Site-Specifically Conjugated Single-Domain Antibody Successfully Identifies Glypican-3-Expressing Liver Cancer by Immuno-PET

Abstract

Primary liver cancer is the third leading cause of cancer-related deaths, and its incidence and mortality are increasing worldwide. Hepatocellular carcinoma (HCC) accounts for 80% of primary liver cancer cases. Glypican-3 (GPC3) is a heparan sulfate proteoglycan that histopathologically defines HCC and represents an attractive tumor-selective marker for radiopharmaceutical imaging and therapy for this disease. Single-domain antibodies are a promising scaffold for imaging because of their favorable pharmacokinetic properties, good tumor penetration, and renal clearance. Although conventional lysine-directed bioconjugation can be used to yield conjugates for radiolabeling full-length antibodies, this stochastic approach risks negatively affecting target binding of the smaller single-domain antibodies. To address this challenge, site-specific approaches have been explored. Here, we used conventional and sortase-based site-specific conjugation methods to engineer GPC3-specific human single-domain antibody (HN3) PET probes. Methods: Bifunctional deferoxamine (DFO) isothiocyanate was used to synthesize native HN3 (nHN3)-DFO. Site-specifically modified HN3 (ssHN3)-DFO was engineered using sortase-mediated conjugation of triglycine-DFO chelator and HN3 containing an LPETG C-terminal tag. Both conjugates were radiolabeled with ⁸⁹Zr, and their binding affinity in vitro and target engagement of GPC3-positive (GPC3⁺) tumors in vivo were determined. Results: Both ⁸⁹Zr-ssHN3 and ⁸⁹Zr-nHN3 displayed nanomolar affinity for GPC3 in vitro. Biodistribution and PET/CT image analysis in mice bearing isogenic A431 and A431-GPC3⁺ xenografts, as well as in HepG2 liver cancer xenografts, showed that both conjugates specifically identify GPC3⁺ tumors. ⁸⁹Zr-ssHN3 exhibited more favorable biodistribution and pharmacokinetic properties, including higher tumor uptake and lower liver accumulation. Comparative PET/CT studies on mice imaged with both ¹⁸F-FDG and ⁸⁹Zr-ssHN3 showed more consistent tumor accumulation for the single-domain antibody conjugate, further establishing its potential for PET imaging. Conclusion: ⁸⁹Zr-ssHN3 showed clear advantages in tumor uptake and tumor-to-liver signal ratio over the conventionally modified ⁸⁹Zr-nHN3 in xenograft models. Our results establish the potential of HN3-based single-domain antibody probes for GPC3-directed PET imaging of liver cancers.

Keywords: GPC3; Nanobody; glypican-3; immuno-PET; liver cancer; molecular imaging.

Graphical abstract

FIGURE 1.

Conventional lysine conjugation vs. site-specific conjugation. Shown are synthetic schema and structures of ⁸⁹Zr-nHN3 (A) and ⁸⁹Zr-ssHN3 (B). HEPES = 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; PBS = phosphate-buffered saline.

FIGURE 2.

Single-domain antibody conjugates retain GPC3 affinity. Shown are biolayer interferometry (left) and radioligand saturation binding assay (right) results for stochastically modified single-domain antibody nHN3-DFO (A) and ssHN3-DFO (B), with determined K_D values. K_D values for biolayer interferometry assays are average of measurements using 4 different HN3 concentrations, whereas K_D values for saturation assays are average of 3 independent replicates.

FIGURE 3.

Single-domain antibody PET tracers successfully image tumors engineered to express GPC3. Shown are representative PET/CT images (A) and calculated SUVs (B) of ⁸⁹Zr-ssHN3 and ⁸⁹Zr-nHN3 in A431 and A431-GPC3⁺ tumor–bearing mice (n = 3) 1 h after injection. Full ex vivo biodistribution data for mice bearing A431 and A431-GPC3⁺ tumors are reported in Supplemental Figures 17 and 18. ***P < 0.005.

FIGURE 4.

Single-domain antibody PET tracers successfully image GPC3⁺ liver tumor xenografts. Shown are selected ex vivo biodistribution of ⁸⁹Zr-ssHN3 and ⁸⁹Zr-nHN3 (A) and tumor-to-tissue ratios of HepG2 tumor-bearing mice (n = 4) (B). Full 12-organ biodistribution results for mice bearing HepG2 tumors are reported in Supplemental Figures 19 and 20. *P < 0.05. **P < 0.01. ***P < 0.005.

FIGURE 5.

⁸⁹Zr-ssHN3 PET tracer is superior to ¹⁸F-FDG for imaging liver tumors. Shown are PET/CT images of mice (n = 4) bearing HepG2 tumors injected with ¹⁸F-FDG (A) and ⁸⁹Zr-ssHN3 (B). At top is SUV comparison, and at bottom is tumor-to-tissue ratios for both tracers. Error bars represent SD. Full SUV analysis of ⁸⁹Zr-ssHN3 and ⁸⁹Zr-nHN3 is reported in Supplemental Figures 21–22. *P < 0.05. ***P < 0.005.