Radiolabeled cCPE Peptides for SPECT Imaging of Claudin-4 Overexpression in Pancreatic Cancer

Abstract

Overexpression of tight-junction protein claudin-4 has been detected in primary and metastatic pancreatic cancer tissue and is associated with better prognosis in patients. Noninvasive measurement of claudin-4 expression by imaging methods could provide a means for accelerating detection and stratifying patients into risk groups. Clostridium perfringens enterotoxin (CPE) is a natural ligand for claudin-4 and holds potential as a targeting vector for molecular imaging of claudin-4 overexpression. A glutathione S-transferases (GST)-tagged version of the C terminus of CPE (cCPE) was previously used to delineate claudin-4 overexpression by SPECT but showed modest binding affinity and slow blood clearance in vivo. Methods: On the basis of the crystal structure of cCPE, a series of smaller cCPE_194-319 mutants with putatively improved binding affinity for claudin-4 was generated by site-directed mutagenesis. All peptides were conjugated site-specifically on a C-terminal cysteine using maleimide-diethylenetriamine pentaacetate to enable radiolabeling with ¹¹¹In. The binding affinity of all radioconjugates was evaluated in claudin-4-expressing PSN-1 cells and HT1080-negative controls. The specificity of all cCPE mutants to claudin-4 was assessed in HT1080 cells stably transfected with claudin-4. SPECT/CT imaging of BALB/c nude mice bearing PSN-1 or HT1080 tumor xenografts was performed to determine the claudin-4-targeting ability of these peptides in vivo. Results: Uptake of all cCPE-based radioconjugates was significantly higher in PSN-1 cells than in HT1080-negative controls. All peptides showed a marked improvement in affinity for claudin-4 in vitro when compared with previously reported values (dissociation constant: 2.2 ± 0.8, 3 ± 0.1, 4.2 ± 0.5, 10 ± 0.9, and 9.7 ± 0.7 nM). Blood clearance of [¹¹¹In]In-cCPE_194-319, as measured by SPECT, was considerably faster than that of [¹¹¹In]In-cCPE.GST (half-life, <1 min). All radiopeptides showed significantly higher accumulation in PSN-1 xenografts than in HT1080 tumors at 90 min after injection of the tracer ([¹¹¹In]In-cCPE_194-319, 2.7 ± 0.8 vs. 0.4 ± 0.1 percentage injected dose per gram [%ID/g], P < 0.001; [¹¹¹In]In-S313A, 2.3 ± 0.9 vs. 0.5 ± 0.1 %ID/g, P < 0.01; [¹¹¹In]In-S307A + N309A + S313A, 2 ± 0.4 vs. 0.3 ± 0.1 %ID/g, P < 0.01; [¹¹¹In]In-D284A, 2 ± 0.2 vs. 0.7 ± 0.1 %ID/g, P < 0.05; [¹¹¹In]In-L254F + K257D, 6.3 ± 0.9 vs. 0.7 ± 0.2 %ID/g, P < 0.001). Conclusion: These optimized cCPE-based SPECT imaging agents show great promise as claudin-4-targeting vectors for in vivo imaging of claudin-4 overexpression in pancreatic cancer.

Keywords: SPECT imaging; claudin-4; early diagnosis; pancreatic ductal adenocarcinoma.

Graphical abstract

FIGURE 1.

(A) Claudin-4 expression in panel of human pancreatic cancer cell lines detected by Western blot. (B) Immunofluorescence staining of claudin-4 expression in human PDAC PSN-1 cell line and claudin-4–nonexpressing HT1080 fibrosarcoma cells (green: claudin-4; blue: 4′,6-diamidino-2-phenylindole, representing cell nuclei; scale bar is 20 μm). (C) Coomassie blue–stained sodium dodecyl sulfate and polyacrylamide gel electrophoresis gel of purified S313A variant. All peptides were obtained with >95% purity. (D) Schematic representation of bioconjugation and ¹¹¹In radiolabeling strategies for cCPE mutants. cCPE peptides were initially reduced with tris(2-carboxyethyl)phosphine hydrochloride and site-specifically conjugated to maleimide-DTPA. Resulting bioconjugates were radiolabeled with ¹¹¹In, with excellent radiochemical purity (>99%) and yield (>95%). RCP = radiochemical purity; RT = room temperature; TCEP = tris(2-carboxyethyl)phosphine hydrochloride.

FIGURE 2.

(A) PSN-1 and HT1080 cells were exposed for 2 h at 4°C to increasing concentrations of ¹¹¹In-labeled cCPE peptides, and extent of cell binding was determined. (B) Experimental results for all tested cCPE radiopeptides in PSN-1 and HT1080 cells.

FIGURE 3.

(A) Immunocytochemistry of claudin-4 (top) and claudin-3 (bottom) in HT1080 cells stably transfected with human claudin-4 (HT1080-hCLDN4). Green fluorescence indicates claudin-4–positive staining, and blue (4′,6-diamidino-2-phenylindole [DAPI]) fluorescence indicates cell nuclei. Scale bar is 20 μm. (B) An excess of cold, unlabeled cCPE.GST (100-fold) was used to block binding of cCPE radiopeptides to PSN-1 and HT1080-hCLDN4 cells. ***P < 0.001.

FIGURE 4.

(A) Representative SPECT/CT images of mice carrying tumor xenografts (white circles) of PSN-1 (claudin-4–positive) or HT1080 (claudin-4–negative) cells 90 min after intravenous injection of [¹¹¹In]In-cCPE^L254F+K257D. Coronal and axial sections through tumor are shown. (B) Ex vivo tumor uptake (%ID/g), tumor-to-blood ratios, and tumor-to-liver ratios of ¹¹¹In-radiolabeled cCPE peptides 2 h after intravenous administration of tracer. *P < 0.05. **P < 0.01. ***P < 0.001. (C) Autoradiography images of PSN-1 and HT1080 tumor xenograft sections of mice injected with [¹¹¹In]In-cCPE^L254F+K257D and corresponding confocal images of immunofluorescence staining of claudin-4 (green: claudin-4; blue: 4′,6-diamidino-2-phenylindole; scale bar is 20 μm). PSN-1 tumor sections showed significantly higher tracer uptake than negative control. (D) Biodistribution of ¹¹¹In-radiolabeled cCPE peptides 2 h after intravenous administration of tracer.