A highly stable human single-domain antibody-drug conjugate exhibits superior penetration and treatment of solid tumors

Abstract

The inefficient tumor penetration of therapeutic antibodies has hampered their effective use in treating solid tumors. Here, we report the identification of a fully human single-domain antibody (UdAb), designated as n501, targeting the oncofetal antigen 5T4. The high-resolution crystal structure indicates that n501 adopts a compact structure very similar to that of camelid nanobodies, and binds tightly to all eight leucine-rich repeats of 5T4. Furthermore, the UdAb n501 exhibits exceptionally high stability, with no apparent activity changes over 4 weeks of storage at various temperatures. Importantly, the UdAb-based antibody-drug conjugate (n501-SN38) showed much deeper tumor penetration, significantly higher tumor uptake, and faster accumulation at tumor sites than conventional IgG1-based antibody-drug conjugate (m603-SN38), resulting in improved tumor inhibition. These results highlight the potential of UdAb-based antibody-drug conjugates as a potential class of antitumor therapeutics with characteristics of high stability and strong tumor penetration for the effective treatment of solid tumors.

Keywords: 5T4; antibody-drug conjugate; oncofetal antigen; single-domain antibody; solid tumors; tumor penetration.

Graphical abstract

Figure 1

Exceptionally high stability of UdAb (A) The extracellular domain of 5T4 antigen that was expressed on the surface of tumor cells was biotinylated and used to biopanning from a human single-domain antibody (UdAb) library. (B) Experimental BLI setup and representative response curves fitted with a global 1:1 binding model to quantify the kinetics (k_on, k_off) and binding affinity (K_D) of n501 to 5T4. The recombinant human 5T4-Fc was immobilized on AHC biosensors and the antibodies were then assessed for association and dissociation over time. The red line indicates the fitted curve. (C) Aggregation of UdAb and nanobody estimated by analytical SEC-HPLC. (D) Thermal stabilities of antibodies as measured by CD spectroscopy. (E) Representative plot of the kinetics of monomer loss for UdAb and nanobody at its melting temperature T_m, measured in a centrifugation assay. (F and G) Binding capacity of UdAb and nanobody in the pH-resistance assay and long-term storage assay.

Figure 2

Structure of 5T4 in complex with the UdAb n501 (A) Schematic representation of UdAb n501 and a camelid nanobody against TNF (PDB: 5M2J), with CDR1, CDR2, and CDR3 regions highlighted. (B) The superimposed nanobodies show the different length of CDR3 between UdAb and camelid nanobody. (C) Overall structure of 5T4-n501 complex. 5T4 is represented in blue. n501 is shown in yellow, with the 3 CDR regions (CDR1–3) highlighted in gray, green, and red, respectively. Residues involved in the interaction are shown as sticks. (D–F) Interaction patterns of 5T4 with CDR1 (D), CDR2 (E), and CDR3 (F) on n501. The residues involved in interactions between 5T4 and n501 are shown as sticks. The red and black dashed lines indicate the salt bridge and hydrogen bond, respectively.

Figure 3

Design of UdAb for ADC conjugation (A) Ribbon structure of n501 overlaid with site locations of engineered cysteine mutations evaluated herein. (B) The binding kinetics of various n501 variants to 5T4, as measured by BLI. The red line indicates the fitted curve. (C) k_on (association rate constant), k_off (dissociation rate constant), and K_D (equilibrium dissociation constant) values of n501 variants binding to 5T4 were determined by fitting curves to a global 1:1 binding model. (D) Schematic representation depicts the discovery of a human monoclonal antibody (called m603) from human naive antibody phage-display library. (E) The binding capacity of m603 and isotype control (G12) to 5T4 was measured by ELISA. (F) Binding kinetics of m603 to 5T4 protein were determined by BLI. The k_on, k_off, and K_D values are shown. The red line indicates the fitted curve.

Figure 4

Deep tumor penetration of UdADC in both patient-derived organoids and tumor spheroids (A) Illustration of conventional ADC based on IgG (m603-SN38) and UdADC n501-SN38. The cytotoxic payload SN38 was coupled to thiols of antibodies via the CL2A linker, which is a cleavable PEG-8 and triazole-containing PABC-peptide-mc linker. (B) Cell cytotoxicity assays were performed on BxPC-3, Huh-7, and FCHO cell lines. Graphs are representative data derived from different cell lines showing the mean percentage of cell cytotoxicity ± SEM (n = 3). (C) Internalization of ADCs in BxPC-3 cell lines. (D and E) Penetration of ADCs in patient-derived pancreatic tumor organoids (D) and multicellular tumor spheroids (E), determined by confocal microscopy. BxPC-3 cells, established organoids, and spheroids were incubated with Dylight650-labeled ADCs, and cell images were analyzed using a digital laser confocal microscope. Red represents labeled ADCs and blue represents nucleus. Data from 1 representative experiment of 3 independent experiments are shown.

Figure 5

UdADC exhibits more potent antitumor effects in BxPC-3-xenograft mice (A) Athymic nude mice bearing BxPC-3 human tumor xenografts were intravenously injected with DyLight650-labeled n501-SN38 or m603-SN38 and imaged under anesthesia at several time points post-injection using the caliper IVIS Lumina II system. (B) The corresponding values of average fluorescence radiance (p/s/cm²/sr), as determined by Living Image software. (C) The pharmacokinetics of n501-SN38 and m603-SN38 ADCs were calculated in nude mice bearing BxPC-3 tumor xenografts after a single dose by intravenous injection. (D–F) Schematic representation of in vivo experiments. BxPC-3 xenograft tumors were established in mice and treated intravenously 4 times every 3 days with G12-SN38 (red), m603 (green), or m603-SN38 (purple) at 25 mg/kg, and 6 times every 2 days with PBS as vehicle (blue), n501 (orange), or n501-SN38 (magenta). Data are expressed as means ± SEMs of the average tumor areas (mm³) (D), tumor weight (E), and individual tumor areas (F), with n = 6 mice per group.