High-affinity transferrin receptor binding improves brain delivery of bispecific antibodies at tracer dose

Abstract

Background: Transferrin receptor (TfR)-mediated transcytosis is a well-established method for delivering biologic therapeutics and diagnostics to the brain. Although moderate affinity towards TfR is beneficial for TfR-mediated brain delivery at therapeutic doses, emerging evidence has indicated that high TfR affinity may be more beneficial at tracer doses. With the development of antibody-based PET radioligands for neurodegenerative diseases, such as Alzheimer's disease, understanding the pharmacokinetics of TfR-binders at tracer dose is essential. Thus, this study aimed to evaluate the effect of TfR affinity on brain uptake at a tracer dose in both wild-type (WT) and amyloid-beta (Aβ) pathology presenting mice and to demonstrate the usability of TfR-mediated brain delivery of immunoPET diagnostic radioligands to visualize intrabrain Aβ pathology in vivo.

Methods: Three different affinity variants of anti-mouse TfR-binding antibody 8D3, engineered by alanine point mutations, were selected. Bispecific antibodies were designed with knob-into-hole technology with one arm targeting TfR (8D3) and the other arm targeting human Aβ (bapineuzumab). Antibody affinities were measured in an in vitro cell assay. In vivo pharmacokinetic analyses of radioiodinated bispecific antibodies and bapineuzumab in brain, blood and peripheral organs were performed over 7 days post-injection in WT mice and a model of Aβ pathology (App^NL-G-F). The strongest TfR affinity bispecific antibody was also evaluated as a positron emission tomography (PET) radioligand for detecting Aβ pathology in WT and App^NL-G-F mice.

Results: The three bispecific antibodies bound to TfR with affinities of 10 nM, 20 nM and 240 nM. Independent of genotype, stronger TfR-affinity resulted in higher initial brain uptake. The two higher-affinity bispecific antibodies behaved similarly and differentiated between WT and App^NL-G-F mice earlier than the lowest affinity variant. Finally, the 10 nM bispecific antibody was able to clearly differentiate between WT and App^NL-G-F mice when used as a PET radioligand.

Conclusion: This study supports the hypothesis that stronger TfR affinity enhances brain uptake at a tracer dose. With the more effective detection of Aβ pathology, stronger TfR affinity is a crucial design feature for future bispecific immunoPET radioligands for intrabrain targets via TfR-mediated transcytosis.

Keywords: Affinity; Alzheimer’s disease; Amyloid-β; Positron emission tomography (PET); Transferrin receptor (TfR)-mediated transcytosis.

Fig. 1

Generation of bispecific antibodies with varying affinity to mTfR. (A) Schematic antibodies illustrating the KiH design of bispecific Bapi-8D3 and monospecific Bapi. (B) Mean fluorescent intensity (MFI) from the on-cell mTfR affinity assay with Bapi-8D3_WT, Bapi-8D3_Y32A, Bapi-8D3_Y52A, Bapi. EC50 values listed below the legend. (C) ELISA analysis of bivalent (Bapi) and monovalent (Bapi-8D3_Y32A) binding to Aβ with EC50 values listed below the legend

Fig. 2

Ex vivo biodistribution of three bispecific affinity variants and Bapi in WT and App^NL−G−F mice over 168 h post-injection. (A) Brain concentrations (%ID/g) of [¹²⁵I]I-Bapi-8D3_WT, [¹²⁵I]I-Bapi-8D3_Y32A, [¹²⁵I]I-Bapi-8D3_Y52A and [¹²⁵I]I-Bapi in WT and App^NL−G−F mice 4, 24, 72, and 168 h post-injection. The first time when the concentration in App^NL−G−F mice is significantly higher than the concentration in WT mice is indicated by a solid arrow for [¹²⁵I]I-Bapi-8D3_WT and [¹²⁵I]I-Bapi-8D3_Y32A and a dashed arrow for [¹²⁵I]I-Bapi-8D3_Y52A and [¹²⁵I]I-Bapi. (B) Whole blood pharmacokinetics (%ID/g of blood) over 168 h post-injection. (C) Brain-to-blood ratio in WT and App^NL−G−F mice 4, 24, 72, and 168 h post-administration. AUC from the brain (AUC_brain, D) and blood (AUC_blood, E) pharmacokinetic curves of each antibody in WT and App^NL−G−F mice (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p < 0.0001). (F) The ratio of AUC_brain/AUC_blood for each antibody in WT and App^NL−G−F mice. Concentrations (%ID) in the capillary enriched pellet (G) and parenchymal supernatant (H) from capillary depletion of cortices at 4, 24, 72 and 168 h post-administration. (I) Percent in plasma fraction of total blood in WT and App^NL−G−F mice 4, 24, 72, and 168 h post-administration

Fig. 3

Ex vivo autoradiography of three bispecific affinity variants and Bapi in WT and App^NL−G−F mice over 168 h post-injection. Autoradiography in sagittal sections from WT and App^NL−G−F mice injected with [¹²⁵I]I-Bapi-8D3_WT, [¹²⁵I]I-Bapi-8D3_Y32A, [¹²⁵I]I-Bapi-8D3_Y52A and [¹²⁵I]I-Bapi at 4, 24, 72 and 168 h post-injection. The intensity for [¹²⁵I]I-Bapi-8D3_WT and [¹²⁵I]I-Bapi-8D3_Y32A are scaled differently than [¹²⁵I]I-Bapi-8D3_Y52A and [¹²⁵I]I-Bapi

Fig. 4

More even antibody distribution with higher affinity to mTfR. Ex vivo autoradiography and Aβ immunohistochemistry (IHC) in sections from App^NL−G−F mice 72 h post-injection of [¹²⁵I]I-Bapi-8D3_WT, [¹²⁵I]I-Bapi-8D3_Y32A, [¹²⁵I]I-Bapi-8D3_Y52A and [¹²⁵I]I-Bapi. The intensity for [¹²⁵I]I-Bapi-8D3_WT and [¹²⁵I]I-Bapi-8D3_Y32A are scaled differently than [¹²⁵I]I-Bapi-8D3_Y52A and [¹²⁵I]I-Bapi

Fig. 5

PET imaging reveals more intense signal in App^NL−G−F _mice compared to WT. (A) Sagittal PET/CT images acquired at 72 h post administration of [¹²⁴I]I-Bapi-8D3_WT. The antibody was retained more in the brain of the App^NL−G−F mice with Aβ pathology than in WT mice without pathology demonstrating that it could be used as a diagnostic tool to visualize the regional distribution of Aβ pathology in the living brain. (B) Quantification of [¹²⁴I]I-Bapi-8D3_WT brain concentrations based on PET images in different brain regions expressed as %ID/g brain. (C) Brain concentrations of [¹²⁴I]I-Bapi-8D3_WT based on gamma counting and expressed as %ID/g brain. (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p < 0.0001)