Development of a Theranostic Convergence Bioradiopharmaceutical for Immuno-PET Based Radioimmunotherapy of L1CAM in Cholangiocarcinoma Model

Cholangiocarcinoma (cholangiocarcinoma) are malignant tumors originating from the bile duct epithelium and can be classified anatomically into intrahepatic and extrahepatic cholangiocarcinoma (1). To date, diagnostic modalities and absolute criteria have not been established for cholangiocarcinoma. Carbohydrate antigen (CA) 19-9 and CA-125 are used as serum biomarkers for cholangiocarcinoma, but they are nonspecific and can be increased in other disease such as cholangiopathies (2). Current diagnostic modality of cholangiocarcinoma based on endoscopic retrograde cholangiography, percutaneous transhepatic cholangiography, ultrasonography, MRI, and CT scanning are limited as they are invasive or nonspecific diagnostic tools with a varying degree of accuracy (3). Complete surgical resection is currently the only way to treat and cure cholangiocarcinoma. The overall 5-year survival rate, which includes patients having undergone surgical treatment, is less than 5% (4). This poor survival outcome is related to being diagnosed at the advanced stage, and it is also associated with the lack of specific clinical symptoms and screening method and the poor response to conventional chemotherapy and radiation treatment. Thus, a novel diagnostic method, which combines in vivo imaging and biological characterization of cholangiocarcinoma, is highly desired to improve the sensitivity and specificity of cholangiocarcinoma detection and predict the outcome for patients with cholangiocarcinoma. In addition, it is imperative to develop a new therapeutic regimen that overcomes conventional therapy.

L1 cell adhesion molecule (L1CAM) is a 200- to 220-kDa transmembrane glycoprotein of the immunoglobulin (Ig) superfamily (5). L1CAM is highly overexpressed in numerous tumors such as neuroblastomas, melanoma, breast cancer, and ovarian cancer (6, 7), and its expression is also associated with several features of cancer such as tumor growth, tumor cell invasion, metastasis as well as chemoresistance (8, 9). L1CAM is highly expressed in cholangiocarcinoma and involved in tumor progression, migration, and resistance to apoptosis in cholangiocarcinoma (10–12). Previous reports have shown that L1CAM could be a useful diagnostic molecular marker with the potential to serve as a therapeutic target for cholangiocarcinoma.

Chimeric A10-A3 (cA10-A3) is a chimeric IgG1 mAb that binds to L1CAM with high affinity (K_D = 1.9 nmol/L; ref. 13) and is selectively binding to L1CAM-positive cells (10, 11). Structural studies have suggested that homophilic interaction of L1CAM Ig-domain forms a static cell adhesion along the intercellular boundaries (14) and heterophilic interaction of L1CAM with other cell surface protein, such as growth factor receptor and integrins, could drive various functions (9). Previous studies showed cA10-A3 blocks L1CAM homophilic binding and inhibits tumor growth of cholangiocarcinoma in vivo (13, 15).

Molecular imaging of cancer could be a promising tool for detecting diseases in early stages and choosing disease- and patient-specific treatment (16). Especially, immuno-PET using tumor-targeting radiolabeled-molecules has high resolution, sensitivity, and features enabling whole-body, quantitative, and noninvasive detection of target antigen. Several studies reported that immuno-PET could provide an improved diagnostic imaging for assessing target expression and antibody accumulation in tumor lesion, screening patient, and guiding mAb-based therapy (17). Currently, ⁶⁴Cu and ⁸⁹Zr are mainly used for immuno-PET imaging. ⁸⁹Zr has attracted a great attention for the radiolabeling of slowly accumulating targeting vector such as antibody, because of long half-life (78.4 hours), but ⁸⁹Zr-labeled antibody showed higher radiation exposure than ⁶⁴Cu-labeled antibody (18). Immuno-PET with ⁶⁴Cu-labeled antibody could potentially achieve adequate tumor-to-tissue contrast and low radiation exposure with relatively short half-life (19, 20). On the basis of immuno-PET imaging, the therapeutic strategy could be tailored for individual patients that are likely to be responded against therapy.

Radioimmunotherapy is a targeted therapy approach that has the potential to improve the efficacy of mAb-based therapies through delivering radiation to the target cells by linking the radionuclide with a mAb (21). Several mAbs against L1CAM have been developed and shown to inhibit cell proliferation in vitro and reduce tumor mass in vivo (13, 22, 23). Furthermore, anti-L1CAM antibodies labeled with radioisotope such as ⁶⁴Cu, ⁶⁷Cu, ¹⁶¹Tb, or ¹⁷⁷Lu, have shown therapeutic potential against L1CAM-expressing tumor (24–26). Especially, ¹⁷⁷Lu is ideal for use in therapeutic purposes because of low tissue penetration range (∼2 mm) with low-energy β⁻-emission (497 keV) that is favored in treatment of small tumors while limiting irradiation of normal tissue. Because of its proper characteristics, ¹⁷⁷Lu has been increasingly used in preclinical and clinical studies (27, 28).

We previously reported the development of theranostic convergence bioradiopharmaceutical, ⁶⁴Cu-/¹⁷⁷Lu-labeled cetuximab (19, 20). Immuno-PET visualized the tumor and quantified the EGFR expression in EGFR-expressing esophageal squamous cell carcinoma and head and neck squamous cell carcinoma xenografted model. Radioimmunotherapy exhibited marked inhibition of tumor growth.

We now report the development of a novel targeted theranostic convergence bioradiopharmaceutical, based on the targeting L1CAM to cholangiocarcinoma. This radioimmunoconjugate can be radiolabeled with ⁶⁴Cu and ¹⁷⁷Lu via 2-S-(4-isothiocyanatoenzyl)-1,4,7-triazacyclononane-1,4,7-triacetic acid (p-SCN-Bn-NOTA). In this study, we have assessed the feasibility and potential role of theranostic convergence bioradiopharmaceutical, ⁶⁴Cu-/¹⁷⁷Lu-cA10-A3, using immuno-PET imaging-based radioimmunotherapy in cholangiocarcinoma xenografted model.

Choi-CK (adenomatous) and SCK (sarcomatoid) cell lines were established from tumor sample of Korean patients with intrahepatic cholangiocarcinoma (29). SCK-L1 cells were prepared by stable transfection of SCK cells with a lentiviral vector (Macrogen) containing L1CAM cDNA (11). Choi-CK, SCK, and SCK-L1 cells were grown in DMEM. ACHN (human renal cell adenocarcinoma) cell line was purchased from ATCC and maintained in Eagle's Minimum Essential Medium. TFK-1 (extrahepatic cholangiocarcinoma) cell line was obtained from RIKEN Bioresource Center Cell Bank and grown in RPMI1640 medium. JCRB1033 (human gallbladder carcinoma) cell line was purchased from JCRB cell bank and maintained in Williams' E medium (Invitrogen). All cell lines were not authenticated and used for experiments between passage 12 and 30. All media were supplemented with 10% FBS and 1% antibiotics/antimycotics (Gibco). All cell lines were incubated at 37°C in 5% CO₂.

Cell lysates were subjected to SDS-PAGE and transferred to polyvinylidene fluoride membranes using iBlot2 Dry Blotting System (Life Technologies). The membranes were incubated primary antibodies (L1CAM, Novus Biologicals; β-actin, Cell Signaling Technology), followed by horseradish peroxidase–conjugated secondary antibodies (Santa Cruz Biotechnology). The immunoreactive bands were visualized using a chemiluminescent substrate (Thermo Fisher Scientific).

Cells were incubated with chimeric anti-L1CAM IgG1 antibody, cA10-A3 (13) or isotype control (Rituximab; Roche) for 1 hour at 4°C. After repeated washing with PBS containing 1% bovine serum albumin (BSA, Sigma-Aldrich), the cells were incubated with FITC–conjugated anti-human IgG (Sigma-Aldrich) for 1 hour at 4°C. Stained cells were analyzed using FACS Calibur and CellQuest software (BD Immunocytometry System).

Immunoconjugates were prepared by a previous protocol (19). cA10-A3 or rituximab were buffer-exchanged and concentrated to 10 mg/mL in 0.1 mol/L sodium bicarbonate buffer, pH 8.5 using Vivaspin-20 ultracentrifugation tubes (MW cutoff; 50 kDa, Sartorius). A 10-fold molar excess of 2-S-(4-isothiocyanatobenzyl)-1,4,7-triazacyclononane-1,4,7-triacetic acid (p-SCN-Bn-NOTA, Macrocyclics) in DMSO was added to antibodies and conjugation was allowed to proceed at room temperature for 2 hours and continued at 4°C overnight. Unconjugated chelator was removed by dialysis. The immunoconjugates were finally concentrated to 2 mg/mL in 20 mmol/L sodium acetate buffer, pH 6.5. To determine the number of chelate per antibody, MALDI mass spectrometry was performed.

⁶⁴Cu was produced by 50 MeV cyclotron irradiation at KIRAMS. ¹⁷⁷Lu was purchased from ITM AG (Germany). ⁶⁴CuCl₂ (74 MBq) or ¹⁷⁷LuCl₃ (74-740 MBq) were added to NOTA-conjugated antibodies (1 mg) and the reaction mixtures were incubated for 1 h at room temperature (⁶⁴Cu) or 37°C (¹⁷⁷Lu), respectively, with constant shaking. Radiolabeling yield was determined by instant thin-layer chromatography-silica gel (ITLC-sg) analysis. Radiochemical purity was analyzed using size exclusion HPLC.

To evaluate L1CAM expression level of cancer cell lines using radiolabeled immunoconjugates, 1 × 10⁶ cells in cold PBS containing 1% BSA were incubated with ⁶⁴Cu- or ¹⁷⁷Lu-NOTA-cA10-A3 (100 ng) for 1 hour at 4°C. Cells were triple washed with cold PBS containing 1% BSA. To determine nonspecific binding, 100-fold excess cold cA10-A3 antibody was treated 1 hour before ⁶⁴Cu- or ¹⁷⁷Lu-NOTA-cA10-A3 were added. The radioactivity in each sample was measured in a gamma counter (WIZARD 1480, Perkin-Elmer). The cell bound radioactivity (%) was calculated by (total cell bound radioactivity – nonspecific binding radioactivity)/total applied radioactivity × 100.

All animals studied were approved by KIRAMS Institutional Animal Care and Use Committee (IACUC). BALB/c athymic female mice, aged 6 weeks (NARA Biotech) were used. Mice were subcutaneously injected in the right flank with 1 × 10⁷ Choi-CK, SCK, SCK-L1 and JCRB1033 cells in PBS. TFK-1 model was excluded because of very low tumorigenic potential in BALB/c athymic mice.

Tumor-bearing mice (n = 3 or 4) were intravenously injected with ⁶⁴Cu-NOTA-cA10-A3 (3.7 MBq/100 μg) or ¹⁷⁷Lu-NOTA-cA10-A3 (3.7 MBq/100 μg). In biodistribution study of ⁶⁴Cu-NOTA-cA10-A3, the mice were sacrificed at 2, 24, 48, and 72 hours postinjection in SCK-L1 xenografted model and at only 48 h in other xenografted models. In case of ¹⁷⁷Lu-NOTA-cA10-A3, the biodistribution performed at 2 hours, 1, 3, 5, 7, and 14 days postinjection in SCK-L1 xenografted model and at only 5 days in other xenografted models. The blood, various organ and tissues were harvested, weighed and measured in a gamma counter. The accumulated activity represented as the percentage of the injected radioactivity dose per a gram of tissue (%ID/g). Tumor-to-blood (T/B), tumor-to-muscle (T/M), and tumor-to-liver (T/L) ratios were calculated.

Tumor-bearing mice were intravenously administrated with ⁶⁴Cu-NOTA-cA10-A3 (3.7 MBq/100 μg). The mice were imaged for 1 hour at 48 hours postinjection using a small animal PET (microPET R4, Concorde Microsystems). For the blocking experiment, 100-fold excess cold cA10-A3 was intravenously administrated 30 minutes before injecting ⁶⁴Cu-NOTA-cA10-A3. Quantitative data were expressed as standardized uptake value (SUV), which are defined as tissue concentration (MBq/mL)/injected dose (MBq) divided by the body weight (g; ref. 30). After PET imaging, SCK-L1 tumor bearing mice were immediately euthanized, frozen and the digital whole-body autoradiography (DWBA) was performed.

Micro-SPECT/CT imaging was performed for 14 days after the intravenous injection of ¹⁷⁷Lu-NOTA-cA10-A3 (12.95 MBq/100 μg) using an Inveon SPECT/CT system (Siemens Medical solutions) equipped with two NaI detectors and fitted with 1.0-mm mouse whole body multipinhole collimators with 5 pinholes. A total of 30 projections were acquired in a 256 × 256 acquisition matrix with a 180 seconds per projection and 0.5 revolution. Images were reconstructed using a MAP 3D (16 iterations). Before micro-SPECT imaging, cone-beam CT images were acquired using 180 projections, 200 ms/projection, 80 kVp, 500 μA. Coregistration of micro-SPECT and CT images was performed using Inveon Research Workplace software. After SPECT/CT scan, DWBA was performed.

SCK-L1 xenografted models with tumor size of approximately 200 mm³ were divided into four groups (n = 9 or 10/group). Each group was treated with a single intravenous injection of saline, cA10-A3 (5 mg/kg), ¹⁷⁷Lu-NOTA-isotype or ¹⁷⁷Lu-NOTA-cA10-A3 (12.95 MBq, 5 mg/kg). Tumor volume and body weight were monitored three times a week. Tumor volume was calculated using the following formula: tumor volume = long diameter × (short diameter)²/2.

To evaluate glucose metabolic changes by treatment, ¹⁸F-FDG-PET imaging experiments (n = 3/group) were performed at 16 day posttreatment. ¹⁸F-FDG (7.4 MBq) was intravenously injected 1 hour before scanning and static scans were obtained for 20 minutes. PET images were quantified and analyzed as described previously (30).

Tumors were harvested at 9 days posttreatment and fixed in 4% paraformaldehyde solution. Subsequently, apoptotic cells in tissue section were detected by TUNEL, using an In Situ Apoptosis Detection Kit (S7100-KIT, Millipore) on 4-μm thick sections according to the manufacturer's instructions. Tumor sections were stained with Ki-67–specific SP6 rabbit mAb (Abcam) and EnVision+ detection system (Dako) was applied according to the instructions. Nuclear staining of Ki-67 was considered positive. In random fields, TUNEL-positive nuclei and Ki-67 staining index (%) were defined as the percentage of positive nuclei within about 1,000 number of nuclei.

Quantitative data are represented as the mean ± SD. Statistical analysis was performed by one-way ANOVA or Student t test using Prism 5 (GraphPad). P values were considered significant if <0.05.

The relative expression level of L1CAM was determined by Western blot analysis (Fig. 1A) and flow cytometry (Fig. 1B). SCK-L1 expressed relatively high level of L1CAM (MFI 45.3). TFK-1 (MFI 29.2) and JCRB1033 (MFI 19.8) were moderate. SCK (MFI 10.4) and Choi-CK (MFI 12.1) were low. L1CAM expression in ACHN cell line was not detected in these experiments. The expression pattern of L1CAM showed good correlation between western blot and flow cytometry results.

Theranostic convergence bioradiopharmaceutical used in this study was based upon NOTA chelator to provide versatility for incorporating different radionuclides, ⁶⁴Cu and ¹⁷⁷ Lu. As shown in Fig. 1C, NOTA was conjugated with antibodies and radiolabeling of the immunoconjugates was identically performed with ⁶⁴Cu and ¹⁷⁷Lu at different reaction temperature (room temperature vs. 37°C). These data indicate that our theranostic convergence biopharmaceutical is an attractive platform for delivery of radionuclides for PET (⁶⁴Cu) or radioimmunotherapy (¹⁷⁷Lu).

The average number of chelates per cA10-A3 and rituximab as isotype control was determined to 4.1 ± 0.1 and 5.1 ± 0.004, respectively, using MALDI mass spectroscopy (Supplementary Fig. S1). ⁶⁴Cu- or ¹⁷⁷Lu-NOTA-cA10-A3 was prepared successfully at high radiolabeling yield and radiochemical purity (>98%) evaluated by ITLC-sg and size exclusion HPLC analysis (Supplementary Fig. S2). ¹⁷⁷Lu-NOTA-rituximab also prepared with high radiolabeling yield and radiochemical purity. The immunoreactivities of ⁶⁴Cu- and ¹⁷⁷Lu-NOTA-cA10-A3 were measured to be 0.94 and 0.93, respectively (Supplementary Fig. S3A and S3B). There was minimal reduction of binding ability of antibody during conjugation of NOTA and radiolabeling procedure. The radioimmunoconjugates showed favorable serum stability; >97% at 24 hours for ⁶⁴Cu-NOTA-cA10-A3 and >84% at 7 days for ¹⁷⁷Lu-NOTA-cA10-A3, respectively (Supplementary Fig. S3C and S3D).

To identify whether radiolabeled immunoconjugates evaluate the expression of L1CAM, cell-binding assay was performed using ⁶⁴Cu- or ¹⁷⁷Lu-NOTA-cA10-A3. Cell-bound radioactivity (%) of ⁶⁴Cu-NOTA-cA10-A3 was the highest in SCK-L1 cells (45.2% ± 2.2%), moderate in TFK-1 (4.8% ± 0.1%) and JCRB1033 (3.7% ± 0.5%) cells, relatively low in SCK (1.0% ± 0.3%) and Choi-CK (1.1% ± 0.2%), and very low in ACHN (0.4% ± 0.3%; Fig. 1D). The cell-bound radioactivity (%) of ¹⁷⁷Lu-NOTA-cA10-A3 showed similar pattern (Fig. 1E). There was well correlated the cell bound radioactivity (%) of between ⁶⁴Cu- or ¹⁷⁷Lu-NOTA-cA10-A3 and the L1CAM expression by Western blot or flow cytometry analysis.

Biodistribution data of ⁶⁴Cu-NOTA-cA10-A3 are presented in Fig. 2A. The radioactivity of ⁶⁴Cu-NOTA-cA10-A3 in blood was highest at 2 hours with 28.2 ± 2.4 %ID/g and gradually decreased to 7.7 ± 0.9 %ID/g at 72 hours postinjection. The radioactivities of other organ and tissues excluding tumors were decreased in a time-dependent manner. SCK-L1 tumor uptake of ⁶⁴Cu-NOTA-cA10-A3 was peaked at 48 hours, reaching 18.9 ± 2.6 %ID/g. SCK-L1 tumor to blood, tumor to muscle, and tumor to liver ratios at 48 hours were 2.3 ± 0.7, 14.3 ± 2.1, and 4.0 ± 0.3, respectively. The uptake of Choi-CK, SCK, and JCRB1033 tumors at 48 hours were 8.7 ± 1.1, 12.4 ± 1.2, and 13.1 ± 0.8 %ID/g, respectively. The uptake patterns of ⁶⁴Cu-NOTA-cA10-A3 in cholangiocarcinoma tumors were consistent with in vitro cell binding assay data.

Figure 2B showed the biodistribution data of ¹⁷⁷Lu-NOTA-cA10-A3. The radioactivity in blood was 26.0 ± 3.3 %ID/g at 2 hours, followed by relatively fast clearance at 14 days with 0.3 ± 0.2 %ID/g. The radioactivities of other organ and tissues excluding tumors and femur were decreased in a time-dependent manner. Femur uptake increased from 2.6 ± 0.3 %ID/g at 2 hours to 5.0 ± 1.0 %ID/g at 7 day and maintained with 4.3 ± 0.8 at day 14. The SCK-L1 tumor uptake of ¹⁷⁷Lu-cA10-A3 was peaked at 5 day with 44.6 ± 3.2 %ID/g. The uptake of Choi-CK, SCK, and JRCB1033 tumors at 5 day were 7.7 ± 0.5, 18.1 ± 4.2, and 22.2 ± 9.2 %ID/g, respectively. SCK-L1 tumor to blood, SCK-L1 tumor to muscle, SCK-L1 tumor to liver ratios at day 5 were 7.6 ± 2.5, 43.0 ± 14.1 and 7.7 ± 0.8, respectively (Supplementary Table S1). Biodistribution data showed similar pattern with in vitro cell binding assay result, in which the radioactivity of cholangiocarcinoma tumors was proportional to the L1CAM expression level.

The blood pharmacokinetic data were analyzed by biexponential curve fitting based on the biodistribution data of ¹⁷⁷Lu-cA10-A3 (Supplementary Fig. S4). The blood half-life distribution and elimination phase was 2.6 ± 1.8 hours and 65.5 ± 12.4 hours, respectively. Blood mean residence time was 69.0 ± 1.4 hours. The radiation dosimetry of ⁶⁴Cu/¹⁷⁷Lu-cA10-A3 is summarized in Supplementary Tables S2 and S3. The critical organ with the highest radiation dose was spleen (1.5 ± 0.5 cGy/MBq), followed by lung and liver. The radiation dose to bone marrow was very low with 5.7 × 10⁻³ ± 1.6 × 10⁻³ cGy/MBq of ¹⁷⁷Lu administered.

To evaluate the potential of ⁶⁴Cu-NOTA-cA10-A3 as an immuno-PET imaging agent for L1CAM expression level in various xenografted models, we performed immuno-PET imaging at 48 hours postinjection (Fig. 3). There was minimal uptake (SUV 1.7 ± 0.1) in ACHN renal cell carcinoma model (Fig. 3A). The tumor uptake (SUV) order of ⁶⁴Cu-NOTA-cA10-A3 in PET images was SCK-L1 (6.2 ± 0.0) ≫ JCRB1033(2.7 ± 0.2) > SCK (2.6 ± 0.1) = Choi-CK (2.6 ± 0.3) > ACHN. In the blocking experiment, the SUV of SCK-L1 tumor was significantly reduced to 2.0 ± 0.1 by pretreatment of excess cold cA10-A3 antibody. DWBA images showed similar distribution pattern with immuno-PET images. Immuno-PET images were consistent with biodistribution data of ⁶⁴Cu-NOTA-cA10-A3. These results suggest that ⁶⁴Cu labeled cA10-A3 as an immuno-PET imaging agent has the potential role for noninvasive and quantitative in vivo imaging of L1CAM expression in cholangiocarcinoma xenografted models.

Considering the characteristic of ¹⁷⁷Lu emitting both γ and β particles, micro-SPECT/CT imaging was performed to evaluate the feasibility of SPECT imaging by gamma radiation of ¹⁷⁷Lu during ¹⁷⁷Lu-NOTA-cA10-A3 radioimmunotherapy. Representative SPECT/CT images of ¹⁷⁷Lu-NOTA-cA10-A3 in SCK-L1 xenografted model for 14 days were shown in Fig. 4 and Supplementary Fig. S5. ¹⁷⁷Lu-NOTA-cA10-A3 specifically accumulated in SCK-L1 tumor and was relatively low uptake in the liver and bone. The radioactivity of SCK-L1 tumor was retained for 14 days and SCK-L1 tumor volume was markedly decreased at 14 days by administration of therapeutic dose (12.95 MBq) of ¹⁷⁷Lu-NOTA-cA10-A3. DWBA image represented similar distribution pattern with SPECT/CT image (Fig. 4D).

To evaluate the therapeutic efficacy of anti-L1CAM targeted radioimmunotherapy using ¹⁷⁷Lu, SCK-L1 xenografted mice were injected with saline as control, cA10-A3 as immunotherapy, ¹⁷⁷Lu-isotype as radioimmunotherapy control, and ¹⁷⁷Lu-cA10-A3 as L1CAM-targeted radioimmunotherapy. There was no therapeutic efficacy for inhibiting the growth of SCK-L1 tumor by immunotherapy with cA10-A3. ¹⁷⁷Lu-cA10-A3 treated mice showed significantly reduced tumor volume on day 20, compared to the other treatment groups (vs. saline: P < 0.01; vs. cA10-A3: P < 0.01; vs. ¹⁷⁷Lu-isotype; P < 0.01, Fig. 5A). In ¹⁷⁷Lu-cA10-A3 radioimmunotherapy, SCK-L1 tumor volume was decreased after 10 day post-treatment and SCK-L1 tumor volume at 27 day showed 51.4% reduction compared with that before treatment. In ¹⁷⁷Lu-isotype-treated group, SCK-L1 tumor growth was slightly retarded by non-specific targeting, however, tumor growth still increased by time dependent manner. There was no statistical difference between saline or cA10-A3 and ¹⁷⁷Lu-isotype (P > 0.05; ¹⁷⁷Lu-isotype vs. saline or cA10-A3). ¹⁷⁷Lu-cA10-A3 radioimmunotherapy showed greater therapeutic efficacy than ¹⁷⁷Lu-isotype from 20 day to 27 day (P < 0.01). SCK-L1 tumor models were well tolerable to the ¹⁷⁷Lu-cA10-A3 treatment, and no apparent body weight loss was observed (Fig. 5B).

To assess the change of glucose metabolism by treatment, ¹⁸F-FDG-PET imaging was performed at 16 days after treatment (Fig. 6A). There was no difference of tumor ¹⁸F-FDG uptake between saline- (1.03 ± 0.22) and cA10-A3- (1.01 ± 0.20) treated groups. The tumor uptake of ¹⁸F-FDG slightly increased in ¹⁷⁷Lu-isotype-treated group (1.28 ± 0.15), but there was no statistically significant difference (¹⁷⁷Lu-isotype vs. saline- or cA10-A3). ¹⁸F-FDG SUV of ¹⁷⁷Lu-cA10-A3-treated group (0.56 ± 0.10) showed a statistically marked reduction compared to those of other groups (P < 0.01, Fig. 6B). Tumor volume in FDG-PET image of ¹⁷⁷Lu-cA10-A3-treated group was also reduced compared to those of other groups. These results suggested that quantitative ¹⁸F-FDG-PET could be useful for monitoring therapeutic efficacy by radioimmunotherapy.

Tumor tissues were obtained at 9 days posttreatment and histological analysis in the tumors was performed (Fig. 6C). ¹⁷⁷Lu-cA10-A3 treatment led to a significant increase in apoptosis as quantified by TUNEL-positive nuclei (33.0 ± 12.2), compared with saline- (13.9 ± 5.1), cA10-A3- (19.6 ± 4.0), and ¹⁷⁷Lu-isotype- (19.5 ± 7.4) treated groups (P < 0.01; Fig. 6D). The percentage of Ki-67 staining index in ¹⁷⁷Lu-cA10-A3 treatment (42.5 ± 12.4%, P < 0.01) was markedly decreased compared to that in tumors treated with saline (64.5 ± 6.1%), cA10-A3 (69.0 ± 4.1%), and ¹⁷⁷Lu-isotype (69.4 ± 4.5%; Fig. 6E).

Human absorbed doses to normal organs were estimated based on the biodistribution of ⁶⁴Cu-NOTA-cA10-A3 (n = 3); organs with high absorbed doses were the lung (1.1 ± 0.2 × 10⁻² mSv/MBq), lower large intestine (6.5 ± 0.6 × 10⁻³ mSv/MBq), spleen (4.0 ± 0.4 × 10⁻³ mSv/MBq), and liver (3.9 ± 0.1 × 10⁻³ mSv/MBq), as shown in Table S1. The total whole-body effective dose was 3.4 ± 0.4 × 10⁻² mSv/MBq, which was far lower than that previously determined using ⁸⁹Zr-labeled antibodies (31, 32). In case of ¹⁷⁷Lu-NOTA-cA10-A3 (n = 4), organs with high absorbed doses were the spleen (1.5 ± 0.5 × 10⁻¹ mSv/MBq), lower large intestine (7.1 ± 0.3 × 10⁻² mSv/MBq), lung (5.3 ± 0.2 × 10⁻² mSv/MBq), stomach (2.9 ± 0.5 × 10⁻² mSv/MBq), and liver (2.3 ± 0.5 × 10⁻² mSv/MBq), as shown in Supplementary Table S2. The total whole-body effective doses were 3.3 ± 1.1 × 10⁻¹ mSv/MBq.

Cholangiocarcinoma is a deadly aggressive malignancy that tends to be diagnosed when the cancer is more advanced. Despite of radical surgical resection, overall 5-year survival rates have been reported to be 31% to 63% for intrahepatic cholangiocarcinoma and 27% to 37% for extrahepatic cholangiocarcinoma (33). Combined chemotherapy regimens such as gemcitabine and cisplatin for unresectable cholangiocarcinoma have been reported; however, these trials also showed little efficacy and low median survival rate (34, 35). Recently, several trials with additional targeted therapy using anti-EGFR antibody, cetuximab and tyrosine protein kinase inhibitor, sorafenib have been investigated, but targeted therapy did not seem to enhance the activity of chemotherapy and improve the overall survival (36, 37). Therefore, alternative novel therapeutic regimens are needed for the improvement of treatment efficacy and patient survival. Because the overexpression of L1CAM has been reported in cholangiocarcinoma, the assessment of L1CAM overexpression can provide important diagnostic aspect, which influences patient selection and management. In this study, tumor-targeting characteristics of ⁶⁴Cu-cA10-A3 as an immuno-PET imaging agent were successfully evaluated both in vitro and in vivo. Anti-L1CAM immuno-PET provides quantitative information for the L1CAM expression level in tumors. The feasibility of radioimmunotherapy using ¹⁷⁷Lu-labeled cA10-A3 was demonstrated in mice bearing L1CAM-expressing cholangiocarcinoma tumor.

In this study, the NOTA-cA10-A3 immunoconjugates were designed for diagnostic and therapeutic application as theranostic convergence bioradiopharmaceutical for ⁶⁴Cu and ¹⁷⁷Lu labeling. ⁶⁴Cu/¹⁷⁷Lu couple was chosen for convergence bioradiopharmaceutical based on their physical and radiochemical characteristics. The NOTA-cA10-A3 immunoconjugates were successfully labeled with ⁶⁴Cu or ¹⁷⁷Lu under mild conditions (weakly acidic pH and room temperature or 37°C) for 1 hour and then prepared with high radiolabeling yield and radiochemical purity (>98%) without further purification (Supplementary Fig. S2). DOTA has been preferred for labeling radiometals. However, recent investigations suggested that the NOTA conjugate is superior to the DOTA conjugate by showing better reaction kinetics, in vivo stability, and favorable pharmacokinetics and dosimetry (38–40). Although NOTA has been previously reported as a stable chelator for ⁶⁸Ga (41), this is the first report for the successful radiolabeling of an antibody with both ⁶⁴Cu and ¹⁷⁷Lu using NOTA.

In biodistribution study, ⁶⁴Cu-NOTA-cA10-A3 resulted in high tumor uptake, leading specific visualization of cholangiocarcinoma tumor (Figs. 2 and 3). The specificity of immuno-PET imaging was validated by blocking study (Fig. 3). There has been no quantitative PET imaging for L1CAM expression level in cholangiocarcinoma. We firstly provide a proof-of-principle that ⁶⁴Cu-NOTA-cA10-A3 can be used for immuno-PET imaging of L1CAM-expressing cholangiocarcinoma xenografted model. The quantification of targeted L1CAM PET imaging could help to guide clinical trial design by stratifying patients based upon L1CAM expression in the cholangiocarcinoma tumor. The applicability using immuno-PET for cholangiocarcinoma diagnosis still remains questionable because of location of cholangiocarcinoma lesions in liver. The successful detection of cholangiocarcinoma will require highly specific target selection and decreased non-specific liver uptake of the imaging agent. Although diagnosis with SPECT imaging using ¹¹¹In-labeled cG250 antibody for biliary cancer in clinical trial was attempted, this study showed ¹¹¹In-cG250 is not suitable for biliary cancer targeting owing to unfavorable tumor: non-tumor ratio in liver (42). Our study represents that the tumor-to-liver ratios on PET images at 48 hours were over 1 in cholangiocarcinoma tumor models, thus ⁶⁴Cu-NOTA-cA10-A3 immuno-PET imaging is expected to overcome the background liver radioactivity. Further preclinical researches are required to validate ⁶⁴Cu-NOTA-cA10-A3 as an imaging biomarker and to establish PET signals cut-off value for target selection in various orthotopic xenografts with high and low L1CAM expressing cholangiocarcinoma. Immuno-PET imaging potentially provides the information that could be used to optimize the radioimmunotherapy.

Considering L1CAM expression in tumors and the multifunctionality of L1CAM in tumor development and progression, L1CAM seems to be a promising target in anticancer therapy. Several studies have supported the feasibility of immunotherapy using anti-L1CAM antibody (11, 22, 23). However, complete cures were not achieved by antibody treatment alone. Hence, the treatment strategy combined L1CAM antibody targeting with therapeutic radionuclide could lead to the enhanced anti-tumor effect. Advantages of radioimmunotherapy over immunotherapy are direct killing tumor cells as well as delivering radiation even to neighboring cells inaccessible to the antibody by “cross-fire effect”. Therapeutic efficacy of radiolabeled anti-L1CAM antibody has been demonstrated in mice bearing ovarian carcinoma or neuroblastoma (24–26, 43).

In this study, we have set up an SCK-L1 subcutaneous xenograft model to allow the study of the therapeutic efficacy of L1CAM antibody-based radioimmunotherapy. Although there was no antitumor effect of unlabeled cA10-A3 antibody on SCK-L1 tumor, ¹⁷⁷Lu-cA10-A3 specially reduced the tumor volume (Fig. 5A). This therapeutic efficacy was predicted by ⁶⁴Cu-cA10-A3 biodistribution and immuno-PET imaging that showed high accumulation in SCK-L1 tumors. In addition, anti-L1CAM radioimmunotherapy markedly attenuated glucose metabolism (Fig. 6A and B). These data suggest that FDG-PET could be useful for the monitoring of therapeutic efficacy in L1CAM targeted radioimmunotherapy. In IHC analysis, radioimmunotherapy induced the increased apoptosis and decreased proliferation of tumor cells, compared with other treatments (Fig. 6C and D). These data represent that beta irradiation by ¹⁷⁷Lu-cA10-A3 could induce apoptosis and inhibit proliferation in cholangiocarcinoma. The previous study reported the repeated dose of cA10-A3 antibody treatment slightly retarded the tumor growth in Choi-CK tumor xenografted mouse (13). SCK cell line has the upregulated genes associated with tumor progression and metastasis (44), and acquired chemoresistance (45). In addition, SCK-L1 cell line is more progressive than SCK due to L1CAM overexpression. However, in our study, SCK-L1 tumor volume markedly decreased by a single dose of ¹⁷⁷Lu-cA10-A3 antibody.

We demonstrated that ¹⁷⁷Lu-labeled anti-cA10-A3 RIT effectively reduced the SCK-L1 xenografts without bodyweight loss. Nevertheless, the L1CAM expression on other normal cells such as myelomonocytic cells (46) and kidney tubule epithelial cells (47) must be considered in the context of the successful radioimmunotherapy strategy for translating to clinical trial. Because cA10-A3 is chimeric anti-human L1CAM antibody, toxicology profile using mouse/human L1-CAM binding version of cA10-A3 should be assessed in further preclinical studies. Previous studies reported that normal radiosensitive organs such as the kidney, lung, and bone marrow must be received less than 2,000, 1,500, and 100 cGy, respectively (48), and our dosimetry data are below the acceptable radiation limits to each organ (Supplementary Table S1). Compared with phase I trial of ¹⁷⁷Lu-J591, effective dose of ¹⁷⁷Lu-cA10-A3 was comparable with that of ¹⁷⁷Lu-J591 (1.23 vs. 1.3 rad/mCi). We selected ¹⁷⁷Lu as therapeutic radioisotope because it is a low-energy beta emitter with long half-life. ¹⁷⁷Lu-cA10-A3 remained over 10 %ID/g in SCK-L1-tumor at 14 days after administration (Fig. 2B), thus this radioimmunoconjugate is thought to permit the delivery of radiation for a period of time long enough to exhibit a therapeutic effect with minimal tissue penetration owing to the short path length (< 2 mm). Recently, a research was reported for estimating the kinetic parameters and cumulated activity of a diagnostic/therapeutic convergence radiopharmaceutical, ⁶⁴Cu-/¹⁷⁷Lu-labeled cetuximab (49). Although complexation and catabolism of copper and lutetium may be quite different, this study showed that the uptake rate constants of ⁶⁴Cu-cetuximab and ¹⁷⁷Lu-cetuximab are close, and their release rate constants are low in comparison with the uptake rate constants. Similarly, our study may be promising as a tool for determining whether ⁶⁴Cu-labeled anti-L1CAM antibody imaging might reliably predict dosimetry with ¹⁷⁷Lu-labeled anti-L1CAM antibody in clinical applications.

In conclusion, we evaluated the feasibility of quantitative immuno-PET and the efficacy of RIT with theranostic convergence bioradiopharmaceutical, ⁶⁴Cu-/¹⁷⁷Lu-NOTA-cA10-A3 in cholangiocarcinoma model. In vivo evaluation of L1CAM expression using ⁶⁴Cu-NOTA-cA10-A3 immuno-PET imaging could provide a compelling rationale for demanding specific and early diagnostic biomarker to increase the survival rate. Radioimmunotherapy with ¹⁷⁷Lu-NOTA-cA10-A3 showed significant therapeutic effect in L1CAM-expressing cholangiocarcinoma xenograft, so it would be utilized as a novel therapeutic agent for cholangiocarcinoma that lacks effective therapeutic regimens. These findings indicate that the theranostic convergence bioradiopharmaceutical, ⁶⁴Cu-/¹⁷⁷Lu-NOTA-cA10-A3 may be a potential immuno-PET imaging and radioimmunotherapeutic agent that guide clinicians in the selection of therapy for individual cholangiocarcinoma patients followed by personalized radioimmunotherapy.

No potential conflicts of interest were disclosed.

Conception and design: H.J. Hong, T.S. Lee

Development of methodology: I.H. Song, H.J. Hong, Y.S. Park

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): I.H. Song, H.J. Hong, J.I. Shin, B.S. Moon, T.S. Lee

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): I.H. Song, J.I. Shin, S.-K. Woo, J.H. Kang, T.S. Lee

Writing, review, and/or revision of the manuscript: I.H. Song, H.J. Hong, J.H. Kang, T.S. Lee

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.S. Jeong, K.I. Kim, Y.J. Lee, T.S. Lee

Study supervision: Y.S. Park, T.S. Lee

Other (performance of research project): J.H. Kang

This research was funded by a grant from Radiation Technology R&D program (2017M2A2A7A02070983) through NRF and a grant from the Korea Institute of Radiological and Medical Sciences (KIRAMS, 50536-2019 and 50461-2019) funded by Ministry of Science and ICT and Basic Science Research Program through NRF funded by the Ministry of Education (2017R1D1A1B03028106), Republic of Korea.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.