Pharmacokinetics of Engineered Antibodies

1. Antibody technology to enhance brain penetration and brain retention

  To overcome the limitation imposed by the BBB, receptor-mediated transcytosis (RMT) has attracted considerable attention. A representative approach uses the transferrin receptor (TfR): by binding to TfR expressed on brain microvascular endothelial cells that constitute the BBB, antibodies can be taken up and transcytosed more efficiently, and ~10-fold improvements in brain exposure have been reported [1]. However, anti-TfR antibodies also bind TfR expressed in peripheral tissues such as the liver, leading to faster systemic clearance and rapid reduction in the antibody supply to the brain. In addition, antibodies can be eliminated from the brain through reverse transport back to plasma, bulk flow of interstitial fluid (ISF), uptake and degradation by neural cells, and other pathways. As a result, achieving long-lasting antibody activity in the brain after a single administration remains challenging.

  To address this, we conceived a dual-targeting concept that combines (i) improved BBB transport by binding to a receptor involved in BBB transport (e.g., TfR) with (ii) improved retention by binding to a protein expressed in the brain (Figure 1). In other words, the antibody crosses the BBB via TfR binding and then stays in the brain by binding a molecule with slow turnover. As one example, we used MOG (myelin oligodendrocyte glycoprotein) as a brain-retention target, and designed an antibody that integrates both TfR binding and MOG binding in a single molecule (anti-MOG/TfR). MOG is expressed on oligodendrocytes and is a component of the myelin sheath; it is known to have an exceptionally slow turnover, with a reported half-life of approximately two months [2].

  We generated four antibodies (Figure 2): a control antibody that does not bind mouse antigens (Ctrl), an anti-MOG antibody (MOG), a control antibody fused with a TfR-binding domain (Ctrl/TfR), and an anti-MOG antibody fused with a TfR-binding domain (MOG/TfR). After a single intravenous dose in mice, we measured antibody concentrations in the brain. MOG/TfR showed high brain concentrations early after dosing and maintained high levels for at least four months (Day 112) (Figure 2). Ctrl/TfR achieved high brain concentrations immediately after dosing but declined rapidly, whereas MOG gradually accumulated in the brain. In MOG/TfR, the weaknesses of each single approach were compensated, resulting in a clear synergistic effect. We have also observed similar synergy when targeting proteins other than TfR and MOG, suggesting that this concept could enable retention in various brain cell types.

  We also evaluated brain distribution by administering fluorescently labeled antibodies to mice, collecting brains on Day 1 and Day 7, followed by fixation, clearing, and observation using light-sheet microscopy (Figure 3). Only weak signals were detected with Ctrl. Signals were found in Ctrl/TfR on day 1, but decreased from Day 1 to Day 7, consistent with the concentration–time profiles. MOG showed localized distribution (e.g., around the fimbria of the hippocampus), whereas MOG/TfR distributed broadly throughout the brain. These observations suggest that combining TfR and MOG binding not only increases brain exposure but also influences whole-brain distribution and residence—a property that neither TfR binding nor MOG binding alone could achieve.

  Dual-targeting antibodies that combine a BBB transport target with a brain-resident target could serve as a powerful platform for delivering therapeutics to the brain. For example, by fusing therapeutic payloads such as peptides, cytokines, neurotrophic factors, or enzymes, one could deliver these molecules efficiently into the brain, maintain them there for an extended period, and thereby enhance and prolong pharmacological effects. We strongly hope that this technology will contribute to the development of novel therapeutics for central nervous system diseases.

2. Antibody technology to enhance cartilage penetration and retention

  As noted above, articular cartilage is another tissue that is difficult for macromolecules such as antibodies to access. Cartilage is avascular and consists of a high-density extracellular matrix formed by collagen and proteoglycans, which is known to impede diffusion of large molecules. Moreover, molecules administered into the synovial cavity—filled with synovial fluid—can be rapidly cleared from the joint by the transfer of synovial fluid into lymph and blood circulation. This reduces the supply of molecules to cartilage, and exchange of synovial fluid can also promote loss of molecules from cartilage. Therefore, for cartilage as well, it is important both to drive macromolecules into the tissue and to retain them once they have penetrated.

  We therefore developed a concept that combines (i) binding to aggrecan, a major component of the cartilage matrix, with (ii) molecular downsizing of the antibody (conversion to Fab and F(ab’)2 formats) (Figure 4). The idea is that aggrecan binding at the cartilage surface, together with smaller molecular size, increases penetration into cartilage, and once inside, binding to aggrecan within the matrix promotes retention. Aggrecan is known to have a very long half-life as a matrix component (approximately 11–12 years in healthy humans) [3], making it an attractive retention target.

  We generated a control IgG that does not bind to rabbit antigens (Ctrl IgG) and aggrecan-binding antibody formats: IgG, F(ab’)2, and Fab (Aggrecan IgG, Aggrecan F(ab’)2, Aggrecan Fab). After a single intra-articular administration into rabbit knee synovial fluid, we measured antibody concentrations in cartilage over time (Figure 5). Cartilage concentration–time profiles differed markedly by molecular format: Ctrl IgG showed low levels immediately after dosing, whereas Aggrecan IgG achieved ~10-fold higher concentrations. Furthermore, downsizing to F(ab’)2 and Fab increased cartilage concentrations even more. From a retention standpoint, aggrecan binders exhibited longer apparent elimination half-lives than Ctrl IgG (Ctrl IgG: 22.0 h; Aggrecan IgG: 40.5 h; Aggrecan F(ab’)2: 68.2 h; Aggrecan Fab: not calculated because a clear terminal slope could not be determined), indicating improved retention as well [4].

  To further assess intratissue distribution, we tested ex vivo penetration of fluorescently labeled antibodies using cartilage explants collected from rabbit knee joints (Figure 6). With Ctrl IgG, little fluorescence was detected. With Aggrecan IgG, signals were observed at the cartilage surface, but penetration into deeper regions was limited. In contrast, Aggrecan F(ab’)2 reached deeper into cartilage, and the smallest format, Aggrecan Fab, showed the strongest fluorescence throughout the explant, including internal regions—suggesting that more antibody was transferred into the deep cartilage.

  Together, these results demonstrate that introducing aggrecan binding and downsizing the antibody can enhance both cartilage penetration and retention. As a delivery tool for cartilage, this approach could be applicable to small molecules, peptides, and even single-domain antibodies such as VHH. We hope that this technology will eventually contribute to therapies for articular cartilage diseases.

  In summary, we considered tissue distribution in two parts—(1) getting into tissue (transport/penetration) and (2) staying once inside (retention)—and pursued antibody concepts, molecular engineering, and pharmacokinetic studies that address both simultaneously. By combining strategies that promote transport/penetration with strategies that promote retention, we found the potential to sustain sufficient antibody exposure even in tissues where distribution has traditionally been difficult. We expect that such technologies will expand the areas where antibodies can work and lead to new therapeutics. In addition, the research process illustrated here—identifying an essential challenge from a pharmacokinetic perspective and suggesting solutions based on the pharmacokinetic data—is, in our view, one distinctive way that pharmacokinetic scientists can contribute to drug discovery. We would like to continue developing drug discovery approaches that open up further possibilities of modalities from the standpoint of pharmacokinetics.

  Across all four installments, we have introduced pharmacokinetics-related topics in engineered antibodies. Part 1 provided an introduction to the fundamentals of antibody pharmacokinetics; Part 2 covered half-life extension by FcRn-binding engineering; Part 3 discussed recycling and sweeping antibodies; and in Part 4,we presented recent insights on tissue-distribution-enhancing antibodies. We hope these articles provide useful hints for your future research. Thank you very much for reading to the end.