blocking group, and 8.80 6 0.32 in the AGS group.The SUV mean of the stomach at 1 h after injection was 2.43 6 0.09 in the AGS CLDN18.2 group, 1.65 6 0.03 in the AGS CLDN18.2 blocking group, and 1.67 6 0.01 in the AGS group.The SUV mean in the AGS CLDN18.2 , AGS CLDN18.2 blocking, and AGS groups at 1 h after injection was 1.14 6 0.01, 0.54 6 0.01, and 0.42 6 0.03, respectively.The tumorto-muscle ratios at each time point after injection of 68 Ga-NC-BCH were significantly greater than those of the other control groups, and at 2 h after injection, the tumor-tomuscle ratio reached its maximum of 34.86 6 4.68 (Supplemental Fig. 6A).The results of immunohistochemistry revealed high and homogeneous CLDN18.2expression in AGS CLDN18.2 tumors, whereas AGS xenograft tumors were negative for CLDN18.2.The gastric mucosa of AGS CLDN18.2 and AGS tumor-bearing mice showed substantially positive expression of CLDN18.2 (Supplemental Fig. 6C).

Biodistribution, Pharmacokinetics, and Safety Study
The biodistribution of 68 Ga-NC-BCH in AGS CLDN18.2 and AGS tumor-bearing mice and the pharmacokinetics study in Kunming mice are presented in Supplemental Figure 6.At 2 h after injection, the stomachs of the mice in all 3 groups exhibited relatively high uptake (6.04 6 0.66 percentage injected dose [%ID]/g in the AGS CLDN18.2 group, 6.36 6 1.43 %ID/g in the AGS group, and 4.96 6 0.04 %ID/g in the blocking group).The uptake value of the kidney was higher than that of the stomach (194.86 6 5.56 %ID/g in the AGS CLDN18.2 group, 128.79 6 0.64 %ID/g  in the AGS group, and 129.70 6 8.55 %ID/g in the blocking group).
Tumor uptake in the AGS CLDN18.2 tumor-bearing mice was greater (6.61 6 0.41 %ID/g) than that in the AGS group (0.39 6 0.13 %ID/ g) and the blocking group (1.69 6 0.84 %ID/g).The pharmacokinetics study showed that 68 Ga-NC-BCH was cleared quickly from the blood, with a half-life of 22.77 min.After the injection of excess 68 Ga-NC-BCH (18.5 MBq, 925 MBq/kg), no obvious toxicity was observed in terms of body weight (Supplemental Fig. 7), blood biochemical parameters (Supplemental Table 2; Supplemental Fig. 8), or hematoxylin and eosin staining of main organ tissue slides (Supplemental Fig. 9).
68 Ga-NC-BCH Dosimetry and Biodistribution Between July 2022 and November 2022, 11 patients were enrolled: 10 with advanced gastric cancer and 1 with advanced colon cancer.Among all patients, the expression of CLDN18.2 was 40% or above.Patient characteristics are shown in Table 1.
No tracer-related adverse events were observed in any patients after injection of 68 Ga-NC-BCH.
All 11 patients underwent 68 Ga-NC-BCH PET/CT.The injection dose of 68 Ga-NC-BCH was 101.99 6 44.65 MBq (range, 56.61-177.97MBq).The highest organ dose values for 68 Ga-NC-BCH were estimated to be for the kidneys, gallbladder, stomach, spleen, and liver.The effective dose of 68 Ga-NC-BCH was estimated to be 0.042 6 0.02 mSv/MBq (Supplemental Table 3), which was lower than the radiation dose used for conventional 18 F-FDG PET/CT (7.0-14.0mSv) (12).Whole-body PET reduces the total radiation dose and facilitates translational research on this new type of radiopharmaceutical.
Five patients were subjected to dynamic scanning via total-body full-motion PET/CT scans.A 42-min whole-body static PET/CT scan was selected to determine the SUV mean in the major organs of these 5 participants.The dynamic curve showed that the 68 Ga-NC-BCH activity in the selected organs increased rapidly and gradually decreased to a steady state over time except in the stomach and kidneys (Fig. 2A).A static total-body PET/CT scan at 60 min was selected to determine the SUV mean of the main organs in all 5 participants (Fig. 2B).Except for the stomach wall, the SUV mean for 68 Ga-NC-BCH in normal tissues (including the spleen, pancreas, liver, brain, and ovaries) was extremely low, as is consistent with previous reports of immunohistochemistry staining of healthy human tissues (6).Since single-domain antibodies are excreted from the body through the kidneys, the probe exhibited strong renal retention in humans.The SUV mean of the positive lesions gradually increased over time from 0 to 58 min (Fig. 2C).The temporal radioactivity curve of the mediastinal metastatic lymph nodes and stomach wall showed a gradual upward trend, and radiotracer accumulation gradually increased over time (Supplemental Fig. 10B).
Surprisingly, all patients with preserved gastric walls had significantly greater uptake of imaging agents in the gastric mucosa (Supplemental Fig. 10A).To our knowledge, this was the first study in which whole-body PET has shown stomach accumulation of a radioactive tracer targeting CLDN18.2 in humans.
Tumor 68 Ga-NC-BCH Uptake and Correlation In total, 215 CLDN18.2-positivelesions were detected in 9 (9/11, 81.8%) patients by 68 Ga-NC-BCH PET/CT, including 4 patients who had not received CLDN18.2-targetedtherapy before imaging and 5 patients who had.The mean SUV max of positive lesions detected by 68 Ga-NC-BCH PET/CT did not significantly differ among different locations (Fig. 3A).Positive lesions were detected in all patients who had received CLDN18.2targetedtherapy before imaging; these lesions were distributed in the lymph nodes, peritoneum, abdominal muscle, and subcutaneous tissue.The other 4 patients, who had not received CLDN18.2targetedtherapy, had positive lesions distributed among the lymph nodes, peritoneum, liver, bone, and ovaries.The mean SUV max of positive lesions in patients without prior CLDN18.2-targetedtherapy was significantly greater than that in treated patients (Fig. 4B).
The CLDN18.2 expression intensity was 11 or 111 in all patients enrolled in the study. 68Ga-NC-BCH-positive lesions    were detected in 9 patients, including 4 patients with 11 CLDN18.2expression and 5 patients with 111 expression.The positive lesions were distributed in the liver, lymph nodes, bone, peritoneum, pleura, abdominal muscle, and ovaries (Table 2).
68 Ga-NC-BCH uptake was highest in liver metastases, followed by lymph nodes and bone.There were statistically significant differences in SUV max between patients with 11 and 111 CLDN18.2 expression in positive lesions (Fig. 4A).Moreover, there was no significant difference in the SUV max of the lesions that were positive on 18 F-FDG PET/CT regardless of the expression intensity of CLDN18.2 or whether the patients had received CLDN18.2-targetedtreatment.Moreover, CLDN18.2 expression differed among metastatic lesions from the same patient (Fig. 5).
68 Ga-NC-BCH PET/CT and 18 F-FDG PET/CT In total, 225 lesions were positive on 68 Ga-NC-BCH PET/CT, distributed in the liver, lymph nodes, bone, peritoneum, pleura, abdominal muscle, and ovaries.In total, 209 lesions were positive on 18 F-FDG PET/CT, distributed in the liver, lymph nodes, bone, peritoneum, abdominal muscle, and ovaries.There was a significant difference in the tumor-tonontumor (T/NT) ratio between all positive lesions detected by the 2 methods.According to a subgroup analysis of the T/NT ratio comparing metastatic lesions at different sites, the T/NT ratio of the lymph nodes and peritoneal metastases detected by 68 Ga-NC-BCH PET/CT was significantly greater than that detected by 18 F-FDG PET/CT (Figs. 3B and 3C). 68Ga-NC-BCH PET/CT was also able to detect small peritoneal and pleural metastases well.Patient 9, a woman with advanced gastric cancer, had a CLDN18.2expression level of 40%, 21.The left pleural metastatic nodule and thickened peritoneum did not show obvious uptake on 18 F-FDG PET/CT, whereas 68 Ga-NC-BCH PET/CT showed high uptake (SUV max , 3.1 and 5.7). 68Ga-NC-BCH PET/CT was also effective at detecting the expression of CLDN18.2 in colon cancer metastases.Patient 4, a man with colon cancer, had a CLDN18.2expression level of 40%, 31.On 18 F-FDG PET/CT, abdominal metastatic lymph nodes showed mild uptake, whereas the same lymph nodes on 68 Ga-NC-BCH PET/CT showed high uptake (SUV max , 1.4 vs. 5.6) (Figs.6A and 6B).Typical cases are shown in Supplemental Figure 11.
The log-rank method was used to test the difference in survival time distribution between the 2 groups.There was no significant difference in survival time distribution between patients with different CLDN18.2expression levels (P 5 0.2988) and no significant difference in clinical stage between the 2 groups (P 5 0.2599).Moreover, there was no significant difference in survival time between the 2 groups (P 5 0.1865, SUV max $ 2.5 vs. SUV max , 2.5; Fig. 7).

DISCUSSION
Several forms of CLDN18.2-targetingtherapeutic agents, including monoclonal antibodies, antibody-drug conjugates, and chimeric antigen receptor T (CAR-T) cell therapies, are currently undergoing clinical trials worldwide (13,14).Our group previously published interim results of a phase I clinical trial of CLDN18.2-specificCAR-T cells, which showed that the overall response rate and disease control rate of CAR-T cells reached 48.6% and 73.0%, respectively (14).Although CLDN18.2-targetingtherapy has achieved good results in clinical studies, we have also noted that the expression level of CLDN18.2 affects therapeutic efficacy to a certain extent.
In this study, we selected a single-domain antibody as a precursor of the radiotracer and labeled it with the short-half-life nuclide 68 Ga based on its small molecular weight and short cycle time in vivo.Compared with those of traditional antibodies, the rapid tissue penetration and renal clearance rate of single-domain antibodies enable high image contrast to be obtained within 1 h after probe injection, allowing patients to complete the whole imaging workflow within 1 d, greatly increasing compliance and reducing radiation exposure.A preclinical study indicated that 68 Ga-NC-BCH has good affinity for CLDN18.2 and can specifically bind to CLDN18.2-positivecells.68 Ga-NC-BCH was taken up strongly by the gastric mucosa in a mouse model, indicating that the smaller molecular architecture allows it to reach the insidious and dense CLDN18.2epitope on the gastric mucosa.However, since CLDN18.2 is typically buried in the gastric mucosa, neither of the previous monoclonal antibody-based probes-89 Zr-DFO-TST001 and 124 I-18B10(10L)-was available in normal tissue (9,15,16).In addition to gastric organ and positive tumors, probes have a high nonspecific radioactive accumulation in the kidneys, as previously reported (17).The prominent renal uptake is due to excretion of single-domain antibodies through the kidneyurinary system.In addition, nonspecific tubular reuptake after glomerular blood flow is another important factor contributing to elevated renal uptake.Indeed, 68 Ga-NC-BCH exhibited a longer renal retention time (207.666 19.99 %ID/g, 2 h after injection) than did other single-domain antibody tracers described in previous studies (17)(18)(19).The increase in nonspecific radioactive accumulation in the kidneys can lead to high radiation doses in patients and may seriously impede the diagnosis of small perirenal lesions.
Then, we described the results of the first, to our knowledge, 68 Ga-NC-BCH PET/CT study on patients.According to dosimetry studies, the effective radiation dose of 68 Ga-NC-BCH was much lower than that of 124 I-18B10(10L).All results showed that 68 Ga-NC-BCH PET/CT is a safe, noninvasive imaging method for detecting CLDN18.2 in patients receiving CLDN18.2-targetedtherapy.
All 11 patients in the study underwent both 68 Ga-NC-BCH PET/CT and 18 F-FDG PET/CT within 1 wk.There was a significant difference in the T/NT ratio between all lesions positive on the 2 methods.Surprisingly, in the subgroup analysis, the T/NT ratio of the lymph nodes and peritoneal metastases detected by 68 Ga-NC-BCH PET/CT was significantly greater than that detected by 18 F-FDG PET/CT.Lymph nodes and the peritoneum are the most common metastatic sites of advanced gastric cancer.The T/NT ratio of lesions detected by 68 Ga-NC-BCH PET/CT was significantly greater, and this high ratio is more conducive to lesion detection. 68Ga-NC-BCH PET/CT reflects the expression level of CLDN18.2 in tumor lesions because the tracer is a CLDN18.2-targetingsingle-domain antibody.
The study included patients who did or did not receive CLDN18.2-targeted therapy before 68 Ga-NC-BCH PET/CT.Uptake in lesions receiving CLDN18.2-targetedtherapy was higher than that in lesions not receiving CLDN18.2-targetedtherapy (P 5 0.2695).This could be because after the first CLDN18.2-targetedtherapy, the lesions progressed approximately 1 y later; these advanced lesions still highly expressed CLDN18.2, which also provided a basis for the second CAR-T treatment.More significantly, results showed that metastatic lesions with a higher CLDN18.2expression level had a higher SUV max .This provides some basis for noninvasive detection of CLDN18.2 expression levels in the future.
There were several limitations to the study.First, the small sample size of patients may impede the comprehensive performance of the radiotracer.Also, the small sample challenges us to accurately assess 68 Ga-NC-BCH uptake in primary gastric tumors because of interference with physiologic uptake in the gastric mucosa and therapeutic response.

CONCLUSION
We developed a CLDN18.2-specificsingle-domain antibody nuclide probe that enables same-day PET imaging, and we demonstrated this ability using whole-body PET.Uptake of 68 Ga-NC-BCH correlated significantly with the expression level of CLDN18.

FIGURE 2 .
FIGURE 2. (A) Dynamic changes in SUV mean of selected organs at 0-42 min (n 5 5).(B) Rank ordering of 68 Ga-NC-BCH uptake in different organs indicated by SUV mean at 42 min (n 5 5).(C) Pilot translational study on dynamic total-body PET/CT imaging of 68 Ga-NC-BCH illustrating time distribution of radiotracers within tumors to optimize imaging window.

FIGURE 5 .
FIGURE 5. (A) Muscle metastases in right lower abdominal wall with mild elevated uptake.(B-D) Multiple lymph node metastasis showing CLDN18.2expression degree from low (B) to high (D) (left) and imaging of 68 Ga-NC-BCH of patient (right).
-BCH PET/CT has great potential in the selection of CLDN18.2-targetedtherapy strategies and the monitoring of treatment responses by systematically quantifying the systemic expression of CLDN18.2.DISCLOSURE This research was funded by the National Key R&D Program of China (2022YFA0912400), the National Natural Science

FIGURE 7 .
FIGURE 7. Graphs of survival time.(Top) Green depicts group aboveand dark yellow, group below-mean CLDN18.2expression of 40%.(Middle) Green depicts clinical stage of 4, and dark yellow depicts clinical stage , 4. (Bottom) Green depicts group below-and dark yellow, group above-mean SUV max of 2.5.HR 5 hazard ratio.

TABLE 1
Patient Characteristics

TABLE 2
Foundation of China (82272627, 82171973, and 82171980), Capital's Funds for Health Improvement and Research (2022-2Z-2154 and 2022-2Z-2155), and Science Foundation of Peking University Cancer Hospital (2022-14).Intellectual properties protection has been filed by Chengdu AlpVHHs Co. Ltd.No other potential conflict of interest relevant to this article was reported.How can CLDN18.2expression be detected rapidly and accurately in gastrointestinal cancer patients?PERTINENT FINDINGS: We demonstrated the feasibility of 68 Ga-NC-BCH for quantifying CLDN18.2levels in preclinical models and in patients, in whom a high affinity for CLDN18.2 was detected.IMPLICATIONS FOR PATIENT CARE: 68 Ga-NC-BCH can be used to image CLDN18.2-positivetumors and identify patients with high CLDN18.2expression as potential candidates for targeted therapy through noninvasive PET imaging.