Cellular assays to quantify target-independent clearance of therapeutic molecules

Abstract

The present disclosure is directed to in vitro cell-based methods useful for predicting in vivo pharmacokinetic properties of candidate therapeutic proteins. In particular, the cell-based assays described herein are useful to predict in vivo non-target dependent clearance and subcutaneous bioavailability of candidate therapeutic proteins.

Description

CELLULAR ASSAYS TO QUANTIFY TARGET-INDEPENDENT CLEARANCE OF THERAPEUTIC MOLECULES

[0001] The benefit under 35 U.S.C. § 119(e) of U.S. Provisional Application Nos. 63 / 523,868 filed June 28, 2023, and 63 / 547,184 filed November 3, 2023, is hereby claimed, and the disclosures thereof are hereby incorporated by reference herein.FIELD

[0002] The present disclosure relates to cell-based methods for predicting in vivo targetindependent clearance of a biological molecule.INCORPORA TION-BY-REFERENCE OF MATERIAL SUBMITTED ELECTRONICALLY

[0003] Incorporated by reference in its entirety herein is a nucleotide / amino acid sequence listing submitted concurrently herewith and identified as follows: One 8 kilobyte XML document named “10587-W001-SEC_SequenceListing.xml,” created on June 27, 2024.BACKGROUND OF VARIOUS EMBODIMENTS

[0004] The pharmacokinetics (PK) of monoclonal (mAbs) and multispecific antibodies is determined by both target-dependent and target-independent clearance (CL) pathways. Target independent CL (CLind) as pertains to mAbs is described mathematically by first-order, linear kinetics (i.e. (dA / dt) / [Drug] = CLind). Mechanistically, the rate of CLind results from the interplay of at least two competing processes, where non-specific endocytosis is balanced by intracellular salvage / recycling via the neonatal Fc receptor (FcRn) that returns drug to the systemic circulation (Ovacik, M., and Lin, K. (2018) Tutorial on Monoclonal Antibody Pharmacokinetics and Its Considerations in Early Development. Clin Transl Sci 11, 540-552; Ryman, J. T., and Meibohm, B. (2017) Pharmacokinetics of Monoclonal Antibodies. CPT Pharmacometrics Syst Pharmacol 6, 576-588; Huisinga, et al., Target-Driven Pharmacokinetics of Biotherapeutics, in Pharmaceutical Sciences Encyclopedia, pp 1-15, and Meno-Tetang, G. M. L. Target-Driven Pharmacokinetics of Biotherapeutics, in Pharmaceutical Sciences Encyclopedia, pp 1-12). Non-specific endocytosis of therapeutic proteins can be driven by fluid-phase uptake and non-specific adsorptive internalization. Fluid-phase uptake is a mechanism whereby solutes are internalized along with the extracellular fluid whose extent is directly proportional to the solute concentration (Steinman, et al. (1983) Endocytosis and the recycling of plasma membrane. J Cell Biol 96, 1 -27, and Besterman, J. M., and Low, R. B. (1983) Endocytosis: a review of mechanisms and plasma membrane dynamics. Biochem J 210, 1-13). Non-specific adsorptive endocytosis occurs when aprotein non-specifically interacts with the cell membrane, such as charge-based attraction, that results in its internalization. Fluid-phase is constitutive and likely cell-type specific where nonspecific adsorption is protein-dependent and can be influenced by factors that includes local charge (Lloyd and Williams (1984) Non-specific adsorptive pinocytosis. Biochem. Soc. Trans. 12(3): 527-28). FcRn is the alpha chain of anon-covalent, heterodimeric complex with p2-microglobulin (P2m) that is chiefly localized within endosomal membranes under basal condition (Praetor, A., and Hunziker, W. (2002) beta(2)-Microglobulin is important for cell surface expression and pH- dependent IgG binding of human FcRn. J Cell Sci 115, 2389-2397; Simister, N. E., and Mostov, K. E. (1989) An Fc receptor structurally related to MHC class I antigens. Nature 337, 184-187; Antohe, et al. (2001) Expression of functionally active FcRn and the differentiated bidirectional transport of IgG in human placental endothelial cells. Hum Immunol 62. 93-105; Roberts, et al. (1990) Isolation and characterization of the Fc receptor from the fetal yolk sac of the rat. J Cell Biol 111, 1867-1876; Dickinson, et al. (1999) Bidirectional FcRn-dependent IgG transport in a polarized human intestinal epithelial cell line. J Clin Invest 104, 903-911; D'Hooghe. et al. (2017) Cell surface dynamics and cellular distribution of endogenous FcRn. PLoS One 12, eO 182695). FcRn exhibits increased affinity to its endogenous ligands albumin and immunoglobulin G (IgG) at acidic pH (Chaudhury, et al. (2003) The major histocompatibility complex-related Fc receptor for IgG (FcRn) binds albumin and prolongs its lifespan. J Exp Med 197, 315-322; Raghavan, et al. (1995) Analysis of the pH dependence of the neonatal Fc receptor / immunoglobulin G interaction using antibody and receptor variants. Biochemistry 34, 14649-14657; Andersen, et al. (2012) Structure-based mutagenesis reveals the albumin-binding site of the neonatal Fc receptor. Nat Commun 3, 610; Ober, et al. (2001) Differences in promiscuity7for antibody-FcRn interactions across species: implications for therapeutic antibodies. Int Immunol 13, 1551-1559; Martin, et al. (2001) Crystal structure at 2.8 A of an FcRn / heterodimeric Fc complex: mechanism of pH- dependent binding. Mol Cell 7, 867-877), and therefore selectively retains these proteins by way of intracellular trafficking away from lysosomal degradation and towards the plasma membrane to facilitate active recycling (Dickinson, et al. (1999) Bidirectional FcRn-dependent IgG transport in a polarized human intestinal epithelial cell line. J Clin Invest 104. 903-911; Ober, et al. (2004) Visualizing the site and dynamics of IgG salvage by the MHC class I-related receptor, FcRn. J Immunol 172, 2021-2029; Ober, et al. (2004) Exocytosis of IgG as mediated by the receptor, FcRn: an analysis at the single-molecule level. Proc NatlAcad Sci USA 101, 11076-11081; Kim, et al. (2004) Net absorption of IgG via FcRn-mediated transcytosis across rat alveolar epithelial cell monolayers. Am J Physiol Lung Cell Mol Physiol 287. L616-622; Schmidt, etal. (2017) Direct demonstration of a neonatal Fc receptor (FcRn)-driven endosomal sorting pathway for cellularrecycling of albumin. J Biol Chem 292, 13312-13322; Bern, et al. (2020) An engineered human albumin enhances half-life and transmucosal delivery when fused to protein-based biologies. Sci Transl Med 12; Strohl, W. R. (2015) Fusion Proteins for Half-Life Extension of Biologies as a Strategy to Make Biobetters. BioDrugs 29, 215-239). This salvage mechanism extends the serum half-lives of FcRn ligands relative to similarly sized proteins without FcRn-binding capabilities (Chaudhury, et al. (2003) The major histocompatibility complex-related Fc receptor for IgG (FcRn) binds albumin and prolongs its lifespan. J Exp Med 197, 315-322). In contrast, target- mediated CL processes are non-linear, and can dominate overall mAb CL depending on the dose (Ovacik, M., and Lin, K. (2018) Tutorial on Monoclonal Antibody Pharmacokinetics and Its Considerations in Early Development. Clin Transl Sci 11, 540-552; Ryman, J. T., and Meibohm, B. (2017) Pharmacokinetics of Monoclonal Antibodies. CPT Pharmacometrics Syst Pharmacol 6, 576-588; Huisinga, et al., Target-Driven Pharmacokinetics of Biotherapeutics, in Pharmaceutical Sciences Encyclopedia, pp 1-15, and Meno-Tetang, G. M. L. Target-Driven Pharmacokinetics of Biotherapeutics, in Pharmaceutical Sciences Encyclopedia, pp 1-12; and Peletier, L. A., and Gabrielsson, J. (2012) Dynamics of target-mediated drug disposition: characteristic profiles and parameter identification. J Pharmacokinet Pharmacodyn 39, 429-451). Target-mediated drug disposition (TMDD) is driven by high affinity mAb-target binding on the cell surface with subsequent internalization of the mAb-target complex resulting in intracellular catabolism (Ovacik, M.. and Lin, K. (2018) Tutorial on Monoclonal Antibody Pharmacokinetics and Its Considerations in Early Development. Clin Transl Sci 11, 540-552; Ryman, J. T., and Meibohm, B. (2017) Pharmacokinetics of Monoclonal Antibodies. CPT Pharmacometrics Syst Pharmacol 6, 576-588; Huisinga, et al., Target-Driven Pharmacokinetics of Biotherapeutics, in Pharmaceutical Sciences Encyclopedia, pp 1-15. and Meno-Tetang. G. M. L. Target-Driven Pharmacokinetics of Biotherapeutics, in Pharmaceutical Sciences Encyclopedia, pp 1-12; and Peletier, L. A., and Gabrielsson, J. (2012) Dynamics of target-mediated drug disposition: characteristic profiles and parameter identification. J Pharmacokinet Pharmacodyn 39, 429-451).

[0005] Variability in CLind of mAbs, even those with overlapping target specificity, can occur for several reasons. For example, altered FcRn binding due to differences in the physicochemical nature of the variable regions or other non-specific endocytic mechanisms can broadly impact the PK profile of candidate therapeutics (Kelly, et al. (2016) Target-independent variable region mediated effects on antibody clearance can be FcRn independent. MAbs 8, 1269- 1275; Piche-Nicholas, etal. (2018) Changes in complementarity-determining regions significantly alter IgG binding to the neonatal Fc receptor (FcRn) and pharmacokinetics. MAbs 10. 81-94; and Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transportand plasma half-life. iScience 25, 103746). Thus, a key goal within the field of large molecule PK is the ability to predict in vivo PK parameters of early preclinical mAbs through in vitro experimental measures to define in vitro to in vivo correlations. Recent methodologies to categorize mAh CLind have encompassed numerous biophysical interaction assays of both specific and nonspecific binding. For example, proxies for non-specific endocytosis that include heparin chromatography and baculovirus particle binding are commonly employed (Kraft, et al. (2020) Heparin chromatography as an in vitro predictor for antibody clearance rate through pinocytosis. MAbs 12, 1683432; Hotzel, et al. (2012) A strategy for risk mitigation of antibodies with fast clearance. MAbs 4, 753-760; Jain, et al. (2017) Biophysical properties of the clinical-stage antibody landscape. Proc Natl Acad Sci U SA 114, 944-949; and Datta-Mannan, et al. (2016) Aberrant bispecific antibody pharmacokinetics linked to liver sinusoidal endothelium clearance mechanism in cynomolgus monkeys. MAbs 8, 969-982). The evaluation of Fc-FcRn interactions is routinely accomplished using surface plasmon resonance and FcRn column chromatography (Schlothauer, et al. (2013) Analytical FcRn affinity chromatography for functional characterization of monoclonal antibodies. MAbs 5, 576-586; Vaughn, D. E., and Bjorkman, P. J. (1997) High-affinity binding of the neonatal Fc receptor to its IgG ligand requires receptor immobilization. Biochemistry 36, 9374-9380; and Borrok, et al. (2015) pH-dependent binding engineering reveals an FcRn affinity7threshold that governs IgG recycling. J Biol Chem 290, 4282-4290). The inference derived from the application of these techniques is that both non-specific uptake and FcRn affinity influence mAb CLind and should be considered for successful CLind predictions (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746).

[0006] FcRn function can also be directly interrogated using cells that either endogenously or ectopically express human FcRn (hFcRn) and human [32m (h[32m). and multiple reports have highlighted the utility of cell-based platforms to study FcRn trafficking dynamics and biology (Praetor, A., and Hunziker, W. (2002) beta(2)-Microglobulin is important for cell surface expression and pH-dependent IgG binding of human FcRn. J Cell Sci 115, 2389-2397; Antohe, et al. (2001) Expression of functionally active FcRn and the differentiated bidirectional transport of IgG in human placental endothelial cells. Hum Immunol 62, 93-105; Roberts, et al. (1990) Isolation and characterization of the Fc receptor from the fetal yolk sac of the rat. J Cell Biol 111, 1867- 1876; Dickinson, et al. (1999) Bidirectional FcRn-dependent IgG transport in a polarized human intestinal epithelial cell line. J Clin Invest 104, 903-911; Ober, et al. (2004) Visualizing the site and dynamics of IgG salvage by the MHC class 1-related receptor, FcRn. J Immunol 172, 2021- 2029; Ober, et al. (2004) Exocytosis of IgG as mediated by the receptor, FcRn: an analysis at thesingle-molecule level. Proc Natl Acad Set U S A 101, 11076-11081; Kim, et al. (2004) Net absorption of IgG via FcRn-mediated transcytosis across rat alveolar epithelial cell monolayers. Am J Physiol Lung Cell Mol Physiol 287, L616-622; Schmidt, et al. (2017) Direct demonstration of a neonatal Fc receptor (FcRn)-driven endosomal sorting pathway for cellular recycling of albumin. J Biol Chem 292, 13312-13322; Bern, et al. (2020) An engineered human albumin enhances half-life and transmucosal delivery when fused to protein-based biologies. Sci Transl Med 12; Grevys. et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746; Tzaban, et al. (2009) The recycling and transcytotic pathways for IgG transport by FcRn are distinct and display an inherent polarity. J Cell Biol 185, 673-684; Tesar, et al. (2006) Ligand valency affects transcytosis, recycling and intracellular trafficking mediated by the neonatal Fc receptor. Traffic 7, 1127-1142; Claypool, et al. (2004) Bidirectional transepithelial IgG transport by a strongly polarized basolateral membrane Fcgamma-receptor. Mol Biol Cell 15, 1746-1759; Goebl, et al. (2008) Neonatal Fc receptor mediates internalization of Fc in transfected human endothelial cells. Mol Biol Cell 19, 5490-5505; Sockolosky, et al. (2012) Engineering neonatal Fc receptor-mediated recycling and transcytosis in recombinant proteins by short terminal peptide extensions. Proc Natl AcadSci USA 109, 16095- 16100; Ying, et al. (2015) Engineered antibody domains with significantly increased transcytosis and half-life in macaques mediated by FcRn. MAbs 7, 922-930; Chung, et al. (2019) An in vitro FcRn- dependent transcytosis assay as a screening tool for predictive assessment of nonspecific clearance of antibody therapeutics in humans. MAbs 11, 942-955; Liu, et al. (2021) A cell-based FcRn-dependent recycling assay for predictive pharmacokinetic assessment of therapeutic antibodies. Bioanalysis 13, 1135-1144; Jaramillo, et al. (2017) Toward in vitro-to-in vivo translation of monoclonal antibody pharmacokinetics: Application of a neonatal Fc receptor- mediated transcytosis assay to understand the interplaying clearance mechanisms. MAbs 9. 781- 791; Grevys, et al.. (2018) A human endothelial cell-based recycling assay for screening of FcRn targeted molecules. Nat Commun 9, 621; and Gjolberg, et al. (2022) Biophysical differences in IgGl Fc-based therapeutics relate to their cellular handling, interaction with FcRn and plasma halflife. Commun Biol 5. 832). An elaborated in vitro experimental strategy can be used to assess the CLind of mAbs, and now several published studies have demonstrated the advantage of a cell-based approach to provide additional biological context not gleaned from standard biophysical methods (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746; Chung, et al. (2019) An in vitro FcRn- dependent transcytosis assay as a screening tool for predictive assessment of nonspecific clearance of antibody therapeutics in humans. MAbs 11, 942-955; Liu, et al. (2021) A cell-based FcRn-dependent recy cling assay for predictive pharmacokinetic assessment of therapeutic antibodies. Bioanalysis 13, 1083-1155; Jaramillo, et al. (2017) Toward in vitro-to-in vivo translation of monoclonal antibody pharmacokinetics: Application of a neonatal Fc receptor-mediated transcytosis assay to understand the interplaying clearance mechanisms. MAbs 9, 781-791; Grevys, et al. (2018) A human endothelial cell-based recycling assay for screening of FcRn targeted molecules. Nat Commun 9, 621; Gjolberg, et al. (2022) Biophysical differences in IgGl Fc-based therapeutics relate to their cellular handling, interaction with FcRn and plasma half-life. Commun Biol 5, 832; Chung, et al. (2022) Methods for Functional Characterization of FcRn Interactions with Therapeutic Antibodies and Fc-Fusion Proteins. Methods Mol Biol 2313, 295-303; Chung, et al. (2018) Development of a label-free FcRn-mediated transcytosis assay for in vitro characterization of FcRn interactions with therapeutic antibodies and Fc-fusion proteins. J Immunol Methods 462, 101-105). This includes direct measurements of cellular internalization, endosomal FcRn binding, and subsequent intracellular trafficking that has expanded our mechanistic understanding of processes driving the CLind of mAbs (Chung, et al. (2019) An in vitro FcRn- dependent transcytosis assay as a screening tool for predictive assessment of nonspecific clearance of antibody therapeutics in humans. MAbs 11, 942-955; Gjolberg, et al. (2022) Biophysical differences in IgGl Fc-based therapeutics relate to their cellular handling, interaction with FcRn and plasma half-life. Commun Biol 5, 832; Gurbaxani, et al. (2013) Are endosomal trafficking parameters better targets for improving mAb pharmacokinetics than FcRn binding affinity? Mol Immunol 56, 660-674; and Brinkhaus, et al. (2022) The Fab region of IgG impairs the internalization pathway of FcRn upon Fc engagement. Nat Commun 13, 6073).

[0007] Still, human CLind predictive capabilities for mAbs remain insufficient. Biophysical assays alone that are limited in scope to identify physicochemical determinants of mAb PK often fail to capture the biological framework and are often based on subjective assay cut points. Combining readouts from multiple techniques can provide a more robust interpretation and the ability to extend the findings to in vivo behavior (Kraft, et al. (2020) Heparin chromatography as an in vitro predictor for antibody clearance rate through pinocytosis. MAbs 12, 1683432). Cellbased methods can simultaneously collate multiple biological processes and physicochemical aspects into an endpoint that can facilitate a deeper understanding of mAb PK behavior. But limitations still exist. For example, one study demonstrated strong correlations between FcRn- mediated transcytosis and mAb CL whereby higher transcytosis was indicative of elevated CL in vivo. The underlying mechanistic rationale for this trend was unexplored, and the authors highlighted the inability’ of the assay to successfully infer the CL of mAbs with enhanced engineered FcRn affinity (Chung, et al. (2019) An in vitro FcRn- dependent transcytosis assay asa screening tool for predictive assessment of nonspecific clearance of antibody therapeutics in humans. MAbs 11. 942-955). A separate study demonstrated that the extent of FcRn-recy cling is heavily dependent on the amount of mAh internalized by specific or non-specific processes, and that the charge profile of the Fv regions are key contributors to the latter (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746). However, this work was limited in scope with the number of compounds examined.

[0008] The present disclosure is directed at overcoming this problem and other deficiencies in the art.SUMMARY OF VARIOUS EMBODIMENTS

[0009] A first aspect of the present disclosure is directed to a method of predicting in vivo non-target mediated clearance (CLind) of a biological molecule. This method involves providing a preparation of cells, where cells of the preparation do not express a target of the biological molecule. The method further involves incubating the preparation of cells with culture media containing the biological molecule under conditions mimicking in vivo physiological conditions and determining an amount of biological molecule taken up by cells of the preparation after incubation. The method further involves predicting in vivo non-target mediated clearance of the biological molecule based on said determining.

[0010] Another aspect of the present disclosure is directed to a method of predicting in vivo non-target mediated clearance of an FcRn interacting molecule. This method involves providing a first cell preparation, wherein cells of the first preparation do not express human neonatal Fc receptor (hFcRn), and providing a second cell preparation, wherein cells of the second preparation express a heterodimer of hFcRn and human beta-2-microglobulin (h|32m). Cells of the first and second cell preparation do not express a target of the FcRn interacting molecule. This method further involves subjecting the first and second cell preparations to first and second incubation periods, where the first incubation period comprises incubating the cell preparations with media containing the FcRn interacting molecule under acidic and / or non-acidic conditions. The second incubation period comprises incubating said cell preparations, after the first incubation, with media lacking the FcRn interacting molecule under non-acidic conditions. The method further involves determining an amount of the FcRn interacting molecule taken up by cells of the first and second preparations after said subjecting to the first incubation period and / or the second incubation period, and measuring an amount of the FcRn interacting molecule in media following the second incubation period. The method further involves quantifying non-specific endocytosis and FcRn recycling of the FcRn interacting molecule based on the determining and measuring steps. The invivo non-target mediated clearance of the FcRn interacting molecule is predicted based on the quantifying step.

[0011] Another aspect of the disclosure is directed to a method of predicting subcutaneous bioavailability of a biological molecule based on the extent of non-specific endocytosis of the biological molecule that is predicted based on the methods disclosed herein.

[0012] The in vivo clearance mechanisms of biological molecules encompass both target- mediated and target-independent processes. Two distinct determinants of biological molecule clearance separate from target-mediated influences are pH-dependent recycling by the FcRn and non-specific endocytosis, where the rates and extents of each vary between molecules. Approaches to quantify these dynamics have shown strong utility for establishing in vitro-in vivo correlations yet have been limited mechanistically or by analyte numbers.

[0013] Several biophysical techniques, including baculovirus particle binding and heparin chromatography, are ty pically utilized to study the potential for biological molecules like mAbs to undergo non-specific endocytosis. However, because these methods do not provide insight into the specific rates of cellular turnover, a reproducible and robust mammalian cell-based method of quantifying biological molecule non-specific endocytosis is described herein. The data generated with this cell-based method can be used to identify gross pharmacokinetic (PK) liabilities during preclinical drug development as well as inform mechanistic PK models for human translation.

[0014] Also disclosed herein is a functional cell-based FcRn recycling assay using mammalian cells. A series of pH-dependent internalization studies using a model antibody have confirmed proper function of the human FcRn complex in this assay. Furthermore, non-specific endocytosis was observed to be the predominant endocytic pathway of the tested antibody when no receptor was present. These cellular assays were applied to assess FcRn and non-specific interactions in a collection of clinical antibodies, multispecific antibodies, and Fc fusion proteins possessing a range of PK behaviors. The results demonstrate that non-specific endocytosis rates, pH-dependent non-specific interactions, and engagement with FcRn all contribute to the overall recycling efficiency of these therapeutic molecules.

[0015] Highlighting the predictive capacity of the assays described herein, all antibodies that possessed CLind in humans greater than 5 mL / kg / day were successfully identified by the assay. These results demonstrate that a combination of cellular assays can identify the individual mechanisms underlying the overall in vivo recycling efficiency and CLind of biological molecules.BRIEF DESCRIPTION OF THE FIGURES

[0016] Figures 1A-1C depict serum concentration-time profiles of ASA and an anti-IL- 4Ra mAb in wild type mice. Data were obtained from separate studies with different intravenous bolus doses for ASA (3 mg / kg) (Figure 1A) and the anti-IL-4Ra mAb (1 mg / kg) (Figure IB). Figure 1C shows the predicted values in Figures 1 A and IB were obtained via computational fitting using a two-compartment pharmacokinetic model with linear elimination from the central compartment (CL) and distribution between vascular and non-vascular compartments (CLD). Parameter estimates are reported in Table 1. C 1 and V 1 , serum concentration and volume of central compartment; C2 and V2, serum concentration and volume of non-vascular compartment; ASA, anti-streptavidin antibody.

[0017] Figures 2A-2D show the internalization kinetics of the anti-IL-4Ra mAb and ASA in CHO-K1 cells by flow cytometry and high content confocal microscopy. Following uptake studies, cells were washed, trypsinized, stained with a live / dead dye, fixed, permeabilized, stained with an anti -human Fc mouse IgG fluorescently conjugated with Alexa Fluor 647, washed, and analyzed via flow cytometry. A representative gating strategy is shown in Figure 2A. The median fluorescent intensities of single live cell events were obtained for each sample. CHO-K1 cells were incubated with 100 pg / mL of the anti-IL-4Ra mAb or ASA for increasing periods of time at 37°C or 4°C. Biphasic time-dependent (Figure 2B) and linear concentration-dependent (Figure 2C) internalization was observed for the anti-IL-4Ra mAb. but only under 37°C conditions. In contrast, ASA endocytosis was negligible at 37°C. Minimal cell surface binding was detected for the anti- IL-4Ra mAb and ASA at 4°C relative to untreated controls. Figure 2D is a panel of confocal microscopy images of CHO-K1 cells following incubation and similar post-experimental processing with the anti-IL-4Ra mAb (Figure 2D, right image) or ASA (Figure 2D, middle image). Untreated cells are shown in Figure 2D, left panel. Cells were stained with CellMask Blue and Hoechst to visualize whole cells (cytoplasm and nucleus) and fluorescently labeled anti-Fc antibody (red). These images demonstrate the utility of confocal microscopy as an alternative to flow cytometry for high throughput quantitation of therapeutic protein internalization kinetics. ASA, anti-streptavidin antibody: CHO-K1, Chinese Hamster Ovary cells.

[0018] Figures 3A-3D provide an overview of quantitative endocytosis assay development using flow cytometry. Figure 3A shows histograms of the five bead populations overlaid when stained with either 5 or 10 pg / mL of the anti -human Fc mouse IgGl conjugated with Alexa Fluor 647. Also shown are plots of the median fluorescent intensities versus the antibody binding capacities (ABCs) indicating both staining conditions resulted in linear curves with r2values of 0.999. Figure 3B is a graph showing assay reproducibility that was evaluated by conducting uptakestudies in CH0-K1 cells using 100 pg / inL of anti-IL-4Ra mAb or ASA incubated at 37°C for 60 min over several different experimental days. Mean anti-IL-4Ra mAb ABC: 40,839 (SD 6614, CV 1 .2%); mean ASA ABC: 921 (SD 315, CV 34.2%). N = 20 per sample group. Figure 3C is a representative histogram of single ASA and untreated samples with a gate incorporating 99% of the untreated population, corresponding to the ‘negative’ signal. The ASA endocytosis at the experimental settings was so low that -92% of the ASA signal resided within the ‘negative’ gate. Figure 3D is an illustrative overview of the quantitative endocytosis methodology. ASA, antistreptavidin antibody; IgGl, immunoglobulin Gl; CHO-K1, Chinese Hamster Ovary cells; SD, standard deviation; CV, coefficient of variation.

[0019] Figures 4A-4C illustrate the surface charge distribution on the anti-IL-4Ra mAb (Figure 4A) and ASA (Figure 4B) and demonstrate that removing exposed positive charge can reduce non-specific endocytosis. Figures 4A and 4B include Fv model ribbon representations (heavy chain in dark teal, light chain in pale teal) depicting heavy chain CDRs in orange (CDR1, CDR2) and red (CDR3). Light chain CDRs are shown in purple. Hydrophobic patches are shown in green. Negative and positive patches are red and blue, respectively. Sites of targeted mutation are indicated. The anti-IL-4Ra mAb (Figure 4 A) had more positive patches than ASA (Figure 4B), with 2 positive patches in heavy chain CDRs. Various point mutations were performed on the anti-IL-4Ra mAb CDRs and Fv to reduce the surface positive charge within the identified charge patches. The graph of Figure 4C shows substantial reductions in non-specific endocytosis were measured for all anti-IL-4Ra mAb mutants. CDR, complementarity-determining region; WT, wild type anti-IL-4Ra mAb; LC, light chain; CDR Hl, the first complementarity-determining region of the heavy7chain; CDR H3, the third complementarity-determining region of the heavy chain.

[0020] Figure 5 is a panel of serum concentration-time profiles of a second set of preclinical antibodies that all bind the same target (mAbs B1-B5) following single intravenous bolus dose in wild type mice. Data was fit to mean serum concentrations using the two- compartment computational model depicted graphically in Figure 1.

[0021] Figures 6A-6D demonstrate that non-specific endocytosis is conserved across different species and cell types. Uptake studies were conducted in either CHO-K1 (Cricetulus griseus, ovarian epithelial-like cells; Figure 6A) or Vero cells (Cercopithecus aeihiops. kidney epithelial cells; Figure 6B) using the mAb panel of Figure 5 possessing vary ing CL in wild type mice. A range of uptake extents was observed that was independent of pl. These results also demonstrate the superior sensitivity of the endocytosis assay performed at 37°C when compared to cell surface binding alone (i.e., 4°C groups). Non-specific endocytosis into either CHO-K1 orVero cells strongly correlated to CL as shown in Figure 6D, indicating this attribute as an important factor of CLind for these mAbs. Additionally, results in Vero and CH0-K1 cells strongly corresponded with each other (Figure 6C), supporting non-specific endocytosis as the mechanism for mAb internalization. All data plotted as mean ± SD, n = 3-4 per group. CL, clearance; CHO- Kl. Chinese Hamster Ovary cells; CLind, target-independent clearance.

[0022] Figures 7A-7C show co-expression of hFcRn-GFP and h(32m in stably transfected MDCK II cells. Maximum projections of confocal images following anti-FcRn immunofluorescent staining confirmed hFcRn-GFP (green, yellow) expression in sorted, hFcRn- GFP / hp2m MDCK II cells (Figure 7A). Cells were stained with a rabbit anti-FcRn antibody using a goat anti-rabbit secondary antibody conjugated with Alexa Fluor 594 (y ellow). Parental MDCK 11 displayed only positive nuclear staining (Hoechst, blue). Top (i) and bottom (ii) images of Figure 7A are with and without the GFP channel depicted, respectively. Scale bars, 20 pm. Figure 7B shows co-staining of parental and hFcRn-GFP / hp2m MDCK II cells with targeted anti-hFcRn and anti-hp2m antibodies or corresponding isotype controls demonstrating robust transgene coexpression in transfected, post-sorted hFcRn-GFP / hp2m MDCK II cells. Cells were stained either with (total) or without (cell surface) fixation / permeation. A population of GFP-hFcRn-GFP / hp2m MDCK II cells were consistently observed both visually (as seen in Figure 7A) and via flow cytometry' under maintained selective pressure. It constituted less than 5% of the total population as determined by flow cytometry’. Therefore, all subsequent analyses with fluorescence detection, including those in the current figure, were done on GFP+ hFcRn-GFP / hp2m MDCK II cells. The median fluorescent intensities of the hFcRn and hp2m signals from hFcRn-GFP / hp2m MDCK II cells in Figure 7B indicated the majority' of hFcRn-GFP and hp2m w as intracellular as shown in the graph of Figure 7C. N = 3, mean ± SD. GFP, green fluorescent protein; BF, brightfield; hFcRn, human neonatal Fc receptor; hp2m, human P2-microglobulin; MDCK II Madin-Darby canine kidney subclone II cells.

[0023] Figures 8A-8C demonstrate the internalization kinetics of ASAwr-DL650 and HSA-DL650 in hFcRn-GFP / hp2m and parental MDCK II cells. Time-dependent endocytosis studies were performed using 10 or 100 pg / mL ASAwr-DL650 at pH 5.8. 7.4, or 8.0 to confirm the function of hFcRn-GFP / hp2m (Figure 8A). Representative histograms are shown for the 100 pg / mL AS AWT-DL650 time-dependent uptake results at pH 5.8 (Figure 8A, top panel). The hFcRn- GFP / hp2m MDCK II cells exhibited increased endocytosis of ASAWT-DL650 with decreasing pH that was time- and concentration-dependent (Figure 8A, bottom left graph). No differences were observed for ASAwr-DL650 endocytic rates within parental MDCK II cells between the tested pH conditions (Figure 8 A, bottom right graph). FcRn-mediated, concentration-dependent endocytosisstudies with ASAWT-DL650 (Figure 8B) or HSA-DL650 (Figure 8C) were carried out in FcRn- GFP / [32m MDCK II cells for 20 min at pH 5.8 with or without 50 mg / mL unlabeled analyte as a competitive inhibitor. Specific, internalized data expressed as a per minute rate were fit to the Michaelis-Menten equation to provide an estimate for the ASAWT-DL650 and HSA-DL650 concentrations achieving half-maximal uptake velocity (Km) and maximum velocity (Vmax) within the respective experimental systems (see Figure 8B (ii) and Figure 8C (ii)). Inset plots (hi) at the bottom of Figures 8B and 8C are the results from the studies performed in the presence of excess unlabeled protein (Figure 8B (i) and Figure 8C (i)), which demonstrated linear concentrationdependent uptake. Data was fit using linear regression to support the observation of non-specific endocytosis in the absence of receptor-mediated internalization. N = 3-4 per point, mean ± SD. MeFl, median fluorescent intensity; HSA-DL650. human serum albumin fluorescently conjugated with Dy Light 650.

[0024] Figures 9A-9C show intracellular localization of hFcRn-GFP in hFcRn- GFP / h|32m-MDCK II cells following time-dependent incubations with ASAwr-DL650. Human FcRn-GFP / hp2m-MDCK II cells were loaded with a 10 kDa Texas Red dextran (TR-dextran) the day prior to uptake studies to label lysosomes. On the day of imaging, cells were incubated with 10 or 100 pg / rnL ASAWT-DL650 at pH 5.8 or 7.4 at the noted time points. Figure 9 A shows a panel of representative images of hFcRn-GFP / hp2m-MDCK II cells treated with 100 pg / mL ASAWT-DL650 at pH 5.8. Hoechst was used as a nuclear stain. Arrowheads indicate GFP' cells that lacked hFcRn-GFP. They provide further confirmation of proper transgene function because the GFP" populations were consistently negative for ASAWT-DL650 at the tested conditions. Mean fluorescent intensities (MFIs) of intracellular AS AWT-DL650 were obtained for GFP+ cells (Figure 9B), which demonstrated time-, concentration-, and pH-dependence that were consistent with observations using flow cytometry. **** p < 0.0001 indicates treatment vs. untreated control MFI following two-way AN OVA with Tukey’s multiple comparisons. N = 3 per group, mean ± SD. Figure 9C is a Pearson correlation coefficient analysis for hFcRn-GFP, ASAWT-DL650, and lysosomal intracellular signals which resulted in positive colocalization only between hFcRn-GFP and ASAWT-DL650 over the duration of the study. *. ***, **** p < 0.05, < 0.001, < 0.0001 following a one-way ANOVA performed on the hFcRn vs. ASAWT group, with Dunnetf s multiple comparisons relative to the 15 min time point. N = 3 per group, mean ± SD.

[0025] Figures 10A-10E depict the development of the cell-based hFcRn recycling assay as described and claimed herein. Figure 10A is a schematic depiction of the experimental workflow. Parental MDCK II cells were included to provide an assessment of non-specific mAb interactions. Because negligible intracellular trafficking could occur under 4°C conditions, the 4hr recycling phase at this temperature provided confirmation of an active process in the 37°C group in addition to a measurement of the total amount of mAh internalized during the load phase. MDCK II hFcRn recycling studies were performed on ASAWT or a mutant with deleted hFcRn affinity7(AS AAAA). Human FcRn-mediated uptake and recycling was found to be higher for ASAWT at pH 5.8 compared with pH 7.4, essentially zero for the ASAAAA mAh (parental MDCK II no different than hFcRn-GFP / hp2m-MDCK II), and elevated with 37°C incubations compared to 4°C, which supported active cellular trafficking (Figure 10B, left graph). Furthermore, non-specific endocytosis was determined to be negligible for all tested ASA mAbs based on minimal internalization in parental MDCK II at either pH value (Figure 10B, right graph). These findings indicated that a mAb with low non-specific uptake at pH 7.4 (i.e. low endocytosis in parental MDCK II) will exhibit very low hFcRn-mediated recycling because the fraction recycled will be dictated by the initial amount internalized. Therefore, mAbs with favorable PK behavior would be challenging to evaluate solely using pH 7.4 load conditions. Incubations at pH 5.8 provides a means to load enough mAb intracellularly to elucidate hFcRn recycling efficiency. Significance of differences in hFcRn-GFP / h[32m-MDCK II cells were measured by two-way ANOVA with Tukey’s multiple comparisons test, **** signifying adjusted p < 0.0001. Bar graphs represent mean ± SD, n= 6-8 replicates per group. Figures 10C and 10D are graphs showing serum concentration-time profiles for ASAWT and ASAAAA following intravenous administration to Tg32 (Figure 10C) or Tg276 (Figure 10D) hFcRn transgenic mouse models to evaluate their disposition. ASAAAA possessed rapid CL consistent with its lack of hFcRn binding. Mean ± SD, n = 3 per time point and group. The hFcRn recycling score using results from the pH 5.8 load phases for ASAWT was significantly higher when compared to ASAAAA as shown in the graph of Figure 10E. The scoring results are consistent with the in vivo CL determined in Figures 10C and 10D. *** p < 0.001 unpaired t-test, mean ± SD. Dashed green, yellow, and red lines indicate a recycling score 1 0%, 50% or 25% of ASAWT, respectively.

[0026] Figures 11 A-l ID shows poor hFcRn recycling scores correspond to high CLind in humans. Figure 11A shows a representative dataset of mAbl from hFcRn recycling studies in transfected and parental MDCK II cells that encompasses all conditions used to derive hFcRn recycling scores. The amount internalized during the load phase was obtained via the residual 4°C group as negligible intracellular trafficking would be expected at this temperature. The mAbl reference exhibited pH-dependent uptake in hFcRn-GFP / hp2m that aligned with the extent recycled at the matching pH values. Furthermore, low non-specific uptake was measured in parental MDCK II cells. The hFcRn recycling scores (FREMS) were obtained for 8 mAbs and plotted against their human CLind values as shown in Figure 1 IB. A low recycling score indicatedhigh non-specific uptake and / or inefficient hFcRn-mediated recycling, with values below 0.25 strongly indicative of rapid CLind in vivo. The pH 5.8 load conditions were used for scoring as the amount internalized for the mAbs with CLind below 5 mL / d / kg was very low at pH 7.4. MAbl served as the reference mAb for the full mAh panel comparison. *, **, *** p < 0.05, 0.01, 0.001 following one-way ANOVA with Dunnetf s multiple comparisons conducted on FcRn recycling scores versus mAbl. Dashed green, yellow, and red lines indicate a recycling score 100%. 50% or 25% of mAb 1 , respectively. Load phase results from parental MDCK II cells at 4°C (indicative of internalized amount) demonstrated significantly higher anti-IL-4Ra mAb uptake at pH 7.4 when compared to mAbl (Figure 11C) (p < 0.0001, one-way ANOVA with Dunnetf s multiple comparisons). At pH 5.8, both anti-IL-4Ra mAb (fttt, P < 0.0001) and mAb9 (fff, P < 0.001) exhibited significantly higher non-specific uptake relative to mAbl (one-way ANOVA with Dunnetf s multiple comparisons). MAb9 and mAbl also displayed significant rises in non-specific uptake at pH 5.8 when compared to pH 7.4 (*, **** p < 0.05, 0.0001 following multiple t-tests). Non-specific endocytosis of the mAb panel was measured by incubating CHO-K1 cells for 60 min at 37°C with 100 pg / mL mAbs at either pH 5.8 or 7.4, followed by flow cytometry on fixed and permeabilized cells with anti-human Fc detection (Figure HD). Only anti -IL-4 Ra mAb (p < 0.0001) and mAb9 (p < 0.01) exhibited significantly higher non-specific uptake at pH 7.4 when compared to ASAwr (ordinary one-way ANOVA with Dunnetfs multiple comparisons test). Additionally, the extent of anti-IL-4Ra mAb internalization at pH 7.4 was dramatically higher than that of mAb9. At pH 5.8, all mAbs exhibited significantly higher non-specific endocytosis into CHO-K1 cells when compared to pH 7.4 conditions, with mAb9 displaying the largest change (multiple unpaired t-tests, p at least less than 0.05). Taken together, these results demonstrate that the tested mAbs had a significantly increased propensity for non-specific behavior at pH 5.8 relative to pH 7.4.

[0027] Figures 12A-12B provide an illustrative summary. Figure 12A is an image depicting cellular processes for a low CLind mAb. Small amounts of mAb enter a non-targeted vascular endothelial cell at pH 7.4 via non-specific endocytosis due to the lack of high non-specific interactions arising from physiochemical traits of the mAb such as localized charge patches. Endosomes containing internalized mAb will undergo gradual acidification during trafficking. As shown herein all tested mAbs exhibited increased non-specific interactions at an acidic pH relative to pH 7.4. However, these unspecific behaviors likely have a minimal impact on the mAb interaction with hFcRn for low CLind compounds, resulting in high percentages of endosomal hFcRn binding. Bound mAb will be subsequently trafficked to the plasma membrane and dissociate from hFcRn at near-neutral pH. The result is efficient hFcRn recycling, observed as ahigh hFcRn recycling score in the current work. Examples of low7CLind mAbs from this study include mAbl and ASAWT. Figure 12B is an image depicting how multiple cellular mechanisms can drive high mAh CLind. Some mAbs, such as the anti-IL-4Ra mAb, exhibit high rates of nonspecific endocytosis due to adverse physiochemical characteristics including strongly charged local patches at neutral pH. This leads to relatively large amounts of mAh internalization into nontarget cell populations via non-specific endocytosis. Upon endosomal acidification, an elevation of non-specificity may occur. This pH-dependent shift could be dramatic, which was observed in the current study for mAb9. These unspecific aspects can impair binding with hFcRn as w ell as potentially result in other non-specific interactions, such as with endosomal membrane components. The consequences are (1) more total mAb within an endosome per hFcRn receptor, (2) dysfunctional hFcRn binding arising from altered charge states, and / or (3) elevated non-specific interactions which act to offset hFcRn-mAb engagement. Additionally, mAb dissociation from hFcRn can become impaired at the cell surface if its affinity to the receptor is too high at pH 7.4. These factors can all contribute to higher rates of intracellular catabolism and thus high CLind due to a smaller fraction of internalized mAb undergoing hFcRn-mediated recycling. Experimentally, this was resolved in the studies described herein where low hFcRn recycling scores were indicative of one or more of these processes occurring. The tested mAbs not displaying large negative behaviors (e.g. mAb8) were impacted by vary ing degrees of non-specific behavior where the amount of hFcRn interactions was negatively counterbalanced by the amount of non-specificity in relation to the reference mAbl. Image created with BioRender.com

[0028] Figure 13A shows gating scheme for all endocytosis studies with sequential subpopulation order indicated by underlined numbers (upper left comer of each plot). Median fluorescent intensities were measured within single, live cell events for parental (GFP ) or hFcRn- GFP / hp2m (GFP ) MDCK II cells. Histograms from two separate samples were overlaid for plot 4 to demonstrate the differences in GFP expression. At the bottom of Figure 13A are two representative histograms from the separate MDCK II cell ty pes for 100 pg / mL ASAWT-DL650 time-dependent uptake results at pH 5.8. In Figure 13B, uptake results with pH 5.8 conditions were removed to allow visualization of the results at pH 7.4.

[0029] Figures 14A-14B show recycled (Figure 14A) and residual (Figure 14B) amounts of the mAb panel tested in the hFcRn recycling studies in transfected and parental MDCK II cells that encompass all conditions used to derive hFcRn recycling scores. The 37°C and 4°C groups apply to the respective temperatures of the assay plates during the recycling phase of the studies. The amount internalized during the load phase was obtained via the residual 4°C group as negligible intracellular trafficking would be expected at this temperature. The amount internalized(uptake) and recycled as well as residual concentrations after the pH 5.8 load phase were used to calculate non-specific uptake coefficient (NUC) and FcRn recycling efficiency metric (FREM) scores, respectively.

[0030] Figures 15A-15B are the measurements of non-specific endocytosis in CH0-K1 cells for the expanded anti-IL-4Ra mAb panel with single- and double-point mutants at pH 7.4 (Figure 15A) and 5.8 (Figure 15B). The non-specific endocytosis of all the WT and YTE mutants were significantly different when compared to control anti-IL-4Ra wild type mAb at either pH value (ordinary one-way ANOVA with Dunnett’s multiple comparisons test; #: P < 0.001 and *: P < 0.0001). Each bar graph represents mean ± SD, N = 3-4 per group.

[0031] Figures 16A-16D demonstrate hFcRn-mediated, concentration-dependent endocytosis studies with ASAwr-hIgG2-DL650 (Figure 16A-16B) or ASAwr-hIgG2-YTE-DL650 (Figure 16C-16D), which were carried out in hFcRn-GFP / hp2m MDCK II cells for 30 min at pH 5.8. Non-specific endocytosis measurements in parental MDCK II cells were subtracted from hFcRn specific internalized data, with the resultant values expressed as a per minute rate. These data were fit to the Michaelis-Menten equation to provide an estimate for the ASAwr-hIgG2- DL650 (Figure 16B) and ASAwr-hIgG2-YTE-DL650 (Figure 16D) concentrations achieving half- maximal uptake velocity (Km) and maximum velocity (Vmax) within the respective experimental systems. N = 3-4 per point, mean ± SD. MeFI, median fluorescent intensity; ASAwr-hIgG2- DL650, anti-streptavidin human immunoglobulin G2 antibody conjugated with DyLight 650; ASAwT-hIgG2-YTE-DL650, anti-streptavidin human immunoglobulin G2 YTE antibody conjugated with DyLight 650.

[0032] Figures 17A-17C exhibit the ability of the cell-based hFcRn recycling assay to successfully and simultaneously rank-order mAbs with wild type or engineered Fc regions. Figure 17A shows the hFcRn efficiency metric (FREM) scores of the anti-IL-4Ra WT and YTE charge mutant mAbs. The scores were normalized to anti-IL-4Ra mAb-EEES-YTE. The assay was optimized to simultaneously compare both Fc-engineered YTE and wild type Fc (i.e. hIgG2) charge mutant mAbs. The half-life extending YTE mAbs exhibited better FREM scores relative to their WT counterparts. Each bar graph represents mean ± SD. and N= 6-8 replicates per group, except for anti-IL-4Ra mAb-EEES-YTE, where N = 16. YTE mutant mAbs were tested in two batches in two different days, and FREM scores from each day were normalized to the EEES-YTE mutant mAb from that day as the relative control. An ordinary one-way ANOVA with Dunnet’s multiple comparisons test was performed between the groups and significance was compared to the anti-lL-4Ra mAb as a control (1: P < 0.05 and *: P < 0.0001). Figure 17B shows the serum concentration time profiles of the tested anti-IL-4Ra mAb WT and YTE charge mutants in malehomozy gous immunodeficient SCID hFcRn Tg32 transgenic mice (N = 3 per group). Figure 17C shows the comparison between CLind and FREM scores. It was observed that mAbs with relatively lower FREM scores demonstrated higher CLind.

[0033] Figures 18A-18F shows the recycled, uptake, and residual amounts of the anti-IL- 4Ra WT mAh mutants in the hFcRn recycling studies in hFcRn-GFP / p2M transfected (Figures 18A-18C) and parental MDCK II cells (Figures 18D-18F) that were used to derive the FREM scores. The load and recycled phases of the mAbs were tested at 37°C. The amount internalized (uptake) and recycled as well as residual concentrations after the pH 5.8 load phase were used to calculate non-specific uptake coefficients (NUC) and FcRn recycling efficiency metric (FREM) scores, respectively.

[0034] Figures 19A-19F shows the recycled, uptake, and residual amounts of the anti-IL- 4Ra YTE mAb mutants in the hFcRn recycling studies in hFcRn-GFP / |32M transfected (Figures 19A-19C) and parental MDCK II cells (Figures 19D-19F) of EEES-YTE, SSLS-YTE, and WT- YTE mAb mutants that were used to derive FREM scores. All experimental data was collected in one day. The experiment was conducted on a separate day after thawing different cell vials compared to WT mutants in Figure 18. The load and recycled phases of the mAbs were tested at 37°C. The amount internalized (uptake) and recycled as well as residual concentrations after the pH 5.8 load phase were used to calculate non-specific uptake coefficient (NUC) and FcRn recycling efficiency metric (FREM) scores, respectively.

[0035] Figures 20A-20F demonstrate the recycled, uptake, and residual amounts of the anti-IL-4Ra YTE mAb mutants in the hFcRn recycling studies in hFcRn-GFP / p2M transfected (A- C) and parental MDCK II cells (D-F) of four YTE mutants along with EEES-YTE as a control mAb that were used to derive FREM scores. Figure 20 experimental data was collected on a separate day from Figure 19. This experiment was conducted along with the WT mutants mentioned in Figure 18 from the same cell flasks. The load and recycled phases of the mAbs were tested at 37°C. The amount internalized (uptake) and recycled as well as residual concentrations after the pH 5.8 load phase were used to calculate non-specific uptake coefficient (NUC) and FcRn recycling efficiency metric (FREM) scores, respectively.

[0036] Figure 21 A is a graph showing the non-specific endocytosis results from the CHO- K1 uptake assay for a panel of fully human and humanized clinical mAbs or derivatives thereof with known clinical PK data. The mAbs were grouped by possessing either: CLind below 4.5 mL / kg / d in human following an intravenous (IV) dose, CLind above 4.5 mL / kg / d, or CLind estimate unobtained (e.g. dose level not high enough to achieve target saturation). This was done because the currently disclosed assays cannot provide information on mAb target-mediated dynamics. The4.5 mL / kg / d value was chosen because this corresponds to a terminal serum half-life of approximately 10 days in an 80 kg human. Also shown on the graph are generated assay thresholds. The ‘'Low-risk” group contains mAbs with low-risk of high CLind and / or low FSQ (due to non-specific endocytosis). It was defined as the upper bound of the 95% confidence interval of the mean ABC value for the mAbs with FSQ above 50% and / or CLind < 4.5 mL / kg / d. The “Mediumrisk” group contains mAbs that may or may not exhibit high CLind and / or low FSQ and was set as two standard deviations above the mean ABC mentioned directly above. The '‘High-risk” group highlights mAbs with obvious non-specific endocytosis that will likely be detrimental to their disposition in human. All mAbs with CLind estimates from Figure 21 A were binned into their respective risk categories as described herein and plotted in Figure 21B. Gradual elevations in mean CLind values were correlated with an increased risk classification in the CHO-K1 assay. Additionally, mAbs that may have had low CLind but still displayed high non-specific CHO-K1 endocytosis (i.e., high ABC value) were found to have low FSQ in humans (e.g. mAb 28). Thus, no mAb in the high-risk category had favorable PK in human.

[0037] Figures 22A-22B highlight how non-specific endocytosis can identify high-risk multispecific antibodies with increased target-independent clearance. Forty-eight multispecific antibodies were obtained that exhibited a broad range of target-independent clearance (CLind) in Tg32 human FcRn transgenic mice (0.2 to 2000 mL / kg / hr) (Figure 22A). The proteins were derived from various molecular formats. Non-specific endocytosis was measured for the multispecific protein panel by obtaining antibody binding capacities (ABCs) in CHO-K1 cells. The ABCs were then used to bin compounds into low, medium, or high-risk categories (Figure 22B) using the same parameters as for the human mAb panel in Figure 15 (i.e., low in the current example defined as having the same ABC value as in the human mAb panel depicted in Figure 21A). Doing so identified medium and high-risk multispecifics as having an increased propensity for elevated CLind in Tg32 mice where 94% of proteins in the high-risk and 92% of proteins in the medium-risk bins possessed CLind in Tg32 mice above 0.5 mL / kg / hr. However, 63% of low-risk multispecifics exhibited CLind above 0.5 mL / kg / hr in Tg32 mice, which highlights alternative clearance mechanisms aside from non-specific endocytosis. This indicates the need for supportive assays of alternative mechanisms, such as the cellular human FcRn recycling assay defined in the current work.

[0038] Figures 23A-23F demonstrate how non-specific endocytosis is indicative of elevated target-independent clearance across a broad range of protein structures. A panel of Fc- fusion proteins containing five distinct protein structures unrelated to that of human immunoglobulins was generated. Each construct possessed two individual proteins fused to ahuman Fc domain. Single dose studies (2 mg / kg) in wild type mice demonstrated a large range of CLind (0.81 to 1260 mL / hr / kg) as shown in Figure 23 A. All Fc-fusion proteins were analyzed using the CHO-K1 cellular endocytosis method. Antibody binding capacities (ABCs) were obtained at either pH 7.4 (Figure 23B) or pH 5.8 (Figure 23C) and plotted against their corresponding CLind in wild type mice. Average ABC values from CHO-K1 non-specific endocytosis studies conducted at pH 7.4 were then placed into low, medium, or high-risk bins using the ABC cutoff points derived from the human mAb panel depicted in Figure 21 (Figure 23D). Ninety -four percent of Fc-fusion proteins within the high-risk bin possessed CLind in wild type mice above 3 mL / kg / hr. Fc-fusion proteins deemed low or medium-risk following pH 7.4 endocytosis experiments were placed into new risk thresholds based on their pH 5.8 ABC values (Figure 23E). It was found that 3 proteins exhibited extensive jumps in pH-dependent, non-specific behavior: Fc-Fusion 7. 8, and 10. Of these, two out of three possessed very high CLind in mice (Figure 23F), supporting high pH- dependent non-specificity as an unfavorable attribute for therapeutic proteins.DETAILED DESCRIPTION OF THE VARIOUS EMBODIMENTS

[0039] The present disclosure is directed to in vitro cell-based methods useful for predicting in vivo pharmacokinetic properties of candidate therapeutic biological molecules. In particular, the cell-based assays described herein facilitate prediction or estimation of in vivo nontarget mediated clearance of candidate therapeutic molecules. The cell-based assays described herein can also be utilized to rank order the in vivo non-target mediated clearance of candidate molecules to support lead candidate selection and optimization while reducing the number of molecules tested in animal models.

[0040] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.

[0041] Unless otherwise defined herein, scientific and technical terms used in connection with the present disclosure shall have the meanings that are commonly understood by those of ordinary skill in the art. Generally, nomenclatures used in connection with, and techniques of, cell culture, molecular biology, genetics and protein and nucleic acid chemistry and hybridization described herein are those well-known and commonly used in the art. The methods and techniques of the present disclosure are generally performed according to conventional methods well known in the art and as described in various general and more specific references that are cited and discussed throughout the present specification unless otherwise indicated. See, e.g., Sambrook et al. Molecular Cloning: A Laboratory Manual, 2d ed., Cold Spring Harbor Laboratory Press, ColdSpring Harbor. N.Y. (1989) and Ausubel et al.. Current Protocols in Molecular Biology, Greene Publishing Associates (1992), and Harlow and Lane Antibodies: A Laboratory Manual Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. (1990), which are incorporated herein by reference.

[0042] In this disclosure, the use of the singular terms includes pluralities and plural terms include the singular unless specifically stated otherwise. As used herein, the singular forms “a”, “an”, and “the”' include both singular and plural referents unless the context clearly dictates otherwise.

[0043] In this disclosure, the use of “or” means “and / or” unless stated otherwise.Furthermore, the use of the term “including”, as well as other forms, such as “includes” and “included”, is not limiting. Also, terms such as “component” encompass both components comprising one unit and components that comprise more than one subunit unless specifically stated otherwise.

[0044] The terms “comprising”, “comprises” and “comprised of as used herein are synonymous with 'including', ‘includes', ‘containing', or ‘contains', are inclusive or open-ended, and do not exclude additional, non-recited members, compounds, products, elements, or method steps. The expression “consisting essentially of used in the context of a composition or method (e.g, “a method consisting essentially of) means that additional elements, ingredients, or steps may be present but only to the extent that such additional elements, ingredients, or steps do not change / alter the characteristic / activity / functionality of said product, composition, or method. The expression “consisting of used in the context of a composition or method means that the referenced composition or method includes only the elements, steps, or ingredients specifically recited in the particular embodiment or claim.

[0045] In embodiments or claims where the term “comprising” is used as the transition phrase, such embodiments and claims can also be envisioned with replacement of the term “comprising” with the terms “consisting of or “consisting essentially of.

[0046] The recitation of numerical ranges by endpoints includes all numbers and fractions subsumed within the respective ranges, as well as the recited endpoints.

[0047] The term “about” as used herein when referring to a measurable value such as a parameter, an amount, a temporal duration, and the like, is meant to encompass variations of + / - 10% or less, preferably + / -5% or + / -!% from the specified value, insofar such variations are appropriate to perform in the disclosed embodiment. It is to be understood that the value to which the modifier “about” refers is itself also specifically, and preferably, disclosed.

[0048] A first aspect of the present disclosure is directed to a method of predicting in vivo non-target mediated clearance of a biological molecule. This method involves providing a preparation of cells, where cells of the preparation do not express a target of the biological molecule. The method further involves incubating the preparation of cells with culture media containing the biological molecule under conditions mimicking in vivo physiological conditions and determining an amount of the biological molecule taken up by cells of the preparation after incubating. The method further involves predicting in vivo non-target mediated clearance of the biological molecule based on said determining.

[0049] As referred to herein “clearance” or “CL” is defined as the volume of plasma cleared of a drug over a specified period of time. Therefore, the unit of measure for drug clearance is volume / time. Clearance is equal to the rate at which a drug, e.g.. a biological molecule such as a therapeutic protein, is removed from plasma (mg / mL) divided by the concentration of that drug in the plasma (mg / mL). For therapeutic biological molecules, such as antibodies, antigen-binding antibody fragments, antibody derivatives, multispecific engineered antibodies and proteins, and fusion proteins, clearance is determined by target-dependent and non-target-dependent pathways.

[0050] Target-dependent clearance, which occurs as a result of the interaction of the biological molecule with its target, is dependent on the rate in which the therapeutic biological molecule is internalized, the binding target (e.g., antigen) density, binding affinity, and turnover kinetics of the binding target. Non-target-dependent or target independent clearance (CLind) of a biological molecule is mediated by non-specific cellular uptake of the biological molecule via adsorptive endocytosis and / or fluid-phase endocytosis. Once inside the endosomal compartment of the cell, two competing processes, z.e., intracellular lysosomal catabolism and neonatal Fc receptor (FcRn) mediated recycling or transcytosis, dictate clearance of the biological molecule. From a drug development perspective, a candidate therapeutic exhibiting high levels of nonspecific endocytosis and / or poor FcRn capture and recycling will have increased non-target dependent clearance. Additionally, as described herein, high non-specific endocytosis has been identified as a key parameter in predicting low subcutaneous bioavailability. Therefore, the cellbased assays described herein are quantitative tools useful for assessing non-specific endocytosis and FcRn recycling efficiency to identify pharmacokinetic liabilities, including increased non- target dependent clearance and low subcutaneous bioavailability', early in the drug development process.

[0051] In any embodiment, the methods described herein are useful for predicting in vivo non-target mediated clearance of a biological molecule. A “biological molecule” as referred to herein includes, without limitation, any therapeutic protein, polypeptide, or nucleic acid molecule.In any embodiment, the biological molecule interacts with or binds to a specific or defined target, e.g., a cellular target. Such biological molecules of the disclosure are referred to as “binding molecules” and these encompass therapeutic proteins, polypeptides, or nucleic acid molecules that comprise or are coupled to a domain having binding specificity to target molecule, i.e., a “binding target”. The binding target or target molecule can be a toxin, drug, protein, nucleic acid molecule, or other biological molecule. In any embodiment, the binding target is a protein expressed by a cell, for example a cell surface protein or cell surface receptor. In any embodiment, the biological molecule has binding specificity to a human target molecule, e.g., a human cell surface protein or human cell surface receptor.

[0052] Exemplary biological molecules comprise antigen binding proteins, such as antibodies (e.g.. human polyclonal or monoclonal antibodies), multispecific antibodies (e.g, human bispecific antibody proteins), antigen-binding fragments of an antibody (e.g, Fv, Fab', and (Fab'ty), and antibody derivatives (e.g., single-chain antibodies, minibodies, diabodies). Biological molecules also include therapeutic proteins (e.g, multispecific engineered proteins, recombinant proteins, and fusion proteins) and nucleic acid molecules (e.g.. siRNA molecules, antisense oligonucleotides, aptamer molecules, and mRNA molecules). Biological molecules of the present disclosure may also interact with or bind to the neonatal Fc receptor (FcRn). In one embodiment, the biological molecule naturally interacts with or binds to FcRn. In one embodiment, the biological molecule is engineered to bind FcRn. In one embodiment, the biological molecule comprises a fragment crystalhzable (Fc) region that interacts with FcRn. In one embodiment, the biological molecule comprises an Fc region engineered to increase FcRn binding affinity. In one embodiment, the biological molecule comprises a Fc portion engineered to reduce FcRn binding affinity or engineered to remove or delete FcRn binding. In one embodiment, the biological molecule comprises an albumin domain capable of binding to FcRn (e.g., a C-terminal Dill region of human serum albumin) and variants thereof (see e.g., Andersen et al., J. Biol. Chem. 289(19): 13492-13502 (2014), which is hereby incorporated by reference). In one embodiment, the biological molecule comprises an albumin-binding protein domain, e.g., albumin-binding domain B2A3 and B1A2B2A3 from Streptococcal protein G, that indirectly binds to FcRn via binding albumin (see e.g., Andersen et al., J. Biol. Chem. 286(7):5234-5241 (201 1), which is hereby incorporated by reference in its entirety). In one embodiment, the biological molecule comprises an affibody molecule, a small affinity' protein, that directly binds FcRn in a pH dependent manner (see Seijsing et al.. Applied Biol. Sci. 111(48): 17110-17115 (2014), which is hereby incorporated by reference in its entirety).

[0053] According to this and all aspects of the disclosure, the methods of predicting in vivo non-target mediated clearance of a biological molecule described herein involve providing a preparation of cells that do not express a target, e.g., a binding target, of the biological molecule. Suitable cell preparations include any mammalian cell preparation, e.g., a human cell preparation, a primate cell preparation, a canine cell preparation, a feline cell preparation, a porcine cell preparation, a rodent cell preparation, or any other suitable mammalian cell preparation. Suitable cells can be derived from any mammalian tissue, including, but not limited to skin (dermal and epidermal tissue), epithelium, vascular tissue (endothelial cells), heart, lung, kidney, liver, intestine, stomach, pancreas, colon, ovary7, lymph tissue, bone, cartilage, tumor, and can be a primary7cell preparation or a preparation of immortalized cells, e.g., a cell line. Exemplary cell line preparations suitable for use in the methods described herein include, without limitation, Chinese hamster ovary' (CHO) cells, Madin-Darby canine kidney (MDCK) cells, Vero cells, HEK- 293 cells, HeLa cells, 3T6 cells, A549 cells, BHK-21 cells, MCF-7 cells, Saos-2 cells, PC3 cells, HepG2 cells, and human umbilical vein endothelial cells (HUVEC).

[0054] The cell preparation useful in this aspect of the disclosure does not express the target molecule of the biological molecule, i.e., the cells of the preparation do not express the cell surface protein or receptor that the biological molecule binds to or interacts with. The cells of the preparation may naturally not express the target or any protein bound by the biological molecule. Alternatively, the cells of the preparation may express a variant of the target that is not bound by the biological molecule, e.g, a canine cell preparation may express a homolog of the target that is not bound by' a human biological molecule having binding specificity for the human binding target. In another embodiment, the cells of the preparation are modified so as not to express the target of the biological molecule. For example, expression of the target can be silenced using siRNA, shRNA, or other inhibitory nucleic acid molecule. Alternatively, expression of the target can be genetically silenced using gene editing technology'.

[0055] In accordance with this aspect of the disclosure, if the biological molecule is a molecule that interacts with or bind to an Fc receptor, then cells of the preparation also do not express an Fc receptor or protein. For example, when the biological molecule is a human immunoglobulin comprising an Fc region, cells of the preparation do not express a cell surface receptor or protein that binds to the Fc region of the human immunoglobulin. In any embodiment, cells of the preparation naturally do not express a human Fc receptor, such as FcaR, FcaRI, FcyRI, FcyRII, FcyRIII, FcRn, FceRI, or FcsRII, or any species equivalent thereof that binds to the Fc portion of a human immunoglobulin. In any embodiment, cells of the preparation do not express the human FcRn protein and, accordingly, do not express an active human FcRn / hp2m complex.The cells of the preparation may naturally lack expression of an Fc receptor or be engineered to not express such receptor (e.g., genetically modified to knock-down or silence Fc receptor expression). In any embodiment, suitable cell preparations include those engineered to not express hFcRn or hp2m, w here the absence of either protein renders the cells of the preparation incapable of expressing an active FcRn / hp2m complex.

[0056] Cell preparations suitable for use in the methods and assays described herein are cultured with a suitable cell culture medium under standard tissue culture conditions suitable for the survival, growth, and proliferation of the cells. Appropriate growth and culture conditions for various mammalian cell types are well known in the art. The cells of the preparation may be seeded onto and / or within a substrate from a suspension so that they are evenly distributed at a relatively high surface and / or volume density. The cell suspensions may comprise approximately about 1 xlO4to about 5x l07cells / ml of culture medium, or approximately about 2xl06cells / ml to about 2x l07cells / ml, or approximately about 5x l06cells / ml. The optimal concentration and absolute number of cells will vary' with cell type, growth rate of the cells, substrate material, and a variety of other parameters. The suspension may be formed in any physiologically acceptable medium, preferably one that does not damage the cells or impair their ability to adhere to the substrate. Appropriate mediums include standard cell growth media, e.g., DMEM with 10% FBS.

[0057] In carrying out the method of predicting in vivo non-target mediated clearance, the biological molecule of interest is added to the cell culture media and the cells are cultured under otherwise normal culture conditions (e.g., 37°C, 5% CO2, and 20% O2) in the presence of the biological molecule for a suitable duration of time, e.g., 30 minutes, 60 minutes, 90 minutes, 120 minutes, 180 minutes, 3 hours, 4 hours, 5 hours, or longer to allow non-target mediated cellular uptake to occur. In any embodiment, the cells are cultured in the presence of media containing the biological molecule for about 60 to about 90 minutes.

[0058] The cells are cultured in the presence of media containing the biological molecule under conditions mimicking in vivo physiological conditions, including physiological pH. In any embodiment, the physiological pH is a neutral pH, e.g, the pH of blood. A suitable neutral pH of the media containing the biological molecule is a pH of about 7.0 to about 8.0. In any embodiment, the media containing the biological molecule has a pH of about 7.0, a pH of about 7.1, a pH of about 7.2, a pH of about 7.3, a pH of about 7.4, a pH of about 7.5, a pH of about 7.6, a pH of about 7.7, a pH of about 7.8, a pH of about 7.9, or a pH of about 8.0.

[0059] In another embodiment, the physiological pH is an acidic pH to mimic the pH ty pical of endosomal acidification. In any embodiment, the media containing the biological molecule has a pH of about 5.6 to about 6.9. For example, the biological molecule has a pH ofabout 5.6, a pH of about 5.7, a pH of about 5.8, a pH of about 5.9, a pH of about 6.0, a pH of about 6. 1, a pH of about 6.2. a pH of about 6.3, a pH of about 6.4, a pH of about 6.5. a pH of about 6.6, a pH of about 6.7, a pH of about 6.8, or a pH of about 6.9.

[0060] After the cell preparation is cultured in the presence of media containing the biological molecule under conditions mimicking in vivo physiological conditions, the amount of biological molecule taken up by the cells of the preparation is determined. In any embodiment, a first cell preparation is cultured in the presence of media containing the biological molecule, where the media has a neutral pH, and a second cell preparation, that is otherw ise identical to the first cell preparation, is cultured in the presence of media containing the biological molecule, where the media has an acidic pH. A comparison of the amount of biological molecule taken up by the cells exposed to the biological molecule under neutral pH conditions vs. cells exposed to the biological molecule under acidic pH conditions informs the pH dependency of the non-target mediated uptake of the biological molecule.

[0061] In one embodiment, determining the amount of biological molecule taken up by the cells of the preparation involves permeabilizing cells of the preparation and detecting the amount of the biological molecule taken up by the cells. To detect the biological molecule taken up by the cells, the biological molecule can be labeled directly or indirectly with a detectable moiety. For example, in one embodiment, the biological molecule is directly conjugated to a detectable moiety', such as a fluorescent moiety, prior to culturing with the cells, and the amount of the detectable moiety in the permeabilized cells is detected to determine the amount of biological molecule taken up by the cells. The biological molecule can alternatively be directly conjugated or coupled to an enzymatic label or a small molecule label (e.g., biotin), and the amount of biological molecule taken up by the cells is determined by incubating the permeabilized cells with reactants appropriate to detect the enzymatic label or small molecule label. In another embodiment, the biological molecule is indirectly labeled with a detectable moiety by incubating the permeabilized cells with an antibody (e.g. , IgG or IgM molecules), antibody fragment (e.g. , Fab F(ab’), Fc fragments, single domain antibody, gamma chain of IgG, Fc5p of IgM, or Mu chain of IgM), or an antibody derivative (e.g., scFv) having binding specificity for the biological molecule, where the antibody is conjugated to a detectable moiety (i.e., a labeled secondary antibody). Suitable detectable moieties include, without limitation, fluorescent molecules, small molecules, enzy matic labels, radioisotopes, as readily known in the art. Following this incubation, the amount of detectable moiety is detected to determine the amount of biological molecule taken up by the cells.

[0062] Suitable methods of detecting a detectable moiety include those known and readily used in the art and described in the Examples herein. For example, in any embodiment, thedetectable moiety is a fluorescent moiety, e.g., Alexa Fluor dyes, fluorescein. Oregon Green dyes, Rhodamine dyes, Texas Red dyes, and / or derivatives thereof, and the fluorescent moiety is detected via a fluorometer in an immunoassay e.g., ELISA), fluorescence microscopy, or flow cytometry. Other suitable detectable moieties include radioisotopes (e.g.,14C,125I,32P,35S), luminescent proteins, enzymatic proteins (e.g., horseradish peroxidase, alkaline phosphatase, glucose oxidase, etc.), and small molecules (e.g., biotin), that can be detected using immunoassays, flow cytometry, spectrometry, microscopy, confocal microscopy, liquid chromatography with tandem mass spectrometry, or scintigraphy methods that are well known in the art.

[0063] Once the amount of detectable moiety is detected, it is compared to one or more reference values to quantify the amount of biological molecule taken up by the cells, thereby providing a value for non-specific endocytosis. For example, as described herein, a standard curve or calibration curve representative of the biological molecule binding capacity can be generated. The amount of biological molecule detected in the cells is compared or plotted against the standard curve to quantify the amount of biological molecule taken up by the cells. As described herein, a standard curve can be generated using commercially available microspheres coated with an anti- Fc antibody (e.g., an anti-IgG antibody) that binds to the biological molecule of interest. A comparison of the level of non-specific endocytosis of the biological molecule as determined from the calibration curve to the level of non-specific endocy tosis of one or more internal reference antibodies is performed to predict the non-target mediated clearance of the biological molecule. In a preferred embodiment, the amount of biological molecule taken up by non-specific endocytosis means as determined from the calibration curve is compared to corresponding values of at least two reference antibodies, one refence antibody having low non-target clearance (e.g., <4.5 mL / kg / d) and one having a high non-target clearance (e.g., >4.5 mL / kg / d).

[0064] Alternatively, the amount of biological molecule detected in the cells is compared to one or more internal reference antibodies included in the assay and detected alongside of the biological molecule, where the non-specific endocytosis of the reference antibodies and in vivo non-target dependent clearance have been previously quantified and are known. A comparison of the level of non-specific endocytosis of the biological molecule being tested to the level of nonspecific endocytosis of one or more internal reference antibodies is performed to predict the non- target mediated clearance of the biological molecule of interest. In a preferred embodiment, the amount of biological molecule taken up by non-specific endocytosis is compared to at least two reference antibodies, one refence antibody having low non-target clearance (e.g., < 4.5 mL / kg / d in humans) and one having a high non-target clearance (e.g., > 4.5 mL / kg / d in humans).

[0065] The level of non-specific endosomal uptake can also be utilized to inform and predict human subcutaneous bioavailability (FSQ), where a high level of non-specific endocytosis indicates a low FSQ (e.g., < 50%) and a low level of non-specific endocytosis indicates a high FSQ (e.g., > 50%). In any embodiment, threshold values for “high”, “medium”, and “low” levels of non-specific endocytosis can be generated using a plurality of reference antibodies as described herein (see e.g. , Example 9 and Figure 21). For example, a low non-specific endocytosis threshold is identified by defining the average quantitative value of non-specific endocytosis of one or more reference mAbs possessing CLind below 4.5 mL / kg / d and / or FSQ above 50% in human. The upper bound of the 95% confidence interval of this mean can sen e as the reference range for low nonspecific uptake. Medium non-specific uptake can be established via the non-specific endocytosis value corresponding to two standard deviations plus the low CLind endocytosis mean value. A tested biological molecule with non-specific uptake in the medium range is expected to have a higher chance of increased CLind and / or low FSQ. Anything above the medium uptake threshold is considered as possessing high non-specific endocytosis and a corresponding greatly enhanced chance of high CLind and / or low FSQ in human.

[0066] Another aspect of the present disclosure is directed to a method of predicting in vivo non-target mediated clearance of aneonatal Fc receptor (FcRn) interacting molecule. This method assesses both non-specific endocytosis and FcRn recycling to enhance the in vivo prediction of non-target mediated clearance of a biological molecule. This method involves providing a first cell preparation, wherein cells of the first preparation do not express human FcRn, and providing a second cell preparation, wherein cells of the second preparation express a heterodimer of human neonatal Fc receptor (hFcRn) and human (32m (hp2m). Cells of the first and second cell preparation do not express a target of the FcRn interacting molecule. This method further involves subjecting the first and second cell preparations to first and second incubation periods, where the first incubation period comprises incubating the cell preparations with media containing the FcRn interacting molecule under acidic conditions, under non-acidic conditions, or under acidic and non- acidic conditions. The second incubation period comprises incubating said cell preparations, after the first incubation, with media lacking the FcRn interacting molecule under non-acidic conditions. The method further involves determining an amount of FcRn interacting molecule taken up by cells of the first and second preparations after said subjecting to the first incubation period and / or the second incubation period, and measuring an amount of the FcRn interacting molecule in media following the second incubation period. The method further involves quantifying non-specific endocytosis and FcRn recycling of the FcRn interacting molecule based on the determining and measuring steps, z.e., non-specific endocytosis is quantified based on the amount of FcRninteracting molecule determined to be taken up by the cells, and FcRn recycling of the FcRn interacting molecule is quantified based on the amount of FcRn interacting molecule measured in the media after the second incubation. The in vivo non-target mediated clearance of the FcRn interacting molecule is predicted based on the quantifying step.

[0067] This method of the present disclosure provides a tool for in vitro evaluation of potential liabilities in non-target mediated clearance of candidate therapeutic molecules. This method facilitates prediction or estimation of in vivo non-target mediated clearance of a particular candidate therapeutic molecule or can be utilized to rank order in vivo non-target mediated clearance of candidate molecules to support lead candidate selection and optimization while reducing the number of molecules tested in animal models.

[0068] In accordance with this aspect of the disclosure, an “FcRn interacting molecule” as referred to herein is any therapeutic protein, polypeptide, or nucleic acid molecule comprising a domain or sequence portion that binds to the neonatal Fc receptor (FcRn). FcRn is an MHC class I-like molecule that is a heterodimer of an alpha chain non-covalently bound to P2-microglobulin. FcRn plays a critical role in maintaining in vivo levels of IgG and albumin by preventing lysosomal degradation of these molecules within cells. Thus exemplary FcRn interacting molecules for purposes of the present disclosure include, without limitation, molecules comprising an Fc domain (e.g., an Fc domain of an IgG), an albumin domain (e.g., full length albumin or a C-terminal Dill fragment thereof), an albumin domain variant having enhanced FcRn binding affinity (see e.g., Andersen et al., J. Biol. Chem. 289(19): 13492-13502 (2014), which is hereby incorporated by reference), an albumin-binding protein domain from Streptococcal protein G or other grampositive bacteria (see e.g., Andersen et al., J. Biol. Chem. 286(7):5234-5241 (2011), which is hereby incorporated by reference in its entirety), or a small affinity' protein (affibody), that directly binds FcRn in a pH dependent manner (see Seijsing et al.. Applied Biol. Sci. 111(48): 17110-17115 (2014), which is hereby incorporated by reference in its entirety).

[0069] In one embodiment, the FcRn interacting molecule comprises a domain or sequence portion that binds to human FcRn. Suitable FcRn interacting molecules for purposes of the present disclosure also encompass molecules comprising an FcRn-binding domain engineered (i.e., comprising one or more amino acid substitutions, insertions, or deletions) to possess enhanced FcRn binding affinity relative to the binding affinity of the non-engineered version of the domain. FcRn interacting molecules also encompass molecules comprising an FcRn-binding domain engineered to possess reduced FcRn binding affinity relative to the FcRn binding affinity’ of the non-engineered version of the domain. FcRn interacting molecules further encompass molecules comprising an FcRn-binding domain engineered to delete or remove FcRn binding. In anyembodiment, the FcRn interacting molecule may also interact with or bind to a non-FcRn cellular target, e.g.. a cell surface receptor or cell surface ligand.

[0070] Exemplary FcRn interacting molecules comprise antigen binding proteins, such as antibodies (e.g, human immunoglobulins including IgGl, IgG2, IgG3, and IgG4 immunoglobulins), multispecific antibodies (e g., human bispecific or trispecific antibody proteins comprising an IgG Fc domain or other FcRn binding domain), antigen-binding fragments of an antibody comprising an IgG Fc domain or other FcRn binding domain, antigen-binding fragments of a multispecific antibody comprising an IgG Fc domain or other FcRn binding domain, and antibody derivatives comprising an IgG Fc region or other FcRn binding domain. FcRn interacting molecules also include therapeutic and candidate therapeutic proteins (e.g., multispecific engineered proteins, recombinant proteins, Fc-fusion proteins, albumin fusion proteins) and nucleic acid molecules (e.g., siRNA molecules, antisense oligonucleotides, aptamer molecules, and mRNA molecules) comprising an FcRn binding domain.

[0071] In accordance with this aspect of the disclosure, suitable first and second cell preparations include any of the mammalian cell preparations described supra, e.g., a human cell preparation, a primate cell preparation, a canine cell preparation, a feline cell preparation, a porcine cell preparation, or a rodent cell preparation. Suitable cells can be derived from any tissue, including, but not limited to skin (dermal and epidermal tissue), epithelium, vascular tissue (endothelial cells), heart, lung, kidney, liver, intestinal, pancreatic, colon, ovarian, lymph tissue, bone, cartilage, tumor, and can be a primary cell preparation or a preparation of immortalized cells, e.g., a cell line. Exemplar}' cell line preparations include, without limitation, Chinese hamster ovary (CHO) cells, Madin-Darby canine kidney (MDCK) cells, Vero cells, HEK-293 cells, HeLa cells, 3T6 cells, A549 cells. BHK-21 cells. MCF-7 cells, Saos-2 cells, PC3 cells, HepG2 cells, human umbilical vein endothelial cells (HUVEC).

[0072] This aspect of the disclosure requires the use of a first and a second cell preparaton, where the first and second preparations differ in their expression of human neonatal Fc receptor (FcRn). For example, cells of the first cell preparation do not express hFcRn, while cells of the second cell preparation express a functional FcRn, i.e., a heterodimer of human FcRn and human 02m (h[32m).

[0073] Cells of the first cell preparation may be cells that naturally do not express hFcRn, e.g., CHO cells are hamster kidney cells that do not express human FcRn. Alternatively, cells of the preparation may be modified to silence expression of the hFcRn for purposes of carrying out this method of the present disclosure. In any embodiment, cells expressing hFcRn can be rendered suitable for use as the first cell preparation by silencing expression of hFcRn and / or h02m usingsiRNA or other inhibitory nucleic acid molecule. Alternatively, cells expressing hFcRn and h(32m can be genetically modified using gene editing technology to "knock-down" or “knock-out” the expression of hFcRn and / or h|32m.

[0074] Cells of the second cell preparation may naturally express a functional hFcRn protein. For example, suitable second cell preparations include human cell preparations that express hFcRn and hp2m. Alternatively, cells of the second preparation may not naturally express a functional hFcRn protein and need to be modified to achieve such expression. In this embodiment, cells of the second preparation may be modified to express one or more heterologous genes, e.g. , the FCGRT gene, which encodes the human neonatal Fc receptor and / or the B2M gene, which encodes the human beta-2 -microglobulin protein.

[0075] In accordance with this aspect of the disclosure, the first and second cell preparations do not express the target of the FcRn interacting molecule, e.g., a binding target of the FcRn interacting molecule. As described supra, suitable cells may naturally not express the target of the FcRn interacting protein. Alternatively, suitable cells include those that express a variant of the target (e.g., a species variant of the target) that does not bind to or interact with the human FcRn interacting molecule. Yet, in another embodiment, suitable cells are those that have been modified to silence the expression of the target.

[0076] As noted above, this method of predicting in vivo non-target mediated clearance of a FcRn interacting molecule involves a first and second incubation period. In the first incubation period, the first and second cell preparations are incubated with culture media comprising the FcRn interacting molecule. This first incubation period is referred to herein as the “Load Phase” where non-target mediated cell uptake of the FcRn interacting molecule is allowed to occur via nonspecific endocytosis. FcRn-mediated uptake, or a combination of both. In one embodiment, this first incubation is carried out under non-acidic, neutral pH conditions. Thus, in one embodiment, a subset of cells from each first and second cell preparation is incubated with culture media comprising the FcRn interacting molecule under neutral pH conditions. Suitable neutral pH conditions include a media pH of about 7.0 to about 8.0, preferably a pH of about 7.2, a pH of about 7.3, a pH of about 7.4. a pH of about 7.5, or a pH of about 7.6.

[0077] In some instances, an FcRn interacting molecule does not exhibit high levels of nonspecific endocytosis at a neutral pH. In other circumstances, it is desirable to use lower amounts of the FcRn interacting molecule, such as during early phases of development where material quantities are limited. In these instances, the Load Phase is carried out under acidic pH conditions to enhance total endocytosis (i.e. non-specific and FcRn-mediated) to ensure a sufficient amount of the FcRn interacting molecule is loaded into the cells to assess the FcRn recycling efficiency ofthe molecule. Thus, in one embodiment, a subset of cells from each first and second cell preparation is incubated with culture media comprising the FcRn interacting molecule under acidic pH conditions. As described supra, acidic pH conditions include a media pH of about 5.6 to about 6.9, preferably a pH of about 5.6, a pH of about 5.7, a pH of about 5.8, a pH of about 5.9, or a pH of about 6.0.

[0078] In instances where the first incubation period is carried out under acidic pH conditions, it is preferable that a concentration of the FcRn interacting molecule that does not saturate the FcRn receptor binding capacity within the FcRn expressing cells is utilized. As shown herein FcRn interacting molecules having high FcRn affinity7(e.g., antibodies comprising YTE substitutions (M252Y / S254T / T256E)), can saturate FcRn binding under acidic conditions, leading to inaccurate results. To accurately assess non-specific endocytosis and FcRn recycling efficiency of these molecules, the concentration of the FcRn interacting molecule added to the media for the first incubation is optimized to achieve <100% FcRn binding (e.g., optimized to achieve -90% FcRn binding occupancy).

[0079] In another embodiment, a subset of cells from each first and second cell preparation is incubated with culture media comprising the FcRn interacting molecule under acidic pH conditions and a second subset of cells from each first and second cell preparation is incubated with culture media comprising the FcRn interacting molecule under non-acidic, neutral pH conditions. Assessing the cellular uptake of a novel FcRn interacting molecule under neutral (non- acidic) and acidic conditions provides a means to determine the molecule’s propensity to exhibit high levels of non-specific endocytosis at physiologically relevant pH and measure FcRn recycling at the same time. As noted supra, the FcRn interacting molecule may be a molecule comprising a FcRn-binding domain that has been engineered to reduce FcRn binding affinity or remove / ablate FcRn binding affinity. These FcRn interacting molecules can be incubated with the first and second cell preparation under both the acidic and non-acidic conditions to confirm the reduction or ablation of FcRn-mediated uptake and / or recycling of the engineered molecule.

[0080] In accordance with the disclosed method of predicting in vivo non-target mediated clearance of an FcRn interacting molecule, the first and second cell preparations are subject to a first incubation period. During this first incubation period the cell preparations are incubated with media containing a concentration of the FcRn interacting molecule that does not saturate FcRn- mediated uptake capacity7of the cells in the preparation. As demonstrated herein, FcRn interacting molecules that have been engineered to have enhanced FcRn binding can saturate the FcRn binding more readily under acidic conditions. When the per cell FcRn binding capacity is saturated, nonspecific endocytosis and FcRn recycling parameters cannot be adequately or accurately assessed.Therefore, in one embodiment, the concentration of the FcRn interacting molecule in the media during the first incubation is a concentration that achieves <100% of FcRn occupancy per cell of the cells of the second preparation. In one embodiment, the concentration of the FcRn interacting molecule in the media is a concentration that achieves <99% of FcRn occupancy per cell of the cells of the second preparation. In one embodiment, the concentration of the FcRn interacting molecule in the media is a concentration that achieves >25% FcRn occupancy per cell of the cells of the second preparation. In one embodiment, the concentration of the FcRn interacting molecule in the media is a concentration that achieves between 25% and 99% FcRn occupancy per cell of the cells of the second preparation. In one embodiment, the concentration of the FcRn interacting molecule in the media is a concentration that achieves between 25% and 90% FcRn occupancy per cell of the cells of the second preparation. In one embodiment, the concentration of the FcRn interacting molecule in the media is a concentration that achieves between 50% and 99% FcRn occupancy per cell of the cells of the second preparation. In one embodiment, the concentration of the FcRn interacting molecule in the media is a concentration that achieves between 50% and 90% FcRn occupancy per cell of the cells of the second preparation.

[0081] A suitable concentration of FcRn interacting molecule can be determined by conducting a concentration-dependent FcRn interaction study, where the study informs the concentration of FcRn interacting molecule that will achieve a defined level of FcRn occupancy (e.g. 50% FcRn occupancy per cell of cells within a cell preparation). This is preferably determined for each type or format of FcRn molecule being tested, where the type or format refers to the FcRn interacting portion of the molecule (e.g., the Fc portion). Exemplary types or formats of an FcRn interacting molecule include, without limitation, wildtype IgG Fc portions, IgG Fc portions engineered to have enhanced FcRn binding, IgG Fc portions engineered to have reduced or ablated FcRn binding. When the assay described herein is utilized to compare non-target mediated cellular uptake of FcRn interacting molecules having different formats, it is important to optimize the concentration of each molecule format being tested to equalize the amount of each FcRn interacting molecule that is being loaded into the cell by an FcRn-mediated manner.

[0082] In one embodiment, a suitable concentration-dependent FcRn interaction study comprises a concentration-dependent FcRn binding study. The binding affinity curve generated in a concentration-dependent FcRn binding study can be used to determine the equilibrium dissociation constant (KD) between the FcRn-interacting molecule and FcRn. The KD is the concentration of the FcRn interacting molecule at which 50% of the total FcRn receptors are bound by the FcRn interacting molecule. Therefore, in one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is the KD. In one embodiment, a suitableconcentration of FcRn interacting molecule in the media for the first incubation is 1 / 2KD (to achieve at least 25% FcRn occupancy). In one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is 2X the KD (to achieve -90% FcRn occupancy). In one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is between 2X and 10X the KD (to achieve ~90%-99% FcRn occupancy).

[0083] In one embodiment, a suitable concentration-dependent FcRn interaction study comprises a concentration-dependent endocytosis assay as described herein. The cell uptake rate curve generated in a concentration-dependent endocytosis study can be used to determine the Km of the FcRn-interacting molecule. As described herein, the Km is the concentration of half maximum velocity or concentration at which approximately 50% of FcRn binding capacity is reached. Therefore, in one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is the Km determined from the endocytosis study. In one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is l / 2Km (to achieve at least 25% FcRn binding capacity). In one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is 2X the Km (-90% of FcRn binding capacity ). In one embodiment, a suitable concentration of FcRn interacting molecule in the media for the first incubation is between 2X and 10X the Km (-90%- 99% FcRn binding capacity).

[0084] As noted above, the first incubation period in the disclosed method of predicting in vivo non-target mediated clearance of an FcRn interacting molecule is referred to as the “Load Phase” where non-target mediated cell uptake is occurring via non-specific endocytosis, FcRn- mediated uptake, or a combination of both . The first and second incubation periods are carried out for a duration of time that is sufficient for non-specific endocytosis and / or FcRn mediated uptake to occur, for example, about 30 minutes, about 60 minutes, about 90 minutes, about 120 minutes, about 180 minutes, about 3 hours, about 4 hours, or more at 37°C. The amount of FcRn interacting molecule taken up by the cells after the first incubation period is referred to herein as the “total uptake” amount or “uptake concentration”. This amount can be determined in a subset of the first and second cell preparations by permeabilizing the cells to detect and quantify the amount of FcRn interacting molecule present in the cells as described supra.

[0085] In the remaining first and second cell preparations, the media containing the FcRn interacting molecule is removed and cells are washed one or more times with media or physiological saline solution to remove FcRn interacting molecule that has not been taken up by the cells. The cells are then subject to the second incubation period to assess FcRn recycling ofthe FcRn interacting molecule. This second incubation period is the “Recycling Phase'’ and it is carried out for an amount of time sufficient for FcRn capture of the FcRn interacting molecule within an endosome compartment, endosomal trafficking to cell surface, and release / recycling of the FcRn interacting molecule back into the media. This incubation period can be about 30 minutes, about 60 minutes, about 90 minutes, about 120 minutes, about 180 minutes, about 3 hours, about 4 hours, about 5 hours, about 6 hours, or more. Following the second incubation period, the recycling phase media is collected and the amount of FcRn interacting molecule recycled and released into the media is detected and quantified. The amount of FcRn interacting molecule detected is quantified via comparison to an internal reference value or standard curve generated as described above. The amount of FcRn interacting molecule in the media of the second cell preparation (expressing FcRn) after the second incubation minus any amount of FcRn interacting molecule detected and measured in the media of the first cell preparation (not expressing FcRn) is the “recycled amount” or “recycled concentration” of FcRn interacting molecule.

[0086] Finally, the amount of FcRn interacting molecule remaining in the cells of the second cell preparation after the first and second incubations is determined following the second incubation period. This amount of FcRn interacting molecule remaining in the cells is referred to herein as the “residual amount” or “residual concentration” of the FcRn interacting molecule. FcRn interacting molecule remaining in the cell preparations is detected and quantified as described supra, i.e., by labeling the FcRn interacting molecule directly or indirectly with a detectable moiety and detecting the detectable moiety by flow cytometry, microscopy, spectroscopy, scintigraphy, or other immunoassay.

[0087] In any embodiment, the recycled, residual, and uptake concentrations of the FcRn interacting molecule obtained while carrying out the method of the present disclosure can be used to calculate hFcRn recycling efficiency metric (FREM) scores by applying the following formula:

[0088] In the above formula. Rx is the recycled concentration of FcRn interacting molecule ‘X’ from 37°C samples following the load and subsequent recycling phases, and RAx is the residual concentration of FcRn interacting molecule ‘X’ from the 37°C samples following the load and subsequent recycling phases (Grevys, et al. iScience 25, 103746 (2022); and Grevys. et al. Nat Commun 9, 621 (2018), which are hereby incorporated by reference in their entirety.

[0089] The uptake concentrations by the first and second cell preparations, i.e., the FcRn and FcRn1expressing cells, respectively, for each FcRn interacting molecule are used to calculate the non-specific uptake coefficient (NUC) by the following formula: rrhFcRn+ rjhFcRn-NUCUX ~UXIjhFcRn-

[0090] In the above formula, [jFcRn+is the uptake concentration (both FcRn dependent and independent) of the molecule ‘X’ in FcRn+cells (i.e., second cell preparation) and u^FcRn~ is the FcRn independent uptake by the molecule ‘X’ in FcRn cells (i.e., first cell preparation) (Grevys, et al. iScience 25, 103746 (2022), which is hereby incorporated by reference in its entirety).

[0091] In one embodiment, when the NUC value is between 0 and 1 (0 < NUC <1), it is an indication that the cellular uptake process is dominated by FcRn-independent non-specific endocytosis (Grevys, etal. iScience 25, 103746 (2022), which is hereby incorporated by reference in its entirety). In these cases, the FREM score is multiplied by the NUC. When NUC >1, it is an indication that FcRn-mediated uptake of the molecule CX‘) with negligible non-specific endocytosis. For these molecules, the FREM score is not multiplied by the NUC value.

[0092] As demonstrated herein, the FREM score inversely correlates with targetindependent clearance values, where a higher FREM score, i.e., a higher recycling efficiency, predicts low target-independent clearance, and a lower FREM score, i.e., a lower recycling efficiency, predicts a high level of target-independent clearance (see Figure 17C).

[0093] The invention having been described, the following examples are offered by way of illustration, and not limitation.EXAMPLESMaterials and Methods for Examples 1-3

[0094] Cell Culture: Parental Chinese Hamster Ovary-Kl (CHO-K.1) cells were maintained in Ham’s F-12K media (Thermo Fisher, 21127022) containing 10% heat inactivated fetal bovine serum (Thermo, 16140071). Vero cells were purchased from the American Type Culture Collection (ATCC; CCL-81, lot 70016956) and grown in ATCC-formulated EMEM with 10% heat-inactivated fetal bovine serum (ATCC; 30-2003). Cells were cultured without antibiotics in a humidified, 5% CO2 incubator at 37°C and passaged utilizing 0.05% trypsin-EDTA (Thermo, 25300054). Mycoplasma testing was conducted using the MycoAlert Detection Kit (Lonza). Cells were cryopreserved using complete growth media containing 5% (v / v) DMSO.

[0095] Monoclonal antibodies (mAbs)’. Two fully human, wild type mAbs were used for the initial characterization: an IgG2 anti-IL-4Ra mAb (AMG 317) and an IgGl control mAb raisedagainst streptavidin (anti-streptavidin antibody, ASA). Both mAbs along with the anti-IL-4Ra mAb mutant series were constructed by recombinant DNA technology and produced in stably transfected CHO cells using standardized protocols. The preclinical mAb panel included five antibodies raised against a G-protein-coupled receptor (mAbs B1-B5). Only the anti-IL-4Ra mAb and ASA were wild type IgGs. The other mAbs were produced in stably transfected CHO cells on the Amgen patented stable effector functionless Fc backbone as described previously (Liu et al., J. Biol. Chem 292(5): 1876-1883 (2016), which is hereby incorporated by reference in its entirety).

[0096] PK Studies and Analyses: PK data was obtained from three separate studies. Female wild type C57BL / 6J (for ASA and the anti-IL-4Ra mAb) and Balb / C (for mAb B1-B5 panel) mice were purchased from Jackson Laboratory (Bar Harbor, MA). The proteins of interest w ere administered as a 1 mg / kg (anti-lL-4Ra mAb). 2 mg / kg (mAb B1-B5 preclinical panel), or 3 mg / kg (ASA) intravenous bolus dose via the lateral tail vein. Blood specimens were collected at various times post injection, incubated at ambient temperature for approximately 20 minutes or until fully clotted, and then centrifuged to separate the serum. All serum specimens were stored at -70°C (± 10°C) until use in analytical assays. Mice were cared for in accordance with the Guide for the Care and Use of Laboratory Animals, 8th Edition at AAALAC, International accredited facilities. All mice protocols were approved by the Amgen, Inc. Institutional Animal Care and Use Committee (Thousand Oaks, CA).

[0097] Quantitation of proteins in mouse serum was performed via electrochemiluminescent immunoassays on the MSD Sector 600 instrument (Meso Scale Diagnostics, Rockville, MD) using an anti-human Fc antibody as both the capture and detection reagent. In all assays, the analyte serum concentrations were interpolated from standard curves using the corresponding analyte prepared in pooled mouse serum using Watson LIMS software (Thermo).

[0098] The murine PK of all tested mAbs was described using a two-compartment model possessing a central and distribution compartment (Tang et al., J. Parm Sci. 93(9): 2184-204 (2004), which is hereby incorporated by reference in its entirety). Parameters included concentration and volume within the respective compartments, elimination from the central compartment, and distribution between the compartments. Computational modeling was performed in Phoenix (Certara, Princeton, NJ) using mean values of the serum concentration time profiles and plotted using GraphPad Prism (Dotmatics, Boston, MA). This method was chosen because linear clearance was assumed for all mAbs at the doses utilized due to lack of measurable target binding in mouse. Specifically for the reporting of human clearance of the anti-IL-4Ra mAb, the previously reported CLind value was normalized to 80 kg as this body weight was used to modelits PK (Kakkar et al., Pharm Res. 28(10:2530-42 (2011), which is hereby incorporated by reference in its entirety)

[0099] Cell Uptake Studies: CHO-K1 and Vero cells were seeded onto 96-well plates at a density of 150,000 cells per well 24 h prior to experimentation, or 75,000 cells per well 48 h beforehand. On study days, cells were washed twice in either pre-warmed (37°C) or ice-cold (4°C) Ringer’s solution (pH 7.4, 122.5 mM NaCl, 5.4 mM KC1, 1.2 mM CaCh, 0.8 mM MgCh, 0.8 mM Na2HPO4, 0.2 mM NaH2PO4, 5.5 mM d-glucose, and 10 mM HEPES). Fresh Ringer’s solution was added a third time, and cells were equilibrated at either temperature for 30 min. Afterwards, wells were aspirated, and then the treatments were immediately applied. Studies utilized 37°C (enables cell surface binding, internalization, intracellular trafficking) and 4°C (enables cell surface binding) conditions. For the time-dependent study, CHO-K1 cells were incubated with 100 pg / mL mAb for 15, 30, 60, or 120 mins. For the concentration-dependent study, CHO-K1 cells received concentrations diluted 1 :2 spanning 100-25 pg / mL for 60 min. These conditions were based on preliminary data that were found to he below the limit of detection saturation. Following incubations, cells were washed four times on ice with ice-cold Ringer’s, then trypsinized for 4 min at 37°C. Once cells were detached, complete ice-cold growth media was added at a 1 : 1 ratio (y / v) to inhibit the trypsin. Cells were removed from cell culture plates and placed into V-bottom 96- well plates (Fisher #249944), then centrifuged at 300 g for 5 min at 4°C. Cells were stained in 100 pL IX PBS containing 0.5% (v / v) Zombie UV fixable viability dye (BioLegend # 423108) on ice for 20 min. Cells were washed once with 100 pL FACS buffer (IX PBS, 2% w / v BSA ,lmM EDTA, 0.1% w / v sodium azide), centrifuged at 300 g for 5 min at 4°C, then fixed / permeabilized in the dark for 20 min at room temperature using the Cyto-Fast fix / perm buffer (Biolegend # 426803). Cells were then washed twice using IX Cyto-Fast Perm Wash solution (Biolegend # 426803). Samples were next incubated with 5 pg / mL (100 pL per well, Cyto-Fast Perm Wash solution) of mouse anti-human Fc monoclonal IgGl antibody fluorescently conjugated with Alexa Fluor 647 (Ab35-AF647) for 30 min on ice in the dark (Hall et al., J. Immunol. Methods 393(1- 2):70-73 (2013), which is hereby incorporated by reference in its entirety). Samples were washed thrice with Cyto-Fast Perm Wash solution, resuspended in 100 pL Cyto-Fast Perm Wash solution, and analyzed on a BD FACSymphony flow cytometer with an 18 color, 5-laser configuration (UV- 355 nm, violet-405 nm, blue-488 nm, yellow / green-561 nm, red-637 nm) and a BD Biosciences High Throughput Sampler (Catalog #338301). Gating was conducted as depicted within figures, obtaining a target of 10,000 single-cell, Zombie' (live) events using BD Diva software. For Vero cell samples, approximately 5000 single hve cell events were captured due to differences in trypsin dissociation rates between Vero and CHO-K1 cell lines, which led to less Vero cells beingcollected. Data was analyzed using FlowJo software (Becton Dickson, Franklin Lakes, NJ) to obtain median fluorescent intensities of the indicated cell populations.

[0100] Antibody Non-Specific Endocytosis (NSE) with Fluorescence Microscopy: Endocytosis studies were carried out as described directly above using 100 pg / mL of either ASA or the anti-IL-4Ra mAb for 60 min at 37°C, then cells were washed 4 times on ice with ice-cold Ringer’s. Cells were then fixed using 4% paraformaldehyde in IX PBS for 15 min, and then washed twice with IX PBS. Samples were then blocked and permeabilized in IX Cyto-Fast Perm Wash buffer for one hour at room temperature. After aspiration, wells were stained with 0.5 pg / mL Ab35-AF647 in IX Cyto-Fast Perm Wash buffer for 1 hr at room temperature. Next, samples were washed three times with IX Cyto-Fast Perm Wash buffer and then stained with 1 pg / mL Hoechst (Thermo #H1399, diluted in IX PBS) and 2 pg / mL HCS Cell Mask Blue (Thermo # H32720, diluted in DMSO) solutions in IX Cyto-Fast Perm Wash buffer for 30 min at room temperature. Wells were then washed thrice with IX PBS and imaged on an Opera Phenix High Content Screening system (PerkinElmer) using a 40x water objective with Z-stack acquisition. Maximum projections were obtained using Columbus software (PerkinElmer).

[0101] Antibody Binding Capacity: Quantum™ Simply Cellular® Mouse IgG beads were obtained from Bio-Rad (#FCSC815; lot #15515; Hercules, CA). The initial evaluation utilized final Ab35-AF647 concentrations of 5 or 10 pg / mL exactly following the vendor product information with all procedures done on ice or at 4°C. Subsequent work used a concentration of 5 pg / mL to match cell staining experiments as it was demonstrated this concentration saturated each bead population with antibody. For each uptake study, freshly stained beads were analyzed on the same flow cytometer using identical instrument parameters for the cell samples from that day. Bead median fluorescent intensities were obtained using FlowJo software, then plotted against the known antibody binding capacities per bead population supplied by the vendor via GraphPad Prism. The data was fit to a straight line using simple linear regression. The antibody binding capacity for each cell sample was then interpolated from the cell median fluorescent intensities.

[0102] Homology Modeling and Anti-IL-4Ra mAb Mutant Design: Antibody homology models were built using Molecular Operating Environment (MOE v2022.02, Chemical Computing Group; Montreal, Canada) and default settings in the Antibody modeler (Amber 10 forcefield, template search by identity, highest scoring template used in model, number of models = 1). For each antibody, the Fv region was modeled using templates of highest identity for framework and complementarity-determining regions (CDRs). Follow ing structure template searching of the PDB database by framework and CDR regions similarity and identity (Kabat definitions), the highest scored structure template was used for homology modeling. One homology model was then builtfor each antibody. Structure preparation was performed using default settings for charge, protonation, rotomer and steric clash minimization, and energy minimization. On the prepared homology models, protein properties were calculated at pH 7.4 and pH 5.2. Protein properties of focus included Fv charge separation (between heavy and light domains), isoelectric point (pl, structure based), zeta potential, dipole moment, hydrophobicity moment, and mobility. Protein patches larger than 50 Angstrom area (hydrophobic, positive, and negative) were also calculated and visualized. Homology models are represented as ribbon structures with surface patches colored by hydrophobic in green, positive in blue, and negative in red.

[0103] Comparison of the anti-IL-4Ra mAb and ASA homology7models and calculated properties highlighted the high positive charge on the anti-IL-4Ra mAb Fv (+7) and resultant higher pl (8.7). Targeted mutant designs were generated to reduce charge patches and charge separation between variable domains. Specifically, residues in positive charge patches were mutated to negative and neutral residues and protein attributes were calculated to evaluate the effects on charge patch (reduced size) and protein properties (reduced Fv charge and pl). Triple and double mutants showed the largest effect on Fv charge and pl reduction. Single mutants were also explored but were anticipated to exhibit a small effect on overall protein characteristics.Example 1: ASA and the Anti-IL-4Ra mAb Possess Varying Degrees of Non-specific Uptake and Preclinical Clearance

[0104] The anti-IL-4Ra mAb is a therapeutic protein which lacked wild type mouse PK data due to the use of a murine surrogate mAb for previous preclinical rodent studies. To better understand the mechanisms driving the high rate of the anti-IL-4Ra mAb CL, PK studies were performed in wild type animals to compare the non-target mediated elimination of the anti-IL-4Ra mAb and a reference mAb, ASA. Single dose PK studies in wild type mice demonstrated the anti- IL-4Ra mAb possessed roughly 3-times higher CL than ASA (Table 1, Figure 1). This result was likely not due to the difference in IgG subclass between ASA and the anti-IL-4Ra mAb as past studies have reported no significant effect of IgG subclass on mAb linear PK (Grinshpun et al., (2021) “Identifying biophysical assays and in silico properties that enrich for slow clearance in clinical-stage therapeutic antibodies / ’ MAbs 13. 1932230 and Haraya et al.. (2021) “Estimation of Clearance and Bioavailability of Therapeutic Monoclonal Antibodies from Only Subcutaneous Injection Data in Humans Based on Comprehensive Analysis of Clinical Data,” Clin I’harmacokinei. 60 1325-1334, which are hereby incorporated by reference in their entirety). Rather, the observations made in wild type mice indicated a role for non-specific clearance based on the absence of target, which is due to negligible anti-IL-4Ra mAb cross-reactivity with murine IL-4Ra.Table 1: Human and rodent CL values for anti-IL-4Rot mAb and ASACLhuman for anti-IL-4Ra mAb is the estimated linear clearance obtained via pharmacokinetic modeling from Kakkar et al (2011). Murine parameter estimates reported with coefficient of variation. CLhuman, human clearance; N / A, not applicable; CL, central compartment clearance in wild type mice; CLD distribution clearance in wild type mice; VI , volume of central compartment; V2, volume of non-vascular compartment; ASA, anti-streptavidin antibody.

[0105] A previous analysis of the anti-IL-4Ra mAb Phase 1 and 2 studies demonstrated the anti-IL-4Ra mAb possessed non-dose proportionality in plasma exposure indicative of target- mediated dynamics at the evaluated doses. The clinical results were mathematically described via a two-compartment population PK model with parallel linear (target independent) and non-linear (target-mediated) CL pathways. From this report, the anti-IL-4Ra mAb was shown to display an estimated CLind of 10.5 mL / d / kg (Table 1), which is more than twice the average CLind value for the 64 clinical mAbs evaluated by Grinshpun et al. (4.84 ± 4.72 SD) (Kakkar et al., “Population PK and IgE pharmacodynamic analysis of a fully human monoclonal antibody against IL4 receptor,'’ Pharm Res 28 (2011) 2530-2542; Grinshpun et al.. “Identifying biophysical assays and in silico properties that enrich for slow clearance in clinical-stage therapeutic antibodies,” MAbs 13 (2021) 1932230, which are hereby incorporated by reference in their entirety). This linear CL was also significantly higher when compared to that of dupilumab, a human IgG4 monoclonal antibody binding the same target antigen as the anti-IL-4Ra mAb (human IL-4Ra) (Kovalenko et al., “Exploratory Population PK Analysis of Dupilumab, a Fully Human Monoclonal Antibody Against IL-4Ralpha. in Atopic Dermatitis Patients and Normal Volunteers,” CPT Pharmacometrics Syst Pharmacol 5 (2016) 617-624, which is hereby incorporated by reference in its entirety)- When taken together, these findings indicate target-mediated disposition was not the sole reason for the high CL of the anti-IL-4Ra mAb in humans.

[0106] To determine if non-specific endocytosis was contributing to the rapid CLind of the anti-IL-4Ra mAb, internalization kinetics of ASA and the anti-IL-4Ra mAb were conducted usingparental CH0-K1 cells. This line was chosen because it is highly amenable to flow cytometry, easy to grow, widely used, and non-human. Of the two mAbs tested, the anti-IL-4Ra mAb displayed pronounced endocytosis with biphasic time-dependent (Figure 2B) and linear concentrationdependent (Figure 2C) uptake under the conditions tested. ASA was nearly undiscemible from untreated controls such that both populations exhibited strong overlap (Figure 2A, bottom left panel). Additionally, very low signal was observed for both mAbs at 4°C that indicated negligible cell surface binding respective to the degree of internalization at 37°C. Confocal microscopy performed following a separate 60 min uptake study at 37°C using 100 pg / mL of ASA or the anti- IL-4Ra mAh confirmed extensive internalization of the anti-IL-4Ra mAh when compared to ASA and untreated controls (Figure 2D). These combined observations supported extensive non-specific endocytosis of the anti-lL-4Ra mAb relative to ASA in CHO-K1 cells. Furthermore, non-specific uptake of ASA was essentially undetectable when compared to untreated controls. This result highlights a detectable liability that is proposed to be a contributing factor for the rapid elimination of the anti-IL-4Ra mAb in vivo.Example 2: Development of a Quantitative Flow Cytometry Method to Enable Between Day Comparisons

[0107] One of the limitations of the initial endocytosis assay iteration was the inability' to compare the extent of uptake across experimental days. This was due to the readout being median fluorescent intensity, which is a variable impacted by an assortment of factors that include different voltage settings on the cytometer and using separate instruments across days. To standardize the assay to permit cross-day comparisons and reliable databasing of molecule attributes, the quantitative flow cytometry method was optimized using commercially available anti-mouse IgG microspheres. Each population of these beads binds a pre-determined amount of antibody that can be used to compute the antibody binding capacity (ABC) of a cell population (i.e., the amount of antibody associated with a single live cell event). Each bead standard was saturated under the experimental conditions. Histograms for each bead population overlapped when stained with either 5 or 10 pg / mL of an anti -human Fc mouse mAb directly conjugated with Alexa Fluor 647 (Figure 3 A). Both concentrations resulted in standard curves with similar extents of linearity (r2= 0.999).

[0108] A concentration of 5 pg / mL was selected, because this was the same concentration used to stain CHO-K1 cells post-uptake. Assay reproducibility was evaluated by conducting endocytosis studies using the anti-IL-4Ra mAb or ASA over three separate days, two different cytometers, and two different scientists (n = 20 total samples). Antibodies were incubated with cells at 37°C for 60 min at 100 pg / mL based on the initial kinetic characterization (Figure 2). The 60 min incubation time was chosen to enable the positive detection of any future mAbs displayingintermediate non-specific uptake that may not be as rapid or to the same extent as that of the anti- IL-4Ra mAb. Under these conditions, a mean ABC of 40,839 (SD 6614. CV 16.2%) for the anti- lL-4Ra mAb and a mean ABC of 921 (SD 315, CV 34.2%) for ASA was measured (Figure 3B). Population histograms were included for representative ASA and untreated samples to demonstrate the significant amount of signal overlap, signifying the weak internalization of ASA in CHO-K1 cells (Figure 3C). A schematic overview of the assay is included in Figure 3D.Example 3: High Non-specific Adsorptive Endocytosis is Driven by Charge-Dependent Interactions and is Conserved Across Species and Cell Types

[0109] One possible mechanism leading to the high rates of the anti-IL-4Ra mAb uptake is charge. Differences in protein charge have been demonstrated to influence mAb PK, with positive and negative net charge extremes typically associated with the strongest impacts (see e.g., Liu et al., MAbs 13(1): 1993769 (2021); Datta-Mannan et al., MAbs 7(3):483-93 (2015); Igawa et al., Protein Eng. Des. Sei. 23(5):385-92 (2010); Boswell et al., Bioconjug. Chem. 21(12): 2153-63 (2010)). It was previously shown that charge effects in more specific structural areas like the CDR can influence non-specific mAb endocytosis, even without a measurable shift in isoelectric point (pl) (Datta-Mannan et al., MAbs 7(3):483-93 (2015)). To provide a better understanding of the relationship between these variables and NSE, structural modeling examining the charge distributions of the anti-IL-4Ra mAb and ASA was performed (Figures 4A, 4B). The pl of the anti-IL-4Ra mAb and ASA were calculated from their amino acid sequences as 8.75 and 7.24, respectively. Antibody modeling identified several charge patches on the anti-IL-4Ra mAb, including two positive patches in heavy chain CDRs. The anti-IL-4Ra mAb had a higher calculated charge of its Fv (+7) compared to the ASA Fv (+2).

[0110] The extra positive charge patches on the anti-IL-4Ra mAb together with its higher Fv charge thus may have contributed to its high non-specific uptake and CLind. To confirm this, point mutations were performed on the anti-IL-4Ra mAb within its CDRs and the framework of the light chains to mitigate its positive charge attributes. Mutations were placed at amino acid positions either exhibiting clusters of positive charge and / or critical for driving the net charge of the protein (Table 2). ABC values obtained via endocytosis assays in CHO-K1 cells demonstrated each mutation resulted in a substantial reduction in non-specific adsorptive endocytosis (Figure 4C). These findings support positive charge exposed at the surface of the anti-IL-4Ra mAb as a key factor causing the elevated rates of non-specific adsorptive endocytosis.Table 2. Series of Anti-IL-4Ra mAb Point Mutants to Modify ChargeBolded letters correspond to mutant labels within Figures. VH, variable heavy chain; VL, variable light chain[OHl] Non-specific endocytosis assessments were performed on a separate panel of preclinical mAbs (mAbs B1-B5) targeting a G-protein coupled cell surface receptor. The mAbs possessed similar pl values and an expected inability’ to bind to murine target, but dramatically different CLind in wild type mice (Figure 5, Tables 3 and 4). Basic local alignment search tool (BLAST) analysis identified that the mAbs differed at amino acids 27 (light chain), 54, and 56 (both on the heavy' chain) within their CDRs. The amino acid variance would lead to anticipated charge differences of mAb Bl possessing the lowest positive charge and mAb B5 the highest. CHO-K1 endocytosis results aligned with this trend where mAb Bl exhibited the lowest ABC and mAb B5 the highest (Figure 6A). Experiments repeated in green monkey kidney epithelial cells (Vero cells, Cercopithecus aeihiops). a cell unrelated to CHO-K1 in species (Cricetulus griseus) and tissue origin (kidney vs. ovary), directly agreed with results using CHO-K1, which confirmed internalization was non-specific (Figures 6B, 6C). Additionally, uptake studies performed at 4°C supported the lack of target expression in either cell line and also highlighted the magnitude of non-specific adsorptive endocytosis relative to cell surface binding. In vivo-in vitro correlations were then conducted using ABC readouts from either CHO-K1 or Vero cells at 37°C, which demonstrated a strong relationship with CLind, mouse (Figure 6D). Together, these results confirm exposed positive charge at the surface of therapeutic proteins results in higher rates of non-specific adsorptive endocytosis, reveal non-specific adsorptive internalization patterns are shared across distinctly different cell types, and infer the ability to inform on CLind using the currently disclosed method.Table 3. Properties of Preclinical inAb SeriesTable 4. Murine PK Parameter Estimates for the Preclinical mAb B1-B5 SeriesDiscussion of Examples 1-3

[0112] The studies herein have demonstrated the sensitivity of the cell based assay of the present disclosure to identify distinct rates of non-specific endocytosis into mammalian cells by different mAbs.

[0113] The kinetic characterization of ASA and the anti-IL-4Ra mAb support non-specific uptake of a mAb in a cell-based experimental platform lacking target for several reasons. The first was the linear relationship between the concentration of the anti-IL-4Ra mAb and the extent of its endocytosis. Second, a large discrepancy between the anti-IL-4Ra mAb cell-associated amounts at the separate temperatures was observed. Whereas 37°C allowed for cell binding, endocytosis, and trafficking, 4°C permitted only surface binding. The low signal at 4°C thus indicated the anti- IL-4Ra mAb internalization at 37°C as well as supported the conclusion for the absence of targetwithin CH0-K1 cells. Of note was the biphasic time-dependent uptake of the anti-IL-4Ra mAh. This same observation was made previously by others and has been attributed to rapid and constitutive analyte exocytosis following initial non-specific internalization (Besterman et al., Endocytosis: a review of mechanisms and plasma membrane dynamics, Biochem. J. 210 (1983) 1-13; Besterman et al., Exocytosis of pinocytosed fluid in cultured cells: kinetic evidence for rapid turnover and compartmentation, J. Cell Biol. 91 (1981) 716-727; van Deurs et al., Kinetics of pinocytosis studied by flow cytometry, Eur. J. Cell Biol. 34 (1984) 96-102; and Scharschmidt et al., Fluid phase endocytosis by cultured rat hepatocytes and perfused rat liver: implications for plasma membrane turnover and vesicular trafficking of fluid phase markers, Proc. Nat ’I Acad. Sci. U S A 83 (1986) 9488-9492, which are hereby incorporated by reference in their entirety). Therefore, the combined results with the anti-lL-4Ra mAb agree with past reports descnbing nonspecific kinetics of various solutes in several different cell types.

[0114] Aside from non-specific uptake, FcRn-facilitated recycling is another major component of mAb linear PK (Ovacik and Lin, Tutorial on Monoclonal Antibody Pharmacokinetics and Its Considerations in Early Development, Clin. Transl. Sci. 11 (2018) 540- 552 and L. Liu, Pharmacokinetics of monoclonal antibodies and Fc-fusion proteins, Protein Cell 9 (2018) 15-32, which are hereby incorporated by reference in their entirety). FcRn is crucial for the long half-lives of Fc-bearing therapeutics due to its pH-dependent association with the Fc region. Following endocytosis, FcRn binds to its ligands during endosomal acidification and then mediates ligand trafficking to the plasma membrane where dissociation from FcRn occurs due to the low affinity of the FcRn-Fc interaction at near neutral pH (Challa et al., FcRn: from molecular interactions to regulation of IgG pharmacokinetics and functions, Curr. Top. Microbiol. Immunol. 382 (2014) 249-272, which is hereby incorporated by reference in its entirety). Though FcRn is a vital aspect of mAb PK and variations in the preclinical mAb CL could possibly be caused by differences in murine FcRn binding, the efforts here were focused on determining the relationship between non-specific uptake and CL. However, as described herein (Examples 4-8) a combinatorial evaluation of non-specific and FcRn interactions for mAbs can provide additional insight into potential PK shortcomings.

[0115] The cell-based assay described herein is a useful in vitro tool to identify high risk compounds in preclinical large molecule development. The current CHO-K1 format has a moderate throughput where roughly 50 samples can be run per day by an individual user. This can be increased by the removal of the 4°C group, which was used here, in part, to help identify the presence of target binding. This group can be substituted by publicly available transcriptomic / genomic-level information to evaluate target expression.

[0116] Because many variables can influence mAb PK, a cell-based measure of nonspecific uptake can be used in conjunction with accompanying approaches to assist in preclinical candidate selection and de-risking strategies. The current technique provides a per cell quantitative output that can be incorporated into computational models to assist in a deeper, more biologically relevant assessment of preclinical compounds. For example, readouts from the current assay can directly yield analyte-specific, single-cell internalization rates for PK modeling efforts rather than inferring this parameter with biophysical techniques or from general measures of non-specific endocytosis. Furthermore, other cells can be used in place of CHO cells to tailor the internalization kinetic measurements to the tissue being modeled, which can provide a more representative level of site-specific information.Materials and Methods for Examples 4-8

[0117] Antibodies and Corresponding CL Values. ASAWT and ASAAAA were fully human IgGl mAbs raised against streptavidin that contained a wildtype Fc region or H310A,I253A,H435A (ASAAAA) mutations based on previous reports (Kim et al., (1999) Mapping the site on human IgG for binding of the MHC class I-related receptor, FcRn. Eur J Immunol 29, 2819-2825; Qiao et al.. (2008) Dependence of antibody-mediated presentation of antigen on FcRn. Proc Natl Acad Sci U SA 105, 9337-9342, which are hereby incorporated by reference in their entirety). The clinical mAbs comprised commercial antibody products, including an anti-CD20 antibody (mAb5), anti-PD-1 antibody (mAb3) and an anti-CD38 antibody (mAb2), and research analogs of several clinical antibodies, including an anti-PCSK9 antibody (mAb4), anti-IL-15 antibody (mAb6), IL-4R antibody (AMG-317), IL-12 / IL-23 antibodies (mAbl and mAb9), EGFR / HER3 antibody (mAb7), IL-2R alpha antibody (mAb8), and a phosphatidylserine antibody (mAblO) . The mAb analogs were produced in stably transfected CHO cells on the stable effector functionless Fc backbone as described previously (Chaudhury, et al. (2003) The major histocompatibility complex-related Fc receptor for IgG (FcRn) binds albumin and prolongs its lifespan. J Exp Med 197. 315-322. which is hereby incorporated by reference in its entirety’).

[0118] A range of fully human, wild type h!gG2 anti-IL-4Ra WT and M252Y / S254T / T256E (YTE) charge mutant mAbs were generated in addition to the mAbs of Table 2 to study the impact of Fv mutations as well as half-life extension (HLE) Fc-engineered YTE mutations. In addition, two h!gG2 control mAbs, ASAwr-hIgG2-DL650 and ASAwr-hIgG2- YTE were also generated for characterization. All mAbs were constructed by recombinant DNA technology and produced in stably transfected CHO cells using standardized protocols. The list of these mAbs along with their pl and net charge are provided in Table 8.

[0119] The human CL values for many of the clinical mAbs were obtained from either internal data. FDA monographs, or a previous report that estimated the CLind for a variety of mAbs (Grinshpun et al., Identifying biophysical assays and in silico properties that enrich for slow clearance in clinical-stage therapeutic antibodies, MAbs 13 (2021) 1932230, which is hereby incorporated by reference in its entirety). A body weight of 80 kg was used for normalization. Publicly available clinical data for some antibodies is limited. For these cases, the CLind was estimated via reported volume of distribution (VD) and terminal half-life (ti / 2p) values obtained under linear PK conditions using the equation below (Schoenwald, R. D. (2002) Basic Principles, in Pharmacokinetics in Drug Discovery and Development (Schoenwald, R. D. ed.), 1 Ed. pp 3-33, which is hereby incorporated by reference in its entirety): 0.693 x VD

[0120] Cell Cultures. Parental MDCK II cells (Sigma #MTOX1300) were maintained in 5% heat-inactivated FBS EMEM (ATCC #30-2003) and subcultured using 0.25% trypsin-EDTA. Maintaining subconfluency was found to be essential to proper monolayer formation. Optimal subculture conditions were determined to be a density of 2.5-3e6 cells per 75 cm2every 2-3 days to maintain them below 90% confluency. Parental MDCK II cells were seeded at 100,000 cells per well in flat-bottom 96-well plates (Coming #353072). Transfected MDCK II cells (see below) were seeded at 120,000 cells per well in the same plates. Monolayers for both cells were achieved in 2 days. CHO-K1 cells were seeded and grown as described in reference to Example 1-3 herein. All cells were cultured in a 5% CO2-95% O2 humidified incubator at 37°C. Negative mycoplasma was confirmed using the Lonza My coAlert detection kit (Lonza, Walkersville, MD #LT07-318). Cryopreservation media was fully supplemented growth media with 5% DMSO (v / v).

[0121] Plasmids & Transfection. The hFcRn construct was developed on a pcDNA3.1 vector, incorporating an N-terminal HA-tag followed by hFcRn, a G4S linker, and rnEGFP, conferring resistance to G418. The h(32m construct was developed on a pEF6 / V5-His vector containing a blasticidin resistance gene.

[0122] In a sterile 1.5 mL Eppendorf tube A, a molar ratio 1: 1 of the two plasmids with a total of 4 pg of DNA w as diluted in 250 pl of Opti-MEM I Reduced Serum Medium without serum and with gentle mixing. Lipofectamine 2000 (10 pL) was diluted in 250 pL of Opti-MEM I Medium and incubated for 5 minutes at room temperature. After the 5 min incubation, the diluted DNA was combined with the diluted Lipofectamine 2000 (total volume = 500 pl), mixed gently, and incubated for 20 minutes at room temperature. Five hundred microliters of the complexes were added dropwise to each well containing cells with gentle mixing by rocking the plate back and forth. Cells w ere incubated at 37°C in a humidified, CO2 incubator for 48 hours prior to testing fortransgene expression. hFcRn-GFP / hp2m-MDCK II were constantly maintained in selection media (1000 pg / mL G418. 7.5 pg / mL blasticidin) as transgene loss was observed if the cells were cultured in normal growth media.

[0123] Immunostaining with Flow Cytometry. Immunostaining was performed using the fluorescently conjugated antibodies denoted in Table 5. Parental and FcRn-GFP / p2m-MDCK II cells grown to confluency were resuspended in IX PBS containing Zombie UV fixable viability dye (BioLegend # 423108) at the manufacturer recommended conditions, then incubated for 20 min at room temperature. Cells were washed once in FACS buffer (IX PBS, 2% w / v BSA, ImM EDTA, 0.1% w / v sodium azide), then either (1) held on ice for cell surface staining or (2) fixed and permeabilized in the dark for 20 min at room temperature using the Cyto-Fast fix / perm buffer (Biolegend # 426803). Fixed cells were washed twice in IX Cyto-Fast Perm Wash solution (Biolegend # 426803). All cells were then transferred to V-bottom 96-well plates (Fisher #249944) and stained on ice in the dark for 30 min using the antibody amounts indicated in Table 5 (100 pL volumes). Either FACS buffer (unfixed cells) or IX Cyto-Fast Perm Wash solution (fixed cells) was used for staining. Cells were then washed three times with the respective buffers, then analyzed on a BD FACSymphony flow cytometer with an 18 color, 5-laser configuration (UV-355 nm, violet-405 nm, blue-488 nm, yellow / green-561 nm, red-637 nm) and a BD Biosciences High Throughput Sampler (Catalog #338301). Where required, spectral compensation was accomplished using single-stained UltraComp eBeads (Thermo #01-2222-42), ArC™ Amine Reactive Compensation Beads (Thermo # Al 0346), and / or AcGFP Flow Cytometer Calibration Beads (Takara, #632594) based on the vendor instructions.Table 5. List of antibodies and dilutions used for flow cytometry

[0124] Fluorescence-Activated Cell Sorting (FACS). Clone #8 FcRn-GFP / p2m-MDCK II cells displayed high transgene expression based on preliminary assessments and were therefore chosen to expand. Distinct GFP+and GFP" populations (corresponding to hFcRn expression) were initially observed. FACS was thus conducted to enrich the GFP+population. Clone #8 cells were subcultured, resuspended in azide free FACS buffer at 7e6 cells / mL, strained through a 70 pm filter 12mm x 75mm polystyrene tube, and sorted for GFP+cells within the single cell Clone #8population via a BD FACSMelody Cell Sorter (equipped with violet-405 nm. blue-488 nm, and red-640nm lasers). However, subsequent evaluations demonstrated inadequate hp2m expression (multiple populations). GFP+Clone #8 FcRn-GFP / p2m-MDCK II cells were expanded and sorted again for the highest 10% GFP+, hFcRn+, hp2m+single cells via cell surface staining with anti- hFcRn-AF647 and anti-hp2m-PerCP / Cy5.5 (Table 6) on a BDFACSAria Fusion cell sorter (18 color, 5-laser configuration (UV-355 nm, violet-405 nm, blue-488 nm, yellow / green-561 nm, red- 640 nm). Even after FACS and constant selection, -2-3% of the total FcRn-GFP / p2m-MDCK II cell population was consistently GFP". Therefore, internalization analyses using fluorescently conjugated proteins were done on GFP+cells.Table 6. ASA mAb PK Estimates following Single Intravenous Bolus in hFcRn Transgenic MiceAn average mouse body weight of 0.028 kg was used for normalization. Data w as fit using a two compartment model.

[0125] Fluorescent Conjugation. Human serum albumin (Sigma #A3782) and ASAWT (2 mg / mL final concentration per reaction; 1 mL volumes) were conjugated with DyLight 650-NHS (Thermo # 62265) at a 10: 1 proteimdye ratio in 50 mM borate buffer at pH 8.5 (Thermo # 28384) following the manufacturer’s protocol. Proteins were dialyzed in 7 kDa molecular weight cutoff cassettes (Thermo # 66372) against 3 L (IL increments, three total changes) of IX PBS at 4°C overnight. Final protein concentrations were obtained spectrophotometrically as described by the manufacturer.

[0126] ASA Internalization Studies with Flow Cytometry. Parental and FcRn-GFP / p2m- MDCK II cells were seeded into 96-well plates and grown to confluent monolayers as described above. On study days, cells were washed twice with pre-warmed pH 7.4 Ringer's solution (122.5 mM NaCl, 5.4 mM KC1. 1.2 mM CaC12, 0.8 mM MgC12. 0.8 mM Na2HPO4, 0.2 mM NaH2PO4, 5.5 mM d-glucose, and 10 mM HEPES), then incubated for 30 min at 37°C in pH 7.4 Ringer’s. For time dependent studies, cells were aspirated and immediately incubated with 10 or 100 pg / mLDyLight 650-conjugated ASAWT (ASAWT-DL650) in pH 5.8, 7.4, or 8.0 Ringer’s over 2 h. For ASAWT-DL650 and DyLight 650-conjugated human serum albumin (HSA-DL650) concentration dependent studies, cells were incubated at pH 5.8 for 20 min with or without excess unlabeled protein (50 mg / mL; IgG from human serum used for ASAWT-DL650, Sigma # 14506) to confirm receptor-mediated endocytosis. After the treatments were completed, cells were placed onto ice and washed four times with ice-cold pH 7.4 Ringer’s (200 pL / well / wash) and then incubated with 100 pL / well 0.25% EDTA-trypsin for 8-10 min at 37°C. Once cells were rounded, wells were first agitated by gentle yet rapid pipetting, then 100 pL of MDCK II growth media (no selection reagents) was added per well to inhibit the trypsin. Cells were transferred to V-bottom 96-well plates, centrifuged at 300 g for 5 min at 4°C. then stained with 0.5% v / v UV Zombie dye in IX PBS for 20 min on ice. Cells were washed once with 100 pL FACS buffer per well, centrifuged at 300 g for 5 min at 4°C, then fixed at room temperature in the dark with Cyto-Fast fix / perm buffer (50 pL / well). Fixed cells were washed twice in IX Cyto-Fast Perm Wash solution (150 pL / well), resuspended in 100 pL FACS buffer, then analyzed on a BD FACSymphony flow cytometer. Median fluorescent intensities of atargeted 10.000 single cell, live, GFP+(FcRn-GFP / p2m-MDCK II only) events were obtained via BD Diva software. Data files were analyzed using FlowJo (BD). Concentration-dependent, receptor specific uptake data (i.e. total signal minus the average (+) unlabeled conditions per concentration and untreated control median fluorescent intensities) was divided by the total incubation time and then fit in GraphPad Prism (Dotmatics, Boston, MA) using the Michaehs-Menten equation:Vmax is the maximum rate achieved within the experimental system, [S] is the concentration of ASAWT-DL650 or HSA-DL650, and Kmthe concentration achieving half-maximal velocity.

[0127] ASAwT-hIgG2-DL650 and ASAwT-hIgG2-YTE-DL650 concentration dependent studies were performed in a similar way as mentioned above except that cells were incubated at pH 5.8 for 30 min in both parental and hFcRn-GFP / p2m-MDCK II cells. After analyzing the data files using FlowJo (BD), the concentration-dependent, receptor specific uptake data (i.e. total signal in hFcRn-GFP / p2m-MDCK II cells minus the average signal of parental MDCK II cells per concentration and untreated parental control cells' median fluorescent intensities) was divided by the total incubation time and then fit using the Michaelis-Menten equation as mentioned in the previous paragraph where Vmax is the maximum rate achieved within the experimental system, [S] is the concentration of ASAwr-hIgG2-DL650 or ASAwr-hIgG2-YTE-DL650, and Km the concentration achieving half-maximal velocity.

[0128] FcRn-ASAwr^Lysosome Imaging and Colocalization Analysis. Human FcRn- GFP / [32m-MDCK II cells were seeded into PerkinElmer CellCarrier Ultra 96-well plates (catalog # 6055302) as described above. On the day prior to imaging studies, confluent cell monolayers were incubated for 6 h with a 10 kDa Texas Red-conjugated dextran (Thermo # DI 828) at 0.25 mg / mL in selection media. The dextran concentration was based on a pilot titration study the week before. Following the incubation, cells were washed thrice in HBSS++and then incubated in selection media overnight. This approach would allow internalized dextran to accumulate within, and therefore label, lysosomes. The following day, time-dependent ASAWT-DL650 uptake studies (10 and 100 pg / mL) were conducted as described in “ASA Internalization Studies with Flow Cytometry" in pH 5.8 or 7.4 Ringer's. After incubations were completed, FcRn-GFP / [32m-MDCK 11 cells were washed four times with ice-cold pH 7.4 Ringer’s, then fixed in 4% v / v paraformaldehyde in IX PBS for 15 min at room temperature. Cells were washed twice with IX PBS after fixing, then stained with 1 pg / mL Hoechst (Thermo #H1399) diluted in IX PBS for 30 min at room temperature. Cells were washed twice with IX PBS, then stored at 4°C with 200 pL IX PBS per well. Images were captured the following day on an Opera Phenix High Content Screening system (PerkinElmer) using a 63x water objective with Z-stack acquisition. Mean fluorescence intensity (MFI) measurements on the channel for ASAWT-DL650 were analyzed using Columbus software (PerkinElmer). An analysis sequence was prepared to segment the cytoplasmic region of individual cells and collect MFI across the different time points and treatments. Colocalization image analysis was performed via Acapella software (PerkinElmer). Briefly, the Hoechst channel was used for nuclear and cytoplasmic segmentation. Colocalization was measured in the cytoplasmic region between the different channel pairs (ASAwr-DL650, hFcRn-GFP, and lysosomes loaded with Texas Red-dextran) using the Pearson's correlation coefficient. The Costes correction method was used to help distinguish labeled regions from background. MFI and colocalization metrics were plotted using GraphPad software.

[0129] Immunofluorescence and Confocal Microscopy. Parental and hFcRn-GFP / p2m- MDCK II cells were seeded and grown to confluent monolayers in PerkinElmer CellCarrier Ultra as above. Cells were washed twice with IX PBS. fixed in 4% PFA (IX PBS), and blocked / permeabilized for 1 h in IX Cyto-Fast Perm Wash solution at room temperature. All subsequent volumes were 100 pL / well. An anti-FcRn antibody (Sigma #HPA012122, 0.2 mg / mL stock concentration) was added at a 1:5000 dilution in IX Cyto-Fast Perm Wash solution and incubated for 1 h at room temperature. Cells were washed twice with IX Cyto-Fast Perm Wash solution, then incubated for 1 h at room temperature with a goat anti-rabbit secondary antibody conjugated with Alexa Fluor 594 (Thermo #A11037, 2 mg / mL stock concentration, 1 : 1000 dilutionin IX Cyto-Fast Perm Wash solution). Cells were washed twice with IX Cyto-Fast Perm Wash solution, then stained for 30 min at room temperature using 1 pg / mL Hoechst plus 2 pg / mL CellMask Blue (Thermo # H32720), both diluted in IX Cyto-Fast Perm Wash solution. Cells were washed twice with IX PBS and imaged on an Opera Phenix High Content Screening system using a 40x water objective with Z-stack acquisition

[0130] Human FcRn mAb Recycling Study for Examples 5 and 6, and Figures 10 and 11. Parental and FcRn-GFP / |32m MDCK II cells were seeded at le5 cells and 1.2e5 cells per well, respectively, in 200 pL of their individual growth media into 96-well plates and cultured for 48 hrs. Cell media was aspirated, and cells were then washed twice wi th pre-warmed (37°C) Ringer’s solution at pH 7.4 followed by an equilibration for 30 min in Ringer's pH 7.4 supplemented with MEM nonessential ammo acids IX (Cat# 25-025-C1. Coming™) and sodium pyruvate ImM solution (Coming #25-000-CI); referred to as Ringer’s++. After, wells were aspirated, and cells incubated with mAb solutions prepared at 666.67 nM (100 pg / mL) in pH 5.8 and pH 7.4 Ringer’s++for 2 hours at 37°C. This is referred to as the Load Phase. Following incubations, cells were washed at 20°C 4X with 200 pL per well with room temperature Ringer’s pH 7.4. Then, serum-free EMEM growth media (150 pL / well) was added to cells and the plates were incubated at either 37°C or 4°C for 4 hours. This is denoted as the Recycling Phase. As there was no active recycling at 4°C, the residual data at 4°C signified the uptake process following the load phase using 100 pg / mL mAbs for 2 h at 37°C. Next, the supernatants (recycled samples) were collected in 96 deep well plates (Axygen, #P-DW-500-C) and stored at -80°C until analyzed. Subsequently, the cells were washed for an additional 4X with 4°C Ringer’s on ice. The cells were then lysed with 150 pL / well ice-cold MOPS lysis buffer pH 7.4 (20 mM MOPS in 0.1% (v / v) Triton X-100) containing EDTA-free protease inhibitors. To prepare the lysis buffer, 1 EDTA-free mini cOmplete tablet containing protease inhibitor cocktail (Roche #04693159001) was added to each 10 mL of MOPS lysis buffer and vortexed. The plates w ere incubated for 30 minutes at 4°C on ice with shaking at 600 rpm, and then pipeted up and dow n 30 times to homogeneously dislodge the cells from the botom of the plates using a multi-channel pipettor. The lysates were collected in 96 deep well plates and stored in -80°C until further analysis.

[0131] Human FcRn Recycling Study of Anti-IL-4Ra WT and YTE Mutant mAbs for Example 8 and Figure 17: This human FcRn recycling study was conducted at 37°C for loading anti-IL-4Ra WT and YTE mutant mAbs as there was no active recycling observed at 4°C in the previous recycling study. Parental and FcRn-GFP / [32m MDCK II cells were seeded at 75,000 and 85.000 cells per well, respectively, in 200 pL of their individual growth media into 96-well plates and cultured for 72 hrs. Cell media was aspirated, and cells were then washed twice with pre-warmed (37°C) Ringer’s solution at pH 7.4 followed by an equilibration for 30 min in Ringer’s pH 7.4 supplemented with IX GlutaMAX™ supplement (Cat# 35050-061. Gibco™) and 1 mM sodium pyruvate solution (Coming #25-000-CI); referred to as Ringer’s+ / +. This differs from the Ringer’ s++described above via the lack of non-essential amino acids, which were not found to be needed for cell viability'. After, wells were aspirated, and cells were incubated with either anti-IL- 4Ra WT mutant mAb solutions prepared at 666.67 nM (100 pg / mL) or anti-IL-4Ra YTE mutant mAb solutions prepared at 166.67 nM (25 pg / mL) in pH 5.8 Ringer’s+ / +for 2 hours at 37°C. This incubation period is referred to as the uptake phase, which is a net result of both uptake and recycling taking place during this load phase. Following incubations, cells were washed four times on ice with 200 pL per well ice-cold Ringer’s pH 7.4 solution. The cells were then lysed with 200 pL / well ice-cold lysis buffer pH 8.0 (10 mM TrisHCl pH 8.0. 1% 1GEPAL CA-630, 0.5% sodium desoxycholate, 0.1% SDS, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA) containing EDTA-free protease inhibitors (referred to herein as ULS lysis buffer). To prepare the final ULS lysis buffer, 1 EDTA-free mini cOmplete tablet containing protease inhibitor cocktail (Roche #04693159001) was added to each 10 mL of ULS lysis buffer and vortexed. The plates were incubated for 30 minutes at 4°C on ice with shaking at 600 rpm, and then pipetted up and down 50 times in a Tecan Freedom EVO 200 liquid handler controlled by Freedom EVOware® version 2.8 software to homogeneously dislodge the cells from the bottom of the plates. The lysates were collected in 96 deep well plates and stored in -80°C until further analysis. These samples constituted the uptake fractions for the FcRn scoring. To obtain the recycling samples, the exact uptake procedure mentioned above was followed except that after the load phase, cells were washed at 20°C 4X with 200 pL per well room temperature Ringer’s pH 7.4. Then, serum-free EMEM grow th media (200 pL / well) was added to cells and the plates were incubated at 37°C for 4 hours. This was denoted as the recycling phase. Next, the supernatants (recycled samples) were collected in 96 deep well plates (Axygen, #P-DW-500-C) and stored at -80°C until analyzed. Subsequently, the cells were washed for an additional 4X with 4°C Ringer’s on ice. The cells were then lysed the same way as mentioned above after adding 200 pL / well ice-cold ULS pH 8.0 lysis buffer. The lysates were collected in 96 deep well plates and stored in -80°C until further analysis. These lysates are denoted as residual samples.

[0132] Recycling Assay Sample Analysis by MesoScale Discovery Immunoassay. Electrochemiluminescence (ECL) immunoassays were performed using the MesoScale Discovery' (MSD) Sector S 600 instrument (MesoScale Diagnostics, Rockville, MD, USA) similar to a previously developed protocol (91). First, the plates stored in -80°C were thawed on ice. The thawed lysis samples were centrifuged for 10 min at 1,000g at 4°C to remove cellular debris andthe supernatants were transferred to new796- w ell plates. For the MSD experiment, 96- ell MSD Gold® Streptavidin SECTOR® assay plates (MesoScale Diagnostics, #L15SA-1) were coated with 50 pL of a biotinylated, proprietary anti-human Fc antibody (clone 35; Ab35) at 2 pg / mL in blocking buffer (Blocker™ BLOTTO in Tris-buffered saline [TBS], Thermo #37530) for 1 hour with shaking at approximately 600 rpm at room temperature. Then, the plates were washed 4X with IX KPL wash buffer (LGC seracare) in a plate washer (Biotek 405 Select microplate washer). The standards, quality control samples (100, 10, and 1 ng / mL), and diluted recycling samples (1 : 1 with IX PBS) were prepared in the media matching experiments: serum-free EMEM media for the recycled samples, MOPS lysis buffer pH 7.4 (described above), or ULS lysis buffer pH 8.0 (described above) containing EDTA-free complete protease inhibitors for the residual samples. The lysis buffer mixture used to prepare standards and quality control samples for the residual sample analysis were generated using matched amounts of lysates based on the average protein content of the residual samples of that plate. The protein content w as measured by either Bradford (Examples 5 and 6, Bryniarski et al., Am. J. Physiol Renal Physiol. 315(5): Fl 191-F1207 (2018), which is hereby incorporated by reference in its entirety) or BCA following the manufacturer’s protocol (Example 8) assay. To generate the matrix for residual sample standards, an equal amount (~le6 cells) of hFcRn-GFP / hp2M MDCK II and parental MDCK II cells were grown to confluent monolayers on 100 mm round culture dishes (Coming #353003) and lysed with 1 m of either MOPS lysis buffer pH 7.4 or ULS lysis buffer pH 8.0 containing EDTA-free complete protease inhibitors. Then, the mixture was centrifuged for 10 min at 1,000g to remove cellular debris. The supernatant was collected and from this stock an appropriate amount of lysate was added to the lysis buffer mixture (v / v) of standards and quality control samples to match the average protein amount of the residual samples of a plate. The standards, quality control, and recycling assay samples (50 pL / well) were added to the plates and incubated for 2 hours with shaking at room temperature, followed by another 4X washing. After, plates were incubated with 50 pL of 2 pg / mL sulfo-tagged Ab35 (ruthenium) in BLOTTO blocking buffer as a detection reagent for 1 h at room temperature. Next, plates were washed 4X and read in the MSD instrument after adding 150 pL 2X MSD read buffer in each well (diluted using DI water from 4X MSD read buffer T with surfactant, cat#R92TC-l, MesoScale Diagnostics). The analyte concentrations were interpolated from a sigmoidal four-parameter least squares fit of antibody-specific standard curves in GraphPad Prism within the same plates and validated using the quality' control samples.

[0133] Human FcRn Recycling Efficiency Metric (FREM) Score. The recycled, residual, and uptake (4°C recycling phase) concentrations (Examples 5 and 6) derived from theinitial pH 5.8 load phase of mAbs were used to calculate hFcRn recycling efficiency metric (FREM) scores by the following formula:RxFREM score = - , when NU C > 1RAXRxFREM score = -A- X NUC , when 0 < NUC < 1 RAvHere. Rx is the recycled concentration of mAb ‘X’ from 37°C samples after a pH 5.8 load phase (average of the 4°C recycled concentration was subtracted from each sample readout to account for the active recycling process), and RAx is the residual concentration of mAb ‘X’ from the 37°C samples of pH 5.8 load phase (Grevys. et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746; Grevys, et al. (2018) A human endothelial cell-based recycling assay for screening of FcRn targeted molecules. Nat Commun 9, 621, which are hereby incorporated by reference in their entirety). Same equation was used to calculate FREM scores in Example 8, where Rx is the recycled concentration of mAb ‘X’ from 37°C samples after a pH 5.8 load phase and RAx is the residual concentration of mAb ‘X’ from the 37°C samples after a pH 5.8 load phase.

[0134] The uptake concentrations by the FcRn-GFP / hp2m MDCK II and parental MDCK II cells for each mAb were used to calculate the non-specific uptake coefficient (NUC) by the following formula:Uptake(both FcRn dependent and independent) of mAb ‘X’ in FcRn-GFP / hp2m MDCK II cells and UDCK 11is the FcRn independent uptake by mAb ‘X’ in parental MDCK II cells (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746, which is hereby incorporated by reference in its entirety).When the NUC value was between 0 and 1 (0 < NUC <1), it indicated the cellular uptake process was dominated by FcRn-independent non-specific endocytosis. In these cases, the FREM score was multiplied by the NUC. When NUC >1, it indicated FcRn-mediated uptake of mAb ‘X’ with negligible non-specific endocytosis. For these mAbs, the FREM score wasn’t multiplied by the NUC value. GraphPad Prism software was used to identity7the outliers of the assays by using the Grubbs’ method when n=4 (mAb panel) and / or Rout method (when n=8; ASA hlgGls and anti- IL-4Ra WT and YTE mutant mAbs) with an alpha value of 0.05.

[0135] CHO-K1 Non-specific Uptake Studies. Non-specific endocytosis studies were conducted at pH 7.4 or 5.8 as described above in Examples 1-3. The antibody binding capacity7(ABC) value is a quantitative technique to facilitate inter-day comparisons between mAbs via the generation of a standard curve utilizing IgG-binding microspheres.

[0136] ASA and Anti-IL-4Ra Mutant mAbs PK Study and Analyses. Mice were housed in groups at an AALAC, International accredited facility7. Animals were cared for in accordance with the Guide for the Care and Use of Laboratory Animals, 8thEdition. All research protocols were reviewed and approved by the Amgen Institutional Animal Care and Use Committee. Male homozygous Tg32 (Strain # 014565), homozygous Tg276 (Strain # 004919), and immunodeficient homozygous hFcRn SCID Tg32 (Strain # 018441) mice, 6-8 weeks old, were purchased from Jackson Laboratory7(Bar Harbor, MA). ASAwr (1 mg / kg), ASAYTE (1 mg / kg), and ASAAAA (3 mg / kg), anti-IL-4Ra WT and YTE mutant mAbs (each at 1 mg / kg) were administered as an intravenous bolus at the noted doses via the lateral tail vein. Blood specimens were collected at various times post injection, incubated at ambient temperature for approximately 20 minutes or until fully clotted, and then centrifuged to separate the serum. All serum specimens were stored at -70°C (± 10°C) until used in analytical assays.

[0137] Quantitation of proteins in mouse serum was performed via electrochemiluminescent immunoassays on the MSD Sector 600 instrument using the Ab35 anti-human Fc antibody as both the capture and detection reagent similar to a past report (Poon- Andersen et al., (2022) Development of an immunoassay for aglycosylated murine IgGl in mouse serum via generation of a specific tool antibody. Bioanalysis 14, 581-588, which is hereby incorporated by reference in its entirety). In all assays, the analyte serum concentrations were interpolated from standard curves using the corresponding analyte prepared in pooled mouse serum in WatsonLIMS software.

[0138] Linear CL was expected for the ASA variants due to the lack of target expression in mouse. Serum concentration data was fit individually using a two compartment dispositional model possessing central and peripheral compartments (Tang et al., (2004) Pharmacokinetic aspects of biotechnology products. J Pharm Sci 93, 2184-2204, which is hereby incorporated by reference in its entirety ). Parameterization included first-order elimination from the central compartment (clearance, CL), distribution clearance between the compartments (CLD), and volumes of the central (VI) and peripheral (V2) compartments. Modelling was conducted using the Ubiquity package workflow within R (v4.2.2).Example 4: Generation and Functional Characterization of MDCK II Cells with Stable Co-expression of hFcRn-GFP and h(I2m

[0139] MDCK II cells were stably transfected with constructs for h(32m and hFcRn-GFP. A clone co-expressing high amounts of cell surface h[32m and hFcRn-GFP was isolated byfluorescence-activated cell sorting (FACS). Protein expression was then confirmed in the resulting post-sorted cell line (designated hFcRn-GFP / h[32m-MDCK II cells) via immunofluorescence with confocal microscopy (Figure 7A) and flow cytometry (Figures 7B, 7C).

[0140] Because FcRn exhibits pH-dependent affinity for its ligands, internalization studies were conducted at varying pH values to evaluate the function of the co-transfected hFcRn- GFP / h(32m complex. Parental and hFcRn-GFP / hp2m-MDCK II cells were incubated with 10 or 100 pg / mL of a wild type hlgGl anti-streptavidin antibody fluorescently conjugated with DyLight 650 (ASAWT-DL650) at pH 5.8, 7.4, or 8.0 for increasing time periods. Analysis was performed via flow cytometry (Figure 13A). Only the hFcRn-GFP / hp2m-MDCK II cells displayed concentration and time dependent uptake in a pH-dependent manner whereby acidic conditions resulted in extensive ASAWT-DL650 endocytosis (Figure 8A and Figure 13B). This behavior is consistent with the enhanced affinity of FcRn for its ligands within acidified endosomes. Additionally, no difference in ASAWT-DL650 uptake was observed in parental MDCK II cells at any pH, thus indicating an undetectable impact on endocytosis at the varying pH levels (Figure 8A). Therefore, under the tested conditions, these results demonstrate hFcRn-dependent ASAWT- DL650 endocytosis in the co-transfected MDCK II cells that decreased with increased media pH. Additionally, non-specific uptake w as observed as the endocytic mechanism of IgG within parental MDCK II cells.

[0141] To provide further evidence for FcRn-mediated endocytosis, concentrationdependent uptake studies using ASAWT-DL650 or fluorescently labelled human serum albumin (HSA-DL650) were performed at pH 5.8. Both compounds were competitively inhibited by excess unlabeled protein (Figures 8B and 8C). The Michaelis-Menten equation was used as a function to fit the specific uptake data, and concentrations achieving half-maximal uptake within the experimental system were estimated as 50 pg / mL [95% confidence interval] for ASAWT-DL650 and 44 pg / mL for HSA-DL650. These equate to approximately 333 pM ASAWT-DL650 and 660 pM HSA-DL650. Furthermore, non-specific internalization was observed to be linear with concentration, which is in-line with non-specific internalization kinetics (Figures 8B and 8C, bottom plots). When taken together, this data indicated (1) functioning hFcRn-GFP / h[32m in cotransfected MDCK II cells and (2) non-specific endocytosis of albumin and ASAWT in the absence of target receptor.

[0142] The next objective was to assess the cellular localization and trafficking behavior of hFcRn-GFP. To do so, a concentration-, pH-, and time-dependent uptake study was performed in hFcRn-GFP / h[32m-MDCK II cells with high content confocal microscopy. ASAWT -DL650 (10 or 100 pg / mL) internalization was measured at pH 5.8 and 7.4. Imaging was done on fixed cellsfollowing incubations of 15, 30, 60, or 120 mins. Lysosomes were labeled by preloading cells the day before with a 10 kDa Texas Red conjugated dextran (Figure 9A). Intracellular ASAWT-DL650 imaging results were consistent with those reported in Figure 8 in which time-, concentration-, and pH-dependence was quantified (Figure 9B). Colocalization analysis via the Costes’ method for image signal thresholding for hFcRn-GFP, ASAWT -DL650, and Texas Red dextran (i.e. lysosomes) resulted in Pearson correlation coefficient scores indicative of apparent colocalization only for hFcRn-GFP and ASAWT-DL650 (Figure 9C). Minimal lysosomal colocalization was observed for hFcRn-GFP and ASAWT. These findings agree with past evaluations on the behavior of GFP- tagged FcRn and provide further support for proper hFcRn-GFP and hp2m function in the stably transfected MDCK II cells.Example 5: Development of a Quantitative FcRn Recycling Assay using hFcRn- GFP / hp2m-MDCK II Cells

[0143] The hFcRn-GFP / hp2m-MDCK II recycling assay described herein was developed in standard 96-well plates optimized by the ASAWT-DL650 kinetic analyses (Figure 10A). Following a 2 h load phase at pH 5.8 or 7.4, cells were washed and then incubated in serum-free cell media at 37°C or 4°C for 4 h. The 4°C group served as a control to provide the means to estimate the amount of mAb internalized during the load period as intracellular processing would be halted on ice. Additionally, this control group also allowed the demonstration of active recycling when compared to the parallel 37°C group (i.e., 4 h recycling period at 37°C). In summary, the workflow allowed for quantitation of both the specific (hFcRn-GFP / hp2m-MDCK II) and nonspecific (parental MDCK II) amount of test mAh internalized and recycled under various assay conditions.

[0144] Initial studies to assess assay performance were accomplished using ASAWT and a mutant with ablated hFcRn affinity (H310A,I253A,H435A; ASAAAA). FcRn recycling studies in parental and hFcRn-GFP / hp2m-MDCK II cells demonstrated: (1) higher ASAWT uptake / recy cling at pH 5.8 compared to pH 7.4, (2) no difference in ASAAAA uptake / recy cling at either pH condition when compared to parental MDCK II data, (3) higher recycling during the 4 h 37°C conditions vs. 4°C (when detectable uptake was observed), and (4) negligible non-specific uptake of any ASA hlgGl into parental MDCK II cells (Figure 10B). Of note, the very low recycling signal for ASAWT after a pH 7.4 load was not the result of inefficient hFcRn-mediated recycling, but rather due to minimal amounts of mAb entering the cell during the load phase.

[0145] Rodent PK studies were performed to characterize differences in total CL for the ASA mAbs (Figure 10C). For in vivo studies, hFcRn transgenic mice (Tg32) were selected to preserve the species-dependent interaction between the ASA hlgGl s and FcRn and because of pastreports on the fidelity of this preclinical model for translation of human mAh PK. ASAAAA exhibited higher systemic CL in Tg32 mice when compared to ASAWT (Table 6). Matching studies were performed in homozygous Tg276 hFcRn mice to confirm these results. Tg276 possess a separate promoter than that in the Tg32 and exhibit variations in hFcRn tissue expression but can provide an alternative humanized mouse model to delineate changes in hlgG disposition. Again, ASAAA demonstrated much faster CL than ASAWT (Table 6).

[0146] One goal of this project was to generate an in vitro assay that could rank a series of compounds to inform on mAb CLmd. A previous report outlined a metric that facilitated the calculation of hFcRn-mediated recycling efficiency in cultured endothelial cells. Their scoring system was recently expanded to also account for non-specific behaviors because the initial approach could not be used on its own to describe the cellular disposition of mAbs with extensive non-specific characteristics (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746; Grevys, et al. (2018) A human endothelial cell-based recycling assay for screening of FcRn targeted molecules. Nat Commun 9, 621. which are hereby incorporated by reference in their entirety). These published studies utilized pH 7.4 load conditions to recapitulate the physiological conditions within the vasculature. However, a similar scoring approach was applied here in the MDCK II cell studies using pH 5.8 load conditions to improve sensitivity of the ASAWT signal based on the internalization results (Figures 8A, 8B and 10B). The human FcRn recycling efficiency metric (FREM) scores in the current study were calculated as described above by multiplying the recycled fraction by the non-specific uptake coefficient (NUC) (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746, which is hereby incorporated by reference in its entirety). The NUC was only incorporated when an analyzed mAb exhibited elevated non-specific endocytosis in hFcRn-GFP / h(32m-MDCKII cells relative to the uptake in parental MDCK II. The FREM score for ASAAAA was negligible relative to ASAWT, highlighting the lack of hFcRn interaction in the ASAAAA mutant (Figure 10E). These results trended with the CL values obtained from both hFcRn transgenic mouse models (Figures 10C and 10D).

[0147] In summary, to validate the cellular system, both parental and transfected MDCK II cells were used to define the hFcRn-specific and non-specific trafficking of two hlgGl mAbs with vary ing hFcRn affinities. A scoring system that could readily differentiate ASAWT from ASAAAA was applied. Furthermore, calculated hFcRn recycling scores agreed with the CL values measured in the single dose PK studies in hFcRn transgenic mice.Example 6: Human FcRn Recycling Scores from MDCK II Cells Can Identify mAbs with Rapid Human CL

[0148] Next, the ability of the MDCK II cell system to detect mAbs with high non-target mediated CL (i.e. Cl.ind) was examined. A total of 10 mAbs possessing human CL values between 2.45 to 32.2 mL / d / kg were chosen (Table 7). All mAbs were assayed in parental and hFcRn- GFP / h|32m-MDCK II cells under identical conditions to the ASA hlgGl series (Figure 14). A clinical research analog mAb, mAbl, was chosen as the reference for all subsequent comparisons because it was well-behaved in the MDCK II cell studies (proper hFcRn interaction, low nonspecific uptake) and because it is an FDA approved drug with low CL (Figure 11 A, Table 7). As mentioned above for the ASA mAbs, the NUC metric was incorporated only when the NUC was below a value of 1 , which indicated non-specific uptake was occurring in excess to that internalized by hFcRn. FREM scores with means below 25% of mAbl were correlated with human CL values above 4.8 mL / kg / d, thus highlighting the potential of this approach to identify mAbs with PK liabilities (Figure 11B).Table 7. Human CL, MDCK II FcRn Recycling, and Biophysical Assay Metrics for the mAb Panel1CL and pl values determined from relevant FDA labels, journal publications, or internal data.*FREM and NUC scores of the antibodies were normalized to mAblExample 7: Distinct Mechanisms for High mAb CLjnd

[0149] An observation from the hFcRn recycling studies was the association between unspecific interactions (via the NUC values) and the final hFcRn recycling scores (Table 7), indicating a strong role for non-specific behavior in hFcRn recycling efficiency. Examination ofparental MDCK II uptake results revealed the anti-IL-4Ra mAb and mAb9 as unique outliers for non-specific characteristics. When compared to the mAbl analog, the anti-IL-4Ra mAb possessed significantly higher endocytosis into parental MDCK II cells at pH 7.4 (Figure 11 C, #### p < 0.0001). This is consistent with Examples 1-3 demonstrating extensive non-specific cellular uptake of the anti-IL-4Ra mAb at this pH value that may have contributed to its rapid CL in wild ty pe mice and humans. Large charge patches were identified within the anti-IL-4Ra mAb Fv region that were likely a determining factor for this observation. Thus, for the anti-IL-4Ra mAb, widespread non-specific endocytosis from the extracellular fluid (e.g. blood) may drive large amounts of mAb into non-target cell populations, which will increase the probability of intracellular catabolism.

[0150] mAb9 is an anti-interleukin- 12 / 23 antibody with aberrant non-specific and hFcRn interactions that result in its high rate of CL as reported in the literature. In contrast to the anti-IL- 4Ra mAb, mAb9 displayed elevated but not statistically significant parental MDCK II uptake at pH 7.4 relative to mAbl (which is also an anti-IL- 12 / 23 antibody) (Figure 11C). However, at pH 5.8, a significant increase in non-specificity for mAb9 was measured (ff , P < 0.001). It was recently shown that a high positive charge surface area within the mAb9 Fv region distinguishes its anti-IL12 / 23 variable domain from mAbl and is also sensitive to pH, where the separation increases under acidic conditions. This is supported by the current findings where the previously reported Fv domain charge difference between mAbl and mAb9 is likely promoting the elevated non-specific uptake at pH 7.4, becoming larger and significantly different at pH 5.8. Therefore, mAb9 possesses pH-dependent charge patches that may not only influence its association / dissociation from hFcRn and initial non-specific endocytosis rate at pH 7.4 but could also result in increased non-specific interactions within the endosome as it acidifies to impair mAb9 dynamics with hFcRn and decrease its recycling efficiency.

[0151] Furthermore, a small increase in the pH-dependent. non-specific uptake of mAbl in parental MDCK II cells was observed (Figure 11C, *p < 0.05). To better characterize these results across the entire mAb series, follow-up studies directly measuring mAb non-specific endocytosis in CHO-K1 cells (as described supra) were performed (Figure 1 ID). This method sees the incubation of parental CHO-K1 with the mAb(s) of interest at 37°C for 60 min. Analysis of cell-associated mAb (i.e. intracellular and surface) is accomplished via anti-human Fc immunostaining and flow cytometry. A standard curve is generated daily using quantitative microspheres that bind a fixed amount of detection antibody to enable inter-day comparisons (antibody binding capacity, ABC). The use of the CHO-K1 assay provides a means to recapitulate the potential for altered non-specific interactions within the endosomal compartments byperforming incubations at an acidic pH. Higher rates of cellular uptake using this method would suggest increased non-specific interactions at the tested pH value.

[0152] In CHO-K1 cells, elevated non-specific endocytosis was measured for mAb9 (p < 0.01) and the anti-IL-4Ra mAh (p < 0.0001) at pH 7.4 when compared to ASAwr (Figure 11D). At pH 5.8, every mAb displayed increased endocytosis when compared to pH 7.4 conditions, which indicated elevated non-specific interactions with decreasing pH. Similar to the observations in parental MDCK II cells (Figure 11C), mAb9 non-specificity increased to amounts comparable to the anti-IL-4Ra mAb only at pH 5.8. These results provide evidence for two different mechanisms for elevated CLind. The first is pronounced non-specific uptake from the extracellular fluid at pH 7.4 and high general non-specificily within the endosome that likely mitigates the mAb- FcRn interaction (e.g., anti-IL-4Ra mAb). The second (e g.. mAb9) is moderate non-specific endocytosis at pH 7.4 relative to alow CLind mAb due to elevated non-specific interactions outside the cell. Following endocytosis, these non-specific behaviors increase during endosomal trafficking that may cause either unintentional association with the endosomal membrane and / or an impaired endosomal interaction with FcRn that could both lead to increased intracellular catabolism. For the tested mAbs with higher CLind not possessing extreme degrees of these aspects, the rate of extracellular non-specific endocytosis and / or extent of acidic non-specificity likely outweigh the ability7of the mAb to efficiently interact with FcRn within the endosome, leading to diminished FcRn recycling efficiency (Figure 12).Example 8: Cellular hFcRn Recycling Scores can Successfully and Simultaneously Rank WT and Fc-Engineered mAbs in Order of Decreasing Murine CLind

[0153] One approach to decrease the CLind and extend the serum half-life of an Fc- interacting molecule (e.g., an Fc-biologic) is to enhance the affinity of the Fc region to FcRn. Doing so must selectively increase the Fc-FcRn binding at acidic pH values while still maintaining the lowest possible interaction at near-neutral pH to preserve the pH-dependent dissociation of the Fc- region from FcRn (DalFAcqua et al., J. Biol. Chem. 281(33):23514-24 (2006); Borrok et al., J. Biol. Chem. 290(7): 4282-90 (2015); Mackness et al., MAbs 11(7): 1276-1288 (2019)). The cellular FcRn recycling assay described herein provides a crucial in vitro tool to gauge the potential success of Fc engineering strategies. However, to properly assess therapeutic proteins with altered FcRn interactions, the assay conditions should be specifically optimized. Not doing so could result in the failure of the assay to correlate with in vivo behavior. For example, a previous study employing a cell based FcRn transcytosis assay was unable to successfully inform on the CLind of antibodies with enhanced FcRn binding with no clear explanation provided (Chung et al., MAbs 11 (5): 942-955 (2019)).

[0154] To demonstrate the feasibility of the cellular hFcRn recycling assay described herein to accurately inform on the CLind of Fc interacting molecules with enhanced FcRn affinity, the physicochemical diversity of the anti-IL-4Ra mAh panel detailed in Example 3 was expanded. In particular, additional single- and double-point mutants with higher positive charge attributes within the variable (Fv) domains relative to the quadruple mutants in Example 3 (Table 8) were generated. NSE measurements were made in parental CHO-K1 cells at both pH 7.4 and 5.8, which demonstrated a much higher spread of non-specific behavior than what was observed for the original panel (Figure 15A-15B). The M252Y / S254T / T256E (YTE) mutations (EU: European Union numbering) were then incorporated into each anti-IL-4Ra mAb variant as a model of increased FcRn affinity' with clinically proven half-life extension and decreased CLind (Dall’Acqua et al., J. Biol. Chem. 281(33):23514-24 (2006); Borrok et al.. J. Biol. Chem. 290(7): 4282-90 (2015); Mackness et al., MAbs 11(7): 1276-1288 (2019)). The addition of the YTE generally maintained the overall non-speci Pi city trends between mAb-YTEs, but led to increased NSE at pH 7.4 and decreased non-specificity at pH 5.8 when compared to the parent WT Fc mAbs (Figure 15A-15B).

[0155] Enhanced Fc-FcRn interaction strategies like the YTE mutations will lead to higher internalization in FcRn-expressing cells across acidic to near-neutral pH values with net extents driven by relative affinity changes over these conditions. This will lead to higher amounts of mAb- YTE entering hFcRn-GFP / hp2m MDCK II cells at any given concentration relative to a non- YTE mAb. This could saturate the per cell FcRn capacity and result in incorrect assessments whereby YTE mutants would falsely appear to behave worse than their matching WT parents following cellular hFcRn recycling studies. To offset this affinity' impact and equalize intracellular loading conditions, a concentration dependent endocytosis studies was conducted using DyLight 650 conjugated control mAbs ASAwr-hIgG2 and ASAwr-hIgG2-YTE at pH 5.8 (Figure 16A-16D). Uptake studies and the estimation of Kmand Vmax values were performed as described above. It was observed that the Km of the ASAwr-hIgG2-YTE-DL650 (15 [10, 21] pg / mL) was nearly 3.5 times lower than the Kmof ASAwr-hIgG2-DL650 (41 [26, 64] pg / mL). Based on this data, subsequent FcRn recycling studies with hFcRn-GFP / h[32m MDCK II cells utilized roughly 2 x Kmconcentrations for WT (100 pg / mL) and YTE (25 pg / mL) for acidic uptake incubations to ensure similar amounts of WT and YTE mAb were loaded into cells in an FcRn-dependent manner.

[0156] Cell-based hFcRn recycling studies were performed with 6 WT / YTE pairs of anti- IL-4Ra mAbs. The FREM score was calculated by analyzing uptake, recycled, and residual samples, and was based on either / both FcRn recycling efficiency (RE) and non-specific uptake coefficient (NUC) scores as outlined above (Figure 17A). The assay' was optimized tosimultaneously compare both Fc-engineered YTE and WT mAbs by normalizing all results to the anti-IL-4Ra EEES-YTE mAb.Table 8. Series of Anti-IL-4Ra WT and YTE mAb Point Mutants to Modify Charge

[0157] Higher non-spccificity and / or lower FcRn recycled efficiency of mAbs resulted in lower FREM scores, as exemplified by the WT anti-IL-4Ra-mAb. A single point mutation at the R20 position did not improve the FREM score of the 20S-WT mutant (0.006 ± 0.001) significantly compared to the WT (0.004 ± 0.001). However, with the R33S mutation at the heavy chain (HC) CDR1, the FREM score (0.155 ± 0.048) was improved compared to R20S mutation. Double point mutations at both the HC CDR1 and CDR3 drastically improved the FREM score of anti-IL-4Ra mAb-33S,l 10L (0.504 ± 0.085) compared to single point mutations at the mAb Fv regions. Anti- IL-4Ra mAb-SSLS mAb demonstrated slightly lower FREM score (0.400 ± 0.095) compared todouble point mutations. The most exposed charge remediated WT mAb mutant, anti-IL-4Rq mAb- EEES, exhibited a higher FREM score (0.584 ± 0.127) compared to all other WT mAbs. The relative FREM score improvement trends were conserved between the WT and YTE mutant mAb series. When tested alongside WT mAbs, the half-life extending YTE mAbs exhibited better FREM scores relative to their WT counterparts, except minimal improvement for SSLS-YTE mutant compared to its WT counterpart. While compared head-to-head, the anti-IL-4Rq mAb- EEES-YTE demonstrated the highest FcRn recycling efficiency metric score (1.000 ± 0. 196).Table 9. CLind from SCID hFcRn Tg32 Mice, FREM Scores from the MDCK II hFcRnRecycling Assay, and pl Values of the Anti-IL-4 Rq WT and -YTE Mutant mAbs.1 CLind was calculated from serum concentration time profile of mAh mutants using non- compartmental analysis (NCA).* FREM scores were normalized to anti-IL-4Ra mAb-EEES-YTE2pl values were collected from internal data.

[0158] A PK study was performed in male homozygous immunodeficient SCID hFcRn Tg32 transgenic mice. Due to the lack of mature T cells or B cells, this mouse model facilitated the ability to obtain the PK across a prolonged sampling time course without the interference of anti-drug antibodies. Serum concentration-time profiles of the mAbs are shown in Figure I7B. It was evident from the PK results that mAbs with higher NSE and lower FcRn recycling efficiencyhad overall poorer PK (e.g. anti-IL-4Ra mAb). The FREM scores were then utilized to rank order both the WT and YTE mAbs concurrently. We observed a very strong relationship between the CLind and FREM scores for all mAbs, with lower FREM scores being associated with higher CLind (Figure 17C).Discussion of Examples 4-8

[0159] The current work provides a quantitative, mechanistic examination of the endocytosis and intracellular handling of a panel of mAbs to describe differences in hFcRn- mediated recycling efficiencies. The experimental outline provided the means to simultaneously assess hFcRn-dependent endocytosis and recycling and non-specific endocytosis characteristics of evaluated mAbs within a cellular assay. The degree of non-specific interactions strongly dictated the recycling outcome, which was distinguished via the mass balance approach undertaken in this study. These findings highlight the need to account for non-specific endocytosis rates, total mAb internalization, and hFcRn recycling when studying mAb CLind.

[0160] mAb internalization kinetics were measured in parental and hFcRn-GFP / hp2m MDCK II cells using ASAWT as a model compound because it possessed low CLind in wild type mice and negligible non-specific endocytosis in mammalian cells. The current studies demonstrated very low time-dependent uptake of ASAWT-DL650 at pH 7.4 in parental MDCK II when compared to receptor-mediated conditions at pH 5.8 in transfected MDCK II cells. Concentration-dependent endocytosis in hFcRn-GFP / h02m MDCK II cells at pH 5.8 under competitive settings was directly proportional to the extracellular ASAWT-DL650 concentration, consistent with non-specific endocytosis for ASAWT in the absence of receptor-mediated dynamics.

[0161] These measurements were of importance for studying hFcRn-recycling efficiencies because it confirmed a mAb with low CLind should exhibit minimal non-specific uptake at pH 7.4. This would then lead to negligible amounts recycled if 100 pg / mL was used. The data herein indicated pH 5.8 promoted FcRn-mediated endocytosis in hFcRn-GFP / hp2m MDCK II cells and led to substantially more ASAWT-DL650 entering the cell when compared to that internalized via non-specific endocytosis alone. Concentrations necessary to achieve detectable recycling following a pH 7.4 load phase would be unreasonable in terms of material usage for ASAWT or any other mAb exhibiting low non-specific uptake. Therefore, these results informed the later recycling studies by indicating an acidic load would be required to achieve detectable concentrations for mAbs with limited non-specific endocytosis.

[0162] The ASAWT-DL650 kinetic studies also demonstrated a contribution of hFcRn for initial mAb internalization at pH 7.4. ASAwr-DL650 uptake into parental MDCK II cells waslower than transfected cells throughout each time point examined at pH 7.4, and pH 8.0 led to reduced ASAwr-DL650 endocytosis in hFcRn-GFP / h[32m MDCK II cells compared with pH 7.4.

[0163] While cell surface FcRn could contribute to the endocytosis of IgG outside acidic environments, that was not the focus of the current project. The data herein indicates that nonspecific endocytosis is the primary endocytic mechanism for IgG at near neutral pH in the absence of either target or IgG receptors that includes FcRn. The results with ASAWT-DL650 indicated FcRn can to some degree facilitate the internalization of IgG at pH 7.4, but results from the mAb panel recycling study suggested minimal FcRn involvement for internalization at this pH value. The extent of the contribution of FcRn at pH 7.4 in vivo remains unknown. Any future work assessing this question should consider utilizing other physiologically relevant cell systems.

[0164] The success of utilizing acidic incubations was confirmed with the clinically relevant mAb panel such that very low amounts of multiple mAbs were internalized into hFcRn- GFP / hp2m MDCK II cells at pH 7.4, resulting in essentially undetectable concentrations of recycled material following 100 pg / mL incubations. It is important to note that conducting cellbased FcRn recycling studies at near-neutral pH is more physiologic and can be performed, as we have demonstrated above. However, based on the internalization kinetics of ASAWT-DL650 (Figures 8 and 13), at least 30-fold higher concentration would have been required at pH 7.4 for mAbs with low non-specific endocytosis and low CLind to achieve similar internalization levels, and therefore assay resolution, of the same mAbs at pH 5.8. This observation and explanation is a key difference from the work described previously (Grevys, et al. (2022) Antibody variable sequences have a pronounced effect on cellular transport and plasma half-life. iScience 25, 103746; Grevys, et al. (2018) A human endothelial cell-based recycling assay for screening of FcRn targeted molecules. Nat Commun 9, 621, which are hereby incorporated by reference in their entirety).

[0165] Examination of the ASA hlgGl s established the specificity of the MDCK II hFcRn recycling assay. ASAAAA exhibited no discernible differences under any condition between parental and hFcRn-GFP / hp2m MDCK II cells, supporting the deletion of hFcRn binding affinity at the mutated amino acids. Furthermore, inclusion of several experimental groups for MDCK II cell studies fully delineated the mechanisms driving differences in mAb CLind. This framework facilitated the capture of mAb internalization, intracellular trafficking, hFcRn-mediated recycling, residual accumulation, and non-specific interactions. The in vitro hFcRn recycling score was modified from previous reports to provide a means to pool the cellular datasets into a cumulative metric to describe the hFcRn recycling efficiency. An average FREM score above 0.5 relative to the mAbl analog signified a mAb with attributes associated with low rates of elimination in human,which included low non-specific uptake and ability to effectively undergo hFcRn-mediated recycling. A hFcRn recycling score below 0.25 was strongly indicative of a mAb with CL exceeding 5 mL / d / kg. FREM scores between 0.25 and 0.5 suggested potential liabilities that may result in elevated CLind.

[0166] When tested in the in vitro cell-based recycling assay described herein, it was evident that the FREM scores of the YTE molecules were comparatively higher than WT mutants. In general, the target-independent PK profiles of these engineered mAb mutants were well- explained with in vitro cell-based assays discussed in this example. Accounting for both NSE and FcRn engagement via FREM scores strengthened the conclusions and increased the PK-predictive capacity for these engineered proteins. Furthermore, use of acidic conditions necessitated the optimization of the experimental workflow to accommodate the evaluation of Fc-engineered mAbs with increased FcRn affinities. This would have been required even if physiologically relevant endocytosis conditions were used (i.e. -pH 7.4) due to the inherent increase of FcRn affinity across a range of pH values (Dall’Acqua et al., J. Biol. Chem. 281(33):23514-24 (2006); Borrok et al., J. Biol. Chem. 290(7): 4282-90 (2015); Mackness et al., MAbs 11(7): 1276-1288 (2019)).

[0167] The results support the importance of non-specific and hFcRn interactions in determining mAb CLind while also providing further insight into the diversity7of mechanisms dictating these cellular processes (summarized in Figures 12A-12B). Here, a significant impact of pH on the non-specific interactions of all mAbs was demonstrated as indicated by higher rates of non-specific endocytosis into CHO-K.1 cells at pH 5.8 than 7.4. This observation suggests that internalized mAbs may display non-specific interactions at extents and forms different to those exhibited at the cell surface at a pH near 7.4. The isoelectric points (pl) of all tested mAbs exceeded 7.0 which would result in increased positive charge at an acidic pH (Table 7). This may have contributed to the universally elevated non-specific endocytosis due to an increased interaction with the negatively charged plasma membrane. However, pl alone cannot fully explain the large differences in non-specific endocytosis of the anti-IL-4Ra mAb or mAb9 relative to the other mAbs.

[0168] In terms of cellular consequences, the observed pH-dependent characteristics could lead to both detrimental non-specific associations with the endosomal membrane as well as with hFcRn itself that reduce hFcRn recycling efficiency and lead to higher rates of CLind (Figure 12). What can be currently concluded is the importance of incorporating proper experimental controls when assessing mAb cellular disposition and hFcRn interactions. The hFcRn recycling scores were obtained at a pH of 5.8 and utilized the NUC metric to assess the extent of non-specific uptake in relation to the mAb interaction with hFcRn. Where measurement of non-specific interactions aloneoffered partial insight (e.g. CH0-K1 endocytosis), the NUC provided a means to concomitantly define the interplay of mAb with hFcRn and non-specific interactions within the context of the same experimental settings. This was crucial in the ability to outline the relationship between in vitro performance and CLind.

[0169] Single parameter in vitro assessments to predict mAb in vivo behavior have oftentimes failed to directly correlate with mAb CLind. This may be due to their inability to fully recapitulate the biological dynamics important for mAb disposition or because they only focus on a single aspect. A key advantage to the use of a cellular hFcRn recycling assay described herein is its integration of multiple processes important for the disposition of mAbs. The results reported herein agree with the capabilities demonstrated by other cell-based assessments of IgG-FcRn for mAb CLind while further emphasizing the multifaceted relationship between cellular uptake and FcRn-mediated recycling. Cell-based systems offer their readouts within a biological framework, which has been directly shown as advantageous for studying FcRn-IgG and FcRn-Fc interactions compared to standard biophysical approaches. Subsequently, physiologically based PK models will play a crucial role for the extension of in vitro results to the patient by helping to describe the net impact of in vitro behaviors on mAb disposition. Combinatorial methodologies that utilize known drug-like attributes for mAbs will be critically needed to develop low CL Fc-HLEs as the biopharmaceutical industry transitions to more complex structures for their biologic pipelines.Example 9: High Non-specific Endocytosis Predicts Low Subcutaneous Bioavailability in Human

[0170] Subcutaneous dosing remains the preferred injection route for therapeutic proteins because it facilitates drug administration in more convenient physical locations for the patient. However, subcutaneously administered drugs must first be absorbed into the blood from the injection site. The fraction of the dose that reaches the systemic circulation relative to intravenous administration (subcutaneous bioavailability, FSQ) is dependent upon several factors that includes temperature, solubility7, lymph flow at the injection site, and tissue composition (Richter et al., (2012) Mechanistic determinants of biotherapeutics adsorption following SC administration. AAPS J. 14(3): 559-70, which is hereby incorporated by reference in its entirety). The complexity of these factors and the poor mechanistic understandings of their underpinnings has led to a lack of ability7to use in vitro techniques to inform on human FSQ.

[0171] As demonstrated herein, non-specific endocytosis and FcRn recycling are key determinants of the CLind for intravenously dosed monoclonal antibodies (mAbs). Defining the cellular mechanisms dictating these processes and establishing quantitative experimental workflows resulted in an enhanced ability7to identify therapeutic mAbs with rapid CLind in vitro.It was hypothesized that these same factors, particularly non-specific endocytosis. would also be important determinants for FSQ. This is because high rates of non-specific endocytosis at the injection site would result in elevated therapeutic protein catabolism, leading to less drug available for absorption into the circulation. Additionally, the mechanisms driving high non-specific endocytosis may also lead to other in vivo characteristics that are detrimental to FSQ, such as nonspecific interactions with the subcutaneous extracellular matrix. Thus, the use of a cell-based assay to quantify non-specific endocytosis was anticipated to be informative on human FSQ.

[0172] A panel of therapeutic mAbs with clinical pharmacokinetic (PK) data was obtained, of which 17 possessed estimates for target-independent clearance (CLind). The mAbs were binned into three groups: (1) CLind < 4.5 mL / kg / d following an intravenous (IV) dose in human, (2) CLind > 4.5 mL / kg / d, or (3) no CLind estimate obtained. A CLind of 4.5 mL / kg / d translates to an approximate terminal half-life of 10 days in humans for an 80 kg individual.

[0173] Antibodies were analyzed using the CHO-K1 assay, as described supra in Examples 1-3, to obtain cellular antibody binding capacities (ABCs) and a range of behavior was observed as shown in Figure 21 A. Assay thresholds were then generated to better interpret the data. The “Low-risk” group contains mAbs with low-risk of high CLind and / or low FSQ (due to non-specific endocytosis). It was defined as the upper bound of the 95% confidence interval of the mean ABC of all mAbs with FSQ above 50% and / or CLind < 4.5 mL / kg / d. The “Medium-risk” group contains mAbs that may or may not exhibit high CLind and / or low FSQ and was set as two standard deviations above the mean ABC mentioned directly above (i.e. mean + 2 SD). The “High-risk” group highlights mAbs with obvious non-specific endocytosis that will likely be detrimental to their disposition in human.

[0174] The mAbs were then grouped by their performance in the CHO-K1 assay (i.e. low, medium, or high risk) and plotted against their respective CLind values. Figure 2 IB demonstrates that an increased risk as defined via higher rates of non-specific endocytosis was an indication for not only higher CLind following IV administration, but also decreased FSQ in human. Of note, every mAb characterized as high-risk exhibited CLind in human above 4.5 mL / kg / d, FSQ below 50%, or both.Example 10: Non-Specific Cellular Endocytosis Can Identify Therapeutic Proteins with High Target Independent Clearance Across a Broad Range of Structural Modalities

[0175] Multi specific, engineered proteins (MS-ePRs) are protein therapeutics capable of simultaneously engaging two or more distinct targets. These compounds have provided new mechanisms to treat various diseases previously unattainable with monoclonal antibodies (mAbs).They span a broad range of molecular formats that can be tailored for their intended therapeutic use, such as manipulated structures to achieve optimal efficacies (Deshaies R., Nature 580(7803): 329-338 (2020); Labrijn et al., Nat. Rev. DrugDiscov. 18(8):585-608 (2019); Husain and Ellerman, BioDrugs 32(5):441-464 (2018); Spiess et al., Mol. Immunol. 67(2 Pt A): 95-106 (2015); Dickopf et al., Comput. Struct. Biotechnol. J. 18: 1221-1227 (2020)). MS-ePRs can also include an Fc- region or utilize an immunoglobulin G backbone (mAb-ePR) to decrease clearance (CL) and reduce dosing frequency (Deshaies R., Nature 580(7803): 329-338 (2020); Labrijn et al., Nat. Rev. DrugDiscov. 18(8): 585-608 (2019); Husain and Ellerman, BioDrugs 32(5):441-464 (2018); Spiess et al., Mol. Immunol. 67(2 Pt A): 95-106 (2015)). The Fc portion facilitates interaction with FcRn, which recycles internalized immunoglobulin G. albumin, and Fc / albumin-fusion proteins via intracellular. pH-dependent binding to increase serum half-life (Challa et al., Curr. Top. Microbiol. Immunol. 382:249-72 (2014)).

[0176] The plasticity7of MS-ePRs not only broadens the therapeutic landscape but can also lead to development obstacles. Structural arrangements ideal for efficacy may also be challenging to synthesize. Certain engineered designs may be more prone to immunogenicity or encounter stability issues in vivo. Additionally, mAb-ePR drug disposition is much more complex compared to mAbs (Deshaies R., Nature 580(7803): 329-338 (2020); Husain and Ellerman, BioDrugs 32(5):441-464 (2018)). Target-mediated dynamics play a pivotal role in the overall pharmacokinetic (PK) profile of a mAb. Target engagement can result in target-mediated drug disposition (TMDD) whereby mAb CL becomes heavily entwined with the biology of its target, leading to dose-dependent changes in the disposition of the mAb (Ovacik and Lin, Clin. Transl. Sci. l l(6):540-552 (2018); Ryman and Meibohm, CPT Pharmacometrics Syst. Pharmacol. 6(9):576-588 (2017); Peletier and Gabrielsson, J. Pharmacokinet. Pharmacodyn. 39(5):429-51 (2012)). The degree of TMDD inherently grows with mAb-ePRs since at least two separate targets will be present, which can be further complicated by varying sites of their expression and / or binding affinities with the mAb-ePR (Deshaies R., Nature 580(7803): 329-338 (2020)).

[0177] Target independent mechanisms of mAb CL are also important for mAb-ePRs, which include FcRn-mediated recycling and non-specific adsorptive endocytosis into non-targeted cell types as described supra. The complex structures that can arise from protein engineering may lead to inferior interactions with FcRn or higher rates of non-specific adsorptive uptake by unintentional cell types that could result in rapid target independent CL (CLind) of the mAb-ePR (Datta-Mannan, Antibodies (Basel) 11(1):2 (2021); Datta-Mannan et al.. Biochemistry 58(28): 3116-3132 (2019)). However, direct cellular measures of mAb-ePR non-specific endocytosis havenot been reported, especially for a large array of proteins with broad structural characteristics and preclinical CLind values.

[0178] To address this limitation, two distinct panels of proteins were formed and the rates of their non-specific endocytosis was quantified by obtaining ABC values at pH 7.4 and / or pH 5.8 in CHO-K1 cells. The first protein series comprised 48 mAb-ePRs generated on the Amgen patented stable effector functionless Fc backbone (Liu et al., J. Biol. Chem. 295(5): 1876-1883 (2017)). The proteins possessed various molecular formats that included different targeting valency, structural arrangements of the binding moieties, and a wide range of CLind in Tg32 human FcRn transgenic mice following a single intravenous dose (Figure 22A). ABC values were calculated as described within Figure 3. The mAb-ePRs were placed into low, medium, or high- risk bins based off their ABCs by using the same numerical thresholds and approach as outlined supra and shown in Figure 21. Consistent with what was observed for the panel of clinical mAbs, 94% of the mAb-ePRs placed into the high-risk bin (i.e. exhibiting high non-specific endocytosis quantified via their ABCs) and 92% of mAb-ePRs in the medium-risk group exhibited CLind above 0.5 mL / kg / hr in Tg32 mice. Additionally, the high-risk group was able to successfully identify the mAb-ePRs with the highest CLind (Figure 22B).

[0179] The second group of proteins consisted of five separate protein structures fused bivalently to the human Fc-domain which constituted a total of 28 distinct Fc-fusion proteins bearing no semblance to human IgG or typical mAb-ePRs. The primary aim of this panel was to determine if non-specific endocytosis can be used to describe the CLind of any protein, not just those founded on an IgG or corresponding scaffold. The Fc-fusion series exhibited a range of CLind values in C57BL / 6 wild ty pe mice following a single intravenous dose of 2 mg / kg (Figure 23 A). CHO-K1 ABC values were obtained at both pH 5.8 and pH 7.4 and plotted against their CLind (Figures 23B and 23C). The pH 7.4 value directly measures non-specific endocytosis by mammalian cells. The pH 5.8 condition mimics the endosomal environment to quantify the pH- dependent shift in positive charge and non-specificity that would be anticipated to occur from endosomal acidification during intracellular trafficking as discussed supra. Proteins were placed into low, medium, and high-risk groups based on their extent of non-specific endocytosis (calculated using the ABC method) using the same ABC cutoff values for the mAb and mAb-ePR protein panels as described above. Ninety -four percent of proteins in the high-risk bin possessed CLind above 3 mL / kg / hr. Similar to the mAb-ePRs, the proteins with the highest CLind were successfully captured.

[0180] The utility of using the pH 5.8 surrogate measure for endosomal non-specificity in identifying high-risk CLind proteins was evaluated. All proteins exhibiting high non-specificendocytosis at both pH 5.8 and pH 7.4 displayed high CLind (Figures 23B, 23C). We then filtered the proteins by examining the pH 5.8 ABC values of only the Fc-Fusion proteins with low or medium-risk ABCs at pH 7.4 (Figure 23E). Of these, three proteins displayed pronounced increases in non-specific behavior at the lower pH with 2 possessing very high CLind in wild type mice (Figure 23F). This signifies that pH 5.8 conditions can be used to further identify therapeutic modalities at increased risk of elevated CLind.

[0181] Results from the mAb-ePR and Fc-Fusion proteins demonstrate that high nonspecific endocytosis provides an in vitro metric that is strongly associated with high CLind. The data presented herein further demonstrates the ability of the described cell-based endocytosis assay to identify structurally unrelated proteins with excessive CLind in rodents. Additionally, the described approach provides a unique and novel ability to quantify the pH-dependent changes in non-specific behavior of biologies. However, assessments of non-specific endocytosis alone are not enough to successfully identify all compounds with elevated CLind. This was observed in the low7and medium-risk groups, which contained several constructs with elevated CLind in rodents. As detailed above, there are additional factors driving CLind for mAbs. mAb-ePRs, and any other biologic modality. One key factor for Fc-bearing proteins is the FcRn interaction. Thus, combining the cell-based readouts for non-specific endocytosis, pH-dependent shifts in nonspecificity / charge, and engagement with FcRn greatly increase the ability to identify7high-risk biologies that interact with FcRn as well as guide protein engineering efforts to improve the PK attributes of prechmcal candidates.

[0182] Each reference cited herein is hereby incorporated by reference in its entirety for all that it teaches and for all purposes.

[0183] The present invention is not to be limited in scope by the specific embodiments described herein, which are intended as single illustrations of individual embodiments of the invention, and functionally equivalent methods and components are invention. Indeed, various modifications of the invention, in addition to those shown and described herein will become apparent to those skilled in the art from the foregoing description and accompanying drawings. Such modifications are intended to fall within the scope of the appended claims.

Claims

CLAIMS1. A method of predicting in vivo non-target mediated clearance of a biological molecule, said method comprising: providing a preparation of cells, wherein cells of the preparation do not express a target of the biological molecule; incubating the preparation of cells with culture media containing the biological molecule under conditions mimicking in vivo physiological conditions; determining an amount of the biological molecule taken up by the cells of the preparation after said incubating; and predicting in vivo non-target mediated clearance of the biological molecule based on said determining.

2. The method of claim 1, wherein the culture media during said incubating has a pH of about 7.0 to about 8.0.

3. The method of claim 1, wherein the culture media during said incubating has a pH of about 5.6 to about 7.0.

4. The method of claim 1 , wherein said incubating is carried out at 37°C for about 60 minutes.

5. The method of claim 1. wherein said determining comprises: permeabilizing cells of the preparation after said incubating; detecting the amount of biological molecule taken up by the cells of the preparation; and quantifying the amount of biological molecule taken up by the cells of the preparation based on said detecting, wherein predicting the in vivo non-target mediated clearance of the biological molecule is based on said quantifying.

6. The method of claim 5 further comprising: incubating, after said permeabilizing, the permeabilized cells with a detectable moiety that binds to said biological molecule, wherein the amount of detectable moiety bound to said biological molecule is detected.

7. The method of claim 5. wherein said quantifying comprises:comparing the detected amount of biological molecule taken up by the cells to one or more quantitative reference values.

8. The method of claim 7, wherein the one or more quantitative reference values comprise a calibration curve of biological molecule binding capacity.

9. The method of claim 6, wherein the detectable moiety is an anti-Fc antibody coupled to a label.

10. The method of claim 5, wherein said detecting is carried out using flow cytometry, immunoassay, or microscopy.

11. The method of claim 1, wherein the biological molecule is a human antibody, epitope-binding fragment of a human antibody, or a human antibody derivative.

12. The method of claim 1, wherein the biological molecule is a multispecific antibody.

13. The method of claim 1, wherein the biological molecule is a recombinant protein or fusion protein.

14. The method of claim 1, wherein cells of the preparation do not express an Fc receptor.

15. The method of claim 1, wherein cells of the preparation do not express human neonatal Fc receptor (hFcRn) and human 02m (h02m) complex.

16. The method of claim 1, wherein the preparation of cells is a preparation of cells selected from the group consisting of Chinese Hamster Ovary (CHO) Cells, Madin-Darby canine kidney (MDCK) cells, Vero cells, HUVECs. HEK293 cells, and primary endothelial cells.

17. A method of predicting in vivo non-target mediated clearance of an FcRn interacting molecule, said method comprising: providing a first cell preparation, wherein cells of the first preparation do not express hFcRn; providing a second cell preparation, wherein cells of the second preparation express a heterodimer of human neonatal Fc receptor (hFcRn) and human 02m (h02m), and wherein cellsof the first and second preparations do not express a binding target of the FcRn interacting molecule; subjecting the first and second cell preparations to first and second incubation periods, wherein said first incubation period comprises incubating the cell preparations with media containing the FcRn interacting molecule under acidic and / or non-acidic conditions, and wherein said second incubation period comprises incubating the cell preparations, after said first incubation, with media lacking the FcRn interacting molecule under non-acidic conditions; determining an amount of the FcRn interacting molecule taken up by cells of the first and second preparations after said subjecting to the first incubation period and / or the second incubation period; measuring an amount of the FcRn interacting molecule in media from the second incubation period collected at the end of the second incubation period; quantifying non-specific endocytosis and FcRn recycling of the FcRn interacting molecule based on said determining and said measuring, respectively; and predicting the in vivo non-target mediated clearance of the FcRn interacting molecule based on said quantifying.

18. The method of claim 17, wherein said first incubation comprises incubating the first and second cell preparations with media containing the FcRn interaction molecule under acidic condition and non-acidic conditions.

19. The method of claim 17, wherein said first incubation comprises incubating the first and second cell preparations with media containing the FcRn interaction molecule under acidic or non-acidic conditions.

20. The method of any one of claims 17-19, wherein said non-acidic conditions comprise a media pH of 7.0 to 8.0.

21. The method of any one of claims 17-19, wherein said acidic conditions comprise a media pH of 5.6 to 6.9.

22. The method of claim 17, wherein said first incubation period comprises incubating the first and second cell preparations with media containing the FcRn interacting molecule at 37°C for about 60 to about 120 minutes.

23. The method of claim 17, wherein said first incubation period comprises incubating the first and second cell preparations with media comprising a concentration of FcRninteracting molecule that does not saturate FcRn binding occupancy of cells of the second cell preparation.

24. The method of claim 23, wherein the concentration of FcRn interacting molecule in the media achieves between 25-90% FcRn binding occupancy of cells of the second cell preparation.

25. The method of claim 17, wherein said second incubation period comprises incubating the first and second cell preparations with media void of the FcRn interacting molecule at 37°C for about 3 hours to about 5 hours.

26. The method of claim 17, wherein said determining comprises: permeabilizing cells of the first and second cell preparations after said first incubation period and / or the second incubation period; detecting the amount of FcRn interacting molecule taken up by cells of each preparation; and quantifying the amount of FcRn interacting molecule taken up by cells of each preparation based on said detecting, wherein predicting the in vivo non-target mediated clearance of the FcRn interacting molecule is based on said quantifying.

27. The method of claim 26 further comprising: contacting, after said permeabilizing, the permeabilized cells with a detectable moiety that binds to said FcRn interacting molecule, wherein the amount of detectable moiety bound to said FcRn interacting molecule is detected.

28. The method of claim 26, wherein said quantifying comprises: comparing the detected amount of FcRn interacting molecule taken up by the cells to one or more quantitative reference values.

29. The method of claim 28, wherein the one or more quantitative reference values comprise a calibration curve of FcRn interacting molecule binding capacity.

30. The method of claim 27, wherein the detectable moiety is an anti-Fc antibody coupled to a label.

31. The method of claim 26, wherein said detecting is earned out using flow cytometry, immunoassay, or microscopy.

32. The method of claim 17, wherein the FcRn interacting molecule comprises a fragment crystallizable (Fc) region.

33. The method of claim 32, wherein the Fc region of the FcRn interacting molecule has been engineered for enhanced binding to FcRn relative to a corresponding non- engineered Fc region.

34. The method of claim 32, wherein the FcRn interacting molecule comprises a human antibody, an antigen-binding fragment of a human antibody comprising a Fc region, or a human antibody derivative comprising a Fc region.

35. The method of claim 17, wherein the FcRn interacting molecule is an Fc fusion protein or an albumin fusion protein.

36. The method of claim 17, wherein the first and second cell preparations are selected from the group consisting of Chinese Hamster Ovary (CHO) Cells, Madin-Darby canine kidney (MDCK) cells, Vero cells, HUVECs, HEK293 cells, and primary endothelial cells.

Images