Preclinical optimization of antibody-based radiopharmaceuticals for cancer imaging and radionuclide therapy-Model, vector, and radionuclide selection

Abstract

Intact antibodies and their truncated counterparts (eg, Fab, scFv fragments) are generally exquisitely specific and selective vectors, enabling recognition of individual cancer-associated molecular phenotypes against a complex and dynamic biomolecular background. Complementary alignment of these advantages with unique properties of radionuclides is a defining paradigm in both radioimmunoimaging and radioimmunotherapy, which remain some of the most adept and promising tools for cancer diagnosis and treatment. This review discusses how translational potency can be maximized through rational selection of antibody-nuclide couples for radioimmunoimaging/therapy in preclinical models.

Keywords: animal model; antibody; imaging; radionuclide; therapy.

Figure 1

In vivo pharmacokinetics of tumor-targeted antibody-based vectors as a function of their molecular weight and size. (Left to Right) Full-length (intact) antibodies yield the highest tumor uptake (% ID/g) and tumor-to-blood ratios (image contrast/therapeutic index) at later time points (>72 h post-injection). This is mostly attributed to their high molecular weight (~150 kDa) making them ineligible for renal clearance and the in vivo recycling of the antibodies by the neonatal Fc-receptor (FcRn) into the bloodstream – thus increasing their bioavailability to target tumors over multiple passes while within the systemic circulation. Various engineered formats of the antibody such as scFv-Fc-fusion constructs (~100–150 kDa) and minibodies are intermediate antibody fragments that achieve relatively lower uptake in the tumor, but their faster in vivo clearance profile enables them to yield high tumor-to-background contrast at early time points. These constructs have been synthesized for the purpose of next day (24 h p.i.) radioimmunoimaging. Finally, a combination of their structural composition (lack of Fc-mediated recycling), extremely rapid renal clearance due to low molecular weight of the smaller tumor-targeting antibody-based constructs (diabodies and scFv(s)) limits their in vivo bioavailability and uptake in the tumor. However, diabodies (~50 kDa) can be used for same-day imaging and yield high contrast images as early as 2–4 h p.i. Adapted with permission from Knowles and Wu (13).

Figure 2

The influence of immunodeficiency status on tumor uptake of the targeting vector. (A) Volume-rendered PET-CT images of nude (Nu/Nu) – less immunodeficient mice versus NOD-SCID gamma (NSG) – highly immunodeficient mice bearing subcutaneously xenografted small cell lung cancer (SCLC) tumors on the right flank. Each animal was injected with the same amount [7.4–9.3 MBq; (200–250 µCi), 35–44 µg] of a ⁸⁹Zr-labeled humanized monoclonal antibody targeting delta-like protein3 (DLL3) in SCLC (57). However, the remarkably different uptake patterns of the same radiotracer in two different biological backgrounds are evident from the higher uptake (indicated by easier delineation of the organs and a relatively higher intensity of signal) of the antibody-based tracer in the spleen, bones and the liver of the highly immunodeficient NSG mouse strain. Images were scaled individually - though not comprehensible from these images, the tumor uptake in NSG mice is significantly lower than in the nude mice. (B) A graph depicting the inverse correlation between immunodeficiency status of the preclinical mouse model background and tumor uptake versus a progressively increasing non-target accumulation of humanized (non-Fc-silent) antibody-based targeting vectors in mice with highly immunodeficient status.

Figure 3

Decay properties and imaging characteristics of common positron emitters for traditional and pre-targeted immunoPET. (A) Positron energy spectra of selected PET nuclides outlined in Table 2 (data obtained from the RADAR database). (B) Continuous slowing down approximation (CDSA) range of beta particles in selected ICRP reference tissues (data obtained from the NIST database). (C) Comparison of image quality and resolution with six clinically relevant positron emitters in a preclinical Derenzo phantom (63) with rod diameters of 0.8, 0.9, 1.0, 1.1, 1.2, and 1.3 mm (counter-clockwise from bottom-right sector). (D) ¹²⁴I-labeled (top, E_β+,avg = 820 keV) versus ⁸⁹Zr-labeled (bottom; E_β+,avg = 396 keV) αHER2 Fab PET maximum intensity projections (64). Note the variation in clearance route arising from chemical differences imparted to the Fab construct via labeling with different radionuclides. Figure (C) adapted from Bunka et al. (63), with permission. Figure (D) adapted from Mendler et al. (64), with permission.

Figure 4

Relevant considerations for optimization of imaging performance with antibodies and fragments. Adapted from Zhou, et al. (82).

Figure 5

Spatially heterogeneous radiation dose deposition for different therapeutic radionuclides. A) Heterogeneous uptake present in high-resolution preclinical SPECT imaging. B) Segmentation of diseased tissue / source definition for dose calculations (40% max SPECT signal thresholding of tumor-associated activity). C) Radiation dose rate* computed in PHITS v2.88 with CT overlay; note the more uniform dose deposition with ⁹⁰Y but significant irradiation of surrounding healthy tissue. *Non-penetrating component only; photon components are assumed negligible in calculation; dose rates are individually normalized to the maximum voxel value for each image.

Figure 6

DNA damage stemming from interaction of alpha, beta, and Auger emissions. Note the high density of ionization along the alpha particle track, and spatially compact region of ionization events produced via Auger cascade, reflective of the high LET (and cytotoxic potential) of these radiations.

Figure 7

Optimization of radioimmunotherapy at the tissue level. A) Schematic representation of an Fab-dsFv (133). An disulfide bonded Fab is conjugated to an anti-HSA variable fragment (Fv) domain containing an incorporated disulfide bond via an amino acid linker. B) Schematic representation of the pretargeting strategy (25). C) Sagittal PET images of [¹²⁴I]I-8H9 antibody following intrathecal injection via an Ommaya reservoir. Images shows good accumulation at the leptomeningeal tumor sites (119).