How to Attribute Therapeutic Proteins

Abstract

[0003] Provided herein are methods for identifying structures, e.g., attributes, of a therapeutic protein or target that affect the interaction between the therapeutic protein and the target. In exemplary embodiments, the method includes: (a) applying stress to a first sample containing the therapeutic protein or target; (b) contacting the first sample with a second sample containing the target or therapeutic protein to form a mixture containing (i) the therapeutic protein-target complex, (ii) unbound therapeutic protein, and (iii) unbound target; and (c) separating the mixture into at least two fractions, wherein the unbound fraction contains unbound therapeutic protein or unbound target, and the bound fraction contains the therapeutic protein-target complex; and (d) identifying and quantifying the abundance of structures, e.g., attributes, present in the species of the therapeutic protein or target for each of the unbound and bound fractions.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Patent Application No. 62 / 857,637, filed June 5, 2019, entitled "Methods of Identifying Attributes of Therapeutic Proteins," and U.S. Provisional Patent Application No. 62 / 968,682, filed January 31, 2020, entitled "Methods of Identifying Attributes of Therapeutic Proteins," each of which is incorporated by reference in its entirety.

[0002] Incorporation by Reference of Electronically Submitted Materials The computer readable nucleotide / amino acid sequence listing submitted concurrently herewith is incorporated by reference in its entirety and is identified as follows: 268,171 byte ASCII (Text) file named "A-2296-WO-PCT_Seqlisting.txt", created on June 3, 2020. [Background technology]

[0003] The native structure of a protein adapts to changes within its environment. This structural flexibility is required for most, if not all, biological functions of proteins, but it also presents many challenges during the development of therapeutic proteins for pharmaceutical use. For example, therapeutic proteins endure various conditions in many process steps before being administered to a patient. Many process steps include, for example, one or more of the production (e.g., recombinant production), recovery, purification, formulation, filling, packaging, storage, delivery, and final preparation immediately prior to administration to a patient. During each of these steps, the therapeutic protein is subjected to one or more environments that may lead to changes in its structure. Changes in structure may result in the formation of different species of therapeutic proteins resulting in heterogeneous products. Some species retain the ability to bind to their target and thus maintain therapeutic efficacy, while others lose target binding ability and thus become functionally inactive. To maximize and maintain quality control of these pharmaceuticals, the biopharmaceutical industry has invested much effort in trying to understand why some species lose activity and others retain activity. Methodologies now exist to analyze the speciation of therapeutic proteins. However, current techniques are impractical in terms of time and cost, while others do not reflect the therapeutic protein in a physiologically relevant context. Also, some current techniques use covalent cross-linking or labeling agents that can adversely affect the outcome of the assay. Thus, there is a need for methods to identify species of therapeutic proteins that result in loss of binding, and methods to identify conditions that result in the formation of such species. Summary of the Invention [Means for solving the problem]

[0004] Described herein for the first time is the development of methods and validation of such methods for identifying attributes of a therapeutic protein that affect the interaction (e.g., binding interaction) between the therapeutic protein and its target. As described herein, exemplary methods of the present disclosure resulted in the identification of 15 attributes of a first therapeutic protein that were statistically significantly altered in the unbound fraction compared to the bound fraction, and thus such attributes were identified as being associated with the interaction (e.g., binding interaction) between the first therapeutic protein and its target. These exemplary methods of the present disclosure also resulted in the identification of 22 attributes of a second therapeutic protein that were statistically significantly altered in the unbound fraction compared to the bound fraction, and thus such attributes were identified as being associated with the interaction (e.g., binding interaction) between the second therapeutic protein and its target. The data provided herein supports the methods disclosed herein for more accurate and detailed identification of attributes that affect the interaction between a therapeutic protein and its target (e.g., binding target).

[0005] Thus, the present disclosure provides a method for identifying structures, e.g., attributes, of a therapeutic protein or target that affect the interaction between the therapeutic protein and the target. In an exemplary embodiment, the method includes: (a) applying stress to a first sample containing the therapeutic protein or target; (b) contacting the first sample with a second sample containing the target or therapeutic protein to form a mixture containing (i) a therapeutic protein-target complex, (ii) an unbound therapeutic protein, and (iii) an unbound target; (c) separating the mixture into at least two fractions, where the unbound fraction contains the unbound therapeutic protein or unbound target, and the bound fraction contains the therapeutic protein-target complex; and (d) identifying and quantifying the abundance of structures, e.g., attributes, present in the species of the therapeutic protein or target for each of the unbound fraction and the bound fraction. For example, the abundance of each structure present in the species of the therapeutic protein or target can be identified and quantified. In an exemplary embodiment, if the abundance of a structure, e.g., attribute, in the unbound fraction is greater than the abundance of the structure, e.g., attribute, in the bound fraction, then the structure, e.g., attribute, negatively affects the interaction between the therapeutic protein and the target.

[0006] In various embodiments, the mixture is separated into at least two fractions using a technique that separates the components of the mixture based on size, charge, hydrophobicity, affinity for a capture molecule, or a combination thereof. In exemplary embodiments, the technique is size exclusion chromatography (SEC), affinity chromatography, precipitation using beads or cells, free-flow fractionation (FFF), ion exchange chromatography (IEX), hydrophobic interaction chromatography (HIC) or ultracentrifugation (UC). In further or alternative embodiments, the mixture is separated into at least two fractions using a technique that separates the components of the mixture based on size, optionally the technique is size exclusion chromatography (SEC). In various embodiments, the mixture is separated into at least two fractions using a technique that separates the components of the mixture based on affinity for a capture molecule bound to a solid support, optionally a bead or a cell. In exemplary embodiments, the capture molecule is a target or an Fc receptor (e.g., FcRn, Fcγ receptor).

[0007] The present disclosure also provides a method for determining the effect of a known structure, e.g., a known attribute, present in a species of a therapeutic protein or target on the interaction between the therapeutic protein and the target. In an exemplary embodiment, the method includes: (a) contacting a first sample containing a therapeutic protein or target having a known structure, e.g., attribute, with a second sample containing the target or therapeutic protein to form a mixture containing (i) a therapeutic protein-target complex, (ii) an unbound therapeutic protein, and (iii) an unbound target; (b) separating the mixture into at least two fractions, the unbound fraction containing the unbound therapeutic protein or target, and the bound fraction containing the therapeutic protein-target complex; and (c) quantifying the abundance of the known structure, e.g., attribute, for each of the unbound fraction and the bound fraction. In an exemplary embodiment, if the abundance of the known structure, e.g., attribute, in the unbound fraction is greater than the abundance of the known structure, e.g., attribute, in the bound fraction, the known structure, e.g., attribute, has a negative effect on the interaction between the therapeutic protein and the target.

[0008] Further explanation and guidance as well as examples of the disclosed methods are provided herein. [Brief description of the drawings]

[0009] [Figure 1A-1] FIG. 1 is a diagram of an exemplary workflow of a method of the present disclosure for identifying attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target. The therapeutic protein is subjected to stress that results in attribute formation at different amino acid residues. Each therapeutic protein with a different profile of attributes is considered a species. When a target is added to the stressed therapeutic protein, an interaction is formed between the therapeutic protein species and the target. For some species, no interaction is formed between the therapeutic protein species and the target. The mixture of therapeutic protein-target complex, unbound therapeutic protein, and unbound target is separated into different fractions including bound and unbound fractions. Each attribute in each fraction is identified by attribute type (e.g., oxidation, deamidation, etc.) and position (e.g., amino acid residue position). The abundance of each attribute in each fraction is determined. For each attribute, the ratio of the abundance of the attribute in the unbound fraction to the abundance of the attribute in the bound fraction is determined. The method is repeated starting from the separation step to determine a second ratio. The method is repeated again. Each time, a new ratio is determined. The statistical significance of the ratios is determined. If the ratio is statistically significant, the attribute is considered to influence the interaction between the therapeutic protein and the target. [Figure 1A-2]FIG. 1 is a diagram of an exemplary workflow of a method of the present disclosure for identifying attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target. The therapeutic protein is subjected to stress that results in attribute formation at different amino acid residues. Each therapeutic protein with a different profile of attributes is considered a species. When a target is added to the stressed therapeutic protein, an interaction is formed between the therapeutic protein species and the target. For some species, no interaction is formed between the therapeutic protein species and the target. The mixture of therapeutic protein-target complex, unbound therapeutic protein, and unbound target is separated into different fractions including bound and unbound fractions. Each attribute in each fraction is identified by attribute type (e.g., oxidation, deamidation, etc.) and position (e.g., amino acid residue position). The abundance of each attribute in each fraction is determined. For each attribute, the ratio of the abundance of the attribute in the unbound fraction to the abundance of the attribute in the bound fraction is determined. The method is repeated starting from the separation step to determine a second ratio. The method is repeated again. Each time, a new ratio is determined. The statistical significance of the ratios is determined. If the ratio is statistically significant, the attribute is considered to influence the interaction between the therapeutic protein and the target. [Figure 1B] FIG. 1 is a diagram of an exemplary workflow of the disclosed method for identifying attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target. In this figure, the separation step of FIG. 1A is named "fractionation" and the identification and abundance determination are performed by LC-MS peptide mapping. At least the LC-MS peptide mapping step is repeated several times, each time determining an abundance ratio for each attribute, where the abundance ratio is the ratio of the abundance of the attribute in the unbound fraction to the abundance of the attribute in the bound fraction. The statistical significance of the abundance ratio is determined by determining a p-value with a t-test. The statistical significance is plotted against the abundance ratio (or fold change) in a volcano plot. The round dots around fold change=1 in the volcano plot indicate (a) attributes that do not affect the interaction formation between the target and the therapeutic protein, and (b) artifactual attributes caused by sample preparation. Attributes that affect binding are indicated with an asterisk in the upper right corner. [Figure 2A] FIG. 1 is an exemplary size-exclusion chromatogram showing the elution of two targets bound to the arms of Antibody 1, one target bound to Antibody 1, Antibody 1 not bound to any target, and unbound target. [Figure 2B] Graph of statistical significance of abundance plotted as a function of abundance (or fold change). A p-value <0.05 corresponds to a -log10 p-value >1.3, and a fold change >1.78 corresponds to a Log2 fold change >0.83. [Diagram 3] 1 is an exemplary attribute profile listing 39 modifications with p-values ​​<0.05 out of a total of 188 modifications identified in Antibody 1. HC, heavy chain; LC, light chain; FC, fold change (ratio of abundance of attribute in unbound fraction to abundance of attribute in bound fraction), p-value (determined by t-test). [Figure 4] 1 is an exemplary attribute profile listing 15 modifications on 7 residues of the heavy and light chains of Antibody 1 that have a statistically significant effect on binding between Antibody 1 and its target. These modifications have a p-value < 0.002 (-log10 p-value > 2.7) and a fold change > 1.78 (log2 fold change > 0.83). HC, heavy chain; LC, light chain; FC, fold change (ratio of abundance of attribute in unbound fraction to abundance of attribute in bound fraction), p-value (determined by t-test). [Figure 5A] 1 is an exemplary size-exclusion chromatogram showing the elution of antibody protein product 1 (APP1), an Fc fusion protein bound to one target, unbound target, and unbound APP1. [Figure 5B] 1 is a graph of the statistical significance of abundance plotted as a function of abundance (or fold change). A p-value <0.0005 corresponds to a -log10 p-value >3.3, and a fold change >2 corresponds to a Log2 fold change >1. [Figure 6A]FIG. 1 is a diagram of an exemplary workflow of the disclosed method of identifying attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target, where beads or cells bound to a capture molecule and expressing the capture molecule on their surface are used. The stressed therapeutic protein is contacted with the beads or cells. The capture molecule is the target. In the tube, a mixture of therapeutic protein-target complex and unbound therapeutic protein is formed. The beads or cells containing the bound therapeutic protein can be separated from the unbound fraction by centrifugation. The beads or cells with bound therapeutic protein form a pellet with the bound therapeutic protein, and the unbound protein remains in the supernatant. [Figure 6B] 1 is an exemplary workflow of a method of the present disclosure for identifying attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target, where a bead is coupled to a capture molecule. The capture molecule is the target. The beads coupled to the target are loaded onto a column. The stressed therapeutic protein is loaded onto the column and the flow-through comprises the unbound fraction. The bound fraction is eluted separately and collected. The method is performed with a therapeutic protein having a KD of about 10-11 to about 10-9 M for the target. [Figure 7] FIG. 6B shows an exemplary method similar to that shown in FIG. 6A, except that the beads are conjugated to Fc receptors or the cells express Fc receptors. The method is performed with a therapeutic protein that has a KD of about 10-8 to about 10-6 M for the target. [Figure 8A-8C]Figure 1 shows a workflow for identifying key chemical modifications and paratope mapping by size exclusion chromatography (SEC) of stressed antibody-target complexes according to some embodiments. A) The antibody is stressed with one or more of the following conditions: high temperature in the initial formulation, e.g., 40°C for 4 weeks (40C4W) in a weakly acidic formulation (pH 5), physiological conditions (pH 7, 37°C), extreme pH, UV light. In this study, 20 hours of UV light with a total of 200 watt-hours per square meter was used. The target protein was added and mixed with the stressed antibody in the appropriate stoichiometric ratio, typically an antibody:target ratio of 1:2. After a short incubation, the mixture was injected into the SEC system and separated for bound and unbound antibody species as follows: antibodies bound to two target molecules eluted first, followed by antibodies bound to one target molecule, then unbound antibodies and remaining targets. B) SEC fractions containing antibodies bound to two targets (bound) and antibodies bound to one target (partially bound) were collected and subjected to peptide mapping. These fractions were chosen due to their comparable protein concentrations, which is important for equal sample preparation conditions. Proteins in the collected fractions were unfolded, disulfides reduced, alkylated, and the protein backbone was digested into peptides by an enzyme (trypsin) or a combination of enzymes. Chemical modifications in the bound and unbound antibody species were identified and quantified by LC-MS / MS peptide mapping. The whole procedure was repeated several times (n=3) and statistical significance was calculated by performing a t-test using the percentage of modification in bound vs. unbound antibody molecules. A graph (volcano plot) was plotted as statistical significance vs. the modification abundance ratio (fold change) in unbound / bound antibody molecules. Statistically significant data were grouped in the upper right corner bounded by a dashed line for p-values ​​<0.006 (-log10 p-values ​​>2.2) and fold changes >1.78 (Log2 fold change >0.83). Eleven residues were modified with a particular statistical significance and modification ratio (unbound / bound antibody). C) Paratope maps were plotted as modification scores versus residue number. The modification score was chosen as the sum of the fold change and the logarithm of the p-value (Log2 fold change +-log10 p-value).D) Correlation of antibody-target complexes with available crystallography data revealed that 8 of 11 antibody residues with modifications identified by this method are within 5 Å of the target, suggesting that these antibody residues are close to and may interact with the target. [Figure 8D]Figure 1 shows a workflow for identifying key chemical modifications and paratope mapping by size exclusion chromatography (SEC) of stressed antibody-target complexes according to some embodiments. A) The antibody is stressed with one or more of the following conditions: high temperature in the initial formulation, e.g., 40°C for 4 weeks (40C4W) in a weakly acidic formulation (pH 5), physiological conditions (pH 7, 37°C), extreme pH, UV light. In this study, 20 hours of UV light with a total of 200 watt-hours per square meter was used. The target protein was added and mixed with the stressed antibody in the appropriate stoichiometric ratio, typically an antibody:target ratio of 1:2. After a short incubation, the mixture was injected into the SEC system and separated for bound and unbound antibody species as follows: antibodies bound to two target molecules eluted first, followed by antibodies bound to one target molecule, then unbound antibodies and remaining targets. B) SEC fractions containing antibodies bound to two targets (bound) and antibodies bound to one target (partially bound) were collected and subjected to peptide mapping. These fractions were chosen due to their comparable protein concentrations, which is important for equal sample preparation conditions. Proteins in the collected fractions were unfolded, disulfides reduced, alkylated, and the protein backbone was digested into peptides by an enzyme (trypsin) or a combination of enzymes. Chemical modifications in the bound and unbound antibody species were identified and quantified by LC-MS / MS peptide mapping. The whole procedure was repeated several times (n=3) and statistical significance was calculated by performing a t-test using the percentage of modification in bound vs. unbound antibody molecules. A graph (volcano plot) was plotted as statistical significance vs. the modification abundance ratio (fold change) in unbound / bound antibody molecules. Statistically significant data were grouped in the upper right corner bounded by a dashed line for p-values ​​<0.006 (-log10 p-values ​​>2.2) and fold changes >1.78 (Log2 fold change >0.83). Eleven residues were modified with a particular statistical significance and modification ratio (unbound / bound antibody). C) Paratope maps were plotted as modification scores versus residue number. The modification score was chosen as the sum of the fold change and the logarithm of the p-value (Log2 fold change +-log10 p-value).D) Correlation of antibody-target complexes with available crystallography data revealed that 8 of 11 antibody residues with modifications identified by this method are within 5 Å of the target, suggesting that these antibody residues are close to and may interact with the target. [Figure 9] Volcano plot of statistical significance versus modification abundance ratio (fold change) of non-binding / binding antibody molecules according to some embodiments. Statistically significant data were grouped in the upper right corner bounded by dashed lines for p-values ​​< 0.006 (-log10 p-values ​​> 2.2) and fold changes > 1.78 (Log2 fold change > 0.83). A log scale is often used for volcano plots for better visualization. The location of the two vertical dashed lines for the fold change (log2 fold change) thresholds was selected by symmetrically moving the lines outwards until the top left quadrant was clear. The location of the horizontal threshold line for statistical significance (-log10 p-value) was selected by moving it upwards until the top left quadrant was clear. These thresholds define the fold change and p-value in the top right quadrant where the modification reduces binding statistically significantly. [Figure 10A]1 shows paratope mapping plots of A) the heavy chain (including proximity maps in its variable regions) and B) the light chain (including proximity maps in its variable regions) according to some embodiments. The paratope maps were plotted as modification scores versus residue number. The modification scores were chosen as the sum of the logarithm of the fold change and the logarithm of the p-value (Log2 fold change +-log10 p-value), because the scales have similar values ​​and both are important to separate the signal from the noise. Ten of the eleven residues (labeled on the plot) with the highest scores are in the complementarity determining regions (CDRs), regions that are typically involved in binding. Residues in conserved regions that are not typically involved in binding had significantly lower scores. Correlation with available crystallography data of antibody-target complexes revealed that eight of the eleven antibody residues with modifications identified by the method were within 5 Å of the target, indicating that these antibody residues are close to and likely to interact with the target. These results supported the validity of the method. Modifications on cysteine ​​residues (C23) outside the CDRs were also detected. Without being limited by theory, the results suggested that long-range (allosteric) effects may play a role caused by reduction of disulfide bonds. Of the 11 residues, clipping at LC S32 / S33 was indicated by two dots (S32 and S33) for the two detected fragments that were significantly increased in the unbound SEC fraction. Clipping reduces binding to the target, which is an important attribute, but these residues may not be involved in binding to the target. Without being limited by theory, a portion of the antibody molecule may easily be lost as a result of clipping. Paratope mapping can generally be performed without clipping data. [Figure 10B]1 shows paratope mapping plots of A) the heavy chain (including proximity maps in its variable regions) and B) the light chain (including proximity maps in its variable regions) according to some embodiments. The paratope maps were plotted as modification scores versus residue number. The modification scores were chosen as the sum of the logarithm of the fold change and the logarithm of the p-value (Log2 fold change +-log10 p-value), because the scales have similar values ​​and both are important to separate the signal from the noise. Ten of the eleven residues (labeled on the plot) with the highest scores are in the complementarity determining regions (CDRs), regions that are typically involved in binding. Residues in conserved regions that are not typically involved in binding had significantly lower scores. Correlation with available crystallography data of antibody-target complexes revealed that eight of the eleven antibody residues with modifications identified by the method were within 5 Å of the target, indicating that these antibody residues are close to and likely to interact with the target. These results supported the validity of the method. Modifications on cysteine ​​residues (C23) outside the CDRs were also detected. Without being limited by theory, the results suggested that long-range (allosteric) effects may play a role caused by reduction of disulfide bonds. Of the 11 residues, clipping at LC S32 / S33 was indicated by two dots (S32 and S33) for the two detected fragments that were significantly increased in the unbound SEC fraction. Clipping reduces binding to the target, which is an important attribute, but these residues may not be involved in binding to the target. Without being limited by theory, a portion of the antibody molecule may easily be lost as a result of clipping. Paratope mapping can generally be performed without clipping data. [Figures 11A-11D]Volcano plots are shown for A) 275 total chemical modifications revealed by peptide mapping of amino acid residues of stressed therapeutic antibody. B) 27 modifications on 11 residues that statistically significantly affected binding to the target and fulfilled the following criteria: Log2(FC)>0.83 and -Log10(p)>2.2. C) 11 residues of the antibody identified as residues involved in binding to the target (paratope). D) 6 residues with more than 0.3% modifications after 4 weeks of stress (40C4W) at 40°C in liquid formulation. The 40C4W accelerated stress was chosen to simulate reasonable conditions of manufacture and storage during the antibody's shelf life of 2 years at 4°C in a weakly acidic liquid formulation. The modifications of the 6 residues affected binding and had relatively large values ​​under reasonable conditions of manufacture and storage and should be considered important attributes and monitored. The 40C4W stress caused very little modification (<0.01%) in the other 5 residues. UV light caused a higher rate of modification of these residues, which were useful for paratope mapping. [Figure 12-1] Chemical modification profile listing 56 modifications with p-values ​​<0.05, -Log10(p)>1.3 out of a total of 275 modifications identified in the stressed antibody used in the study according to some embodiments herein (see Example 6). Of these, 27 modifications met the following criteria of statistical significance: Log2(FC)>0.83 and -Log10(p)>2.2. These modifications occurred in 11 residues, with some residues having more than one modification. Eleven residues of the antibody were identified as residues involved in binding to the target (paratope). In six residues (including two residues with a clip between them), the modifications exceeded 0.3% after 4 weeks of stress at 40°C (40C4W) in liquid formulation. [Figure 12-2]Chemical modification profile listing 56 modifications with p-values ​​<0.05, -Log10(p)>1.3 out of a total of 275 modifications identified in the stressed antibody used in the study according to some embodiments herein (see Example 6). Of these, 27 modifications met the following criteria of statistical significance: Log2(FC)>0.83 and -Log10(p)>2.2. These modifications occurred in 11 residues, with some residues having more than one modification. Eleven residues of the antibody were identified as residues involved in binding to the target (paratope). In six residues (including two residues with a clip between them), the modifications exceeded 0.3% after 4 weeks of stress at 40°C (40C4W) in liquid formulation. [Figure 12-3] Chemical modification profile listing 56 modifications with p-values ​​<0.05, -Log10(p)>1.3 out of a total of 275 modifications identified in the stressed antibody used in the study according to some embodiments herein (see Example 6). Of these, 27 modifications met the following criteria of statistical significance: Log2(FC)>0.83 and -Log10(p)>2.2. These modifications occurred in 11 residues, with some residues having more than one modification. Eleven residues of the antibody were identified as residues involved in binding to the target (paratope). In six residues (including two residues with a clip between them), the modifications exceeded 0.3% after 4 weeks of stress at 40°C (40C4W) in liquid formulation. [Figure 12-4]Chemical modification profile listing 56 modifications with p-values ​​<0.05, -Log10(p)>1.3 out of a total of 275 modifications identified in the stressed antibody used in the study according to some embodiments herein (see Example 6). Of these, 27 modifications met the following criteria of statistical significance: Log2(FC)>0.83 and -Log10(p)>2.2. These modifications occurred in 11 residues, with some residues having more than one modification. Eleven residues of the antibody were identified as residues involved in binding to the target (paratope). In six residues (including two residues with a clip between them), the modifications exceeded 0.3% after 4 weeks of stress at 40°C (40C4W) in liquid formulation. [Figures 13A-13I]A volcano plot of the individual chemical modifications of the statistical significance versus abundance ratio in the non-bound / combined antibody fractions according to some embodiments. The following modifications were plotted: A) all modifications; B) isomerization and H2O elimination (the latter for succinimide); C) deamidation; D) +13.97 Da (doubly oxidized H, W and H2O elimination modification); E) clip; F) digestion artifacts including unintentional carboxymethylation of H, M and under-alkylation of some C residues; G) unknown modifications including oxidation, hydroxylation, H2O elimination identified in "abnormal residues" and some other modifications only identified by mass change; H) oxidation mainly identified on M, probably caused by dissolved oxygen; I) double oxidation, triple oxidation mainly identified on W, H, and oxidation to quinurenine, likely caused by radicals, metals and light. The plot was used for quality control applying the following rationale. When collecting and processing the SEC non-bound and combined antibody fractions similarly, the noise is expected to be symmetric around the base of the volcano and centered on the vertical axis Log2(FC)=0. The noise includes the region with -log10(p-value) < 2.2 and the region with -log10(p-value) > 2.2 and 0.7 < log2(FC) < 0.83 (shown in gray in the all-modifications plot). If the dots in the noise region are not symmetrically distributed with respect to the vertical axis, different treatments were performed on the bound and non-bound fractions. The dots for double oxidation, triple oxidation, and oxidation to quinurenine are mostly slightly shifted to the right, suggesting that the non-bound fraction may be more exposed to radicals, metals and light. Otherwise, both fractions were treated similarly with respect to factors causing other modifications. [Figure 14-1] A schematic diagram of the SEC binding affinity workflow according to some embodiments. [Figure 14-2] A schematic diagram of the SEC binding affinity workflow according to some embodiments. [Figures 15A-15G] SEC-UV profiles (280 nm) of antibodies, receptors, and their mixtures at different ratios according to some embodiments. The binding ratios used in each experiment are shown in the corresponding panel. [Figures 16A-16C] SEC-UV-MALS profiles of stressed antibody (45C10d), antigen and their mixture with a 1:1 binding ratio according to some embodiments. MALS data are represented as dots with the y-axis of determined molecular weights shown on the right. [Figure 17] SEC-UV (280 nm) profile and collected fractions of 45C10d antibody / antigen mixture at 1:1 ratio according to some embodiments. The UV profile is shown as a blue line and the dashed blue line indicates the area for fraction collection of the antibody 45C10d / antigen mixture. SEC-UV profiles of antibody T0, antibody 45C10d, HER2 and antibody T0 / antigen mixture are shown in black, red, magenta and green for comparison. [Figures 18A-18D] 17 shows the SEC-UV (280 nm) profiles of antibody TO, antibody 45C10d, collected fraction 1 and collected fraction 3 according to some embodiments. Fractions 1 and 3 are collected based on the SEC-UV profile of the 1:1 ratio antibody 45C10d / antigen mixture shown in FIG. [Figure 19] 1 is a volcano plot for the importance assessment of antibody attributes according to some embodiments. The x-axis represents the ln2 value of the ratio between non-binding and binding, and the y-axis represents the log10 value of the p-value. For reference, the vertical dashed line corresponds to a ratio of 1.5, and the three horizontal dashed lines represent the following three p-values: 0.05, 0.005, and 0.05 / 420. [Fig. 20A-20E] 1 is a volcano plot of selected attributes of an antibody by SEC of the antibody and its target, including oxidation (A), isomerization (B), deamidation (C), glycation, mannosylation and N-glycans (D), and artifact digestion (E), according to some embodiments. [Figure 21] 1 shows a relative abundance plot of select attributes of antibodies in binding (white) and non-binding (blue) species according to some embodiments. Modifications with p-values ​​less than 0.05 are marked with an asterisk symbol. [Figure 22A]1 shows the CEX-UV profile of an antibody according to some embodiments, and a characterization of each proteoform observed in the CEX experiment. Potency measurements of the CEX peaks are also shown in the table to the right. [Figure 22B] FIG. 1: Crystal structure of antibody-HER2 and a close-up of the binding region with key attributes shown in the right box. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] The present disclosure provides methods for identifying structures, e.g., attributes, of a therapeutic protein or target that affect the interaction between the therapeutic protein and the target. In an exemplary embodiment, the method includes the steps of: (a) applying a stress to a first sample comprising the therapeutic protein or target; (b) contacting the first sample with a second sample comprising the target or therapeutic protein (e.g., the first sample can comprise the therapeutic protein and the second sample can comprise the target, or the first sample can comprise the target and the second sample can comprise the therapeutic protein) to form a mixture comprising (i) the therapeutic protein-target complex, (ii) unbound therapeutic protein, and (iii) unbound target; (c) separating the mixture into at least two fractions, the unbound fraction comprising unbound therapeutic protein or unbound target, and the bound fraction comprising therapeutic protein-target complex; and (d) identifying and quantifying the abundance of structures present in the species of the therapeutic protein or target for each of the unbound fraction and the bound fraction. For example, the abundance of each structure present in a species of therapeutic protein can be identified and quantified. In an exemplary embodiment, if the abundance of the structure in the unbound fraction is greater than the abundance of the structure in the bound fraction, the structure negatively affects the interaction between the therapeutic protein and the target. In an exemplary example, if the abundance of the structure in the unbound fraction is equal to or less than the abundance of the structure in the bound fraction, the structure does not negatively affect the interaction between the therapeutic protein and the target. Without being limited by theory, it can be inferred that if a structure negatively affects the interaction between a therapeutic protein and a target, the structure may be at or near a position involved in binding between the therapeutic protein and the target and / or allosterically affects binding between the therapeutic protein and the target. By way of example, in the methods described herein, the therapeutic protein can include an antibody and the target can include an antigen for the antibody.

[0011] The present disclosure also provides a method for determining the effect of a known structure (e.g., attribute) present in a species of a therapeutic protein or target on the interaction between the therapeutic protein and the target. In an exemplary embodiment, the method includes: (a) contacting a first sample containing a therapeutic protein or target having a known structure (e.g., attribute) with a second sample containing the target or therapeutic protein to form a mixture containing (i) a therapeutic protein-target complex, (ii) an unbound therapeutic protein, and (iii) an unbound target; (b) separating the mixture into at least two fractions, the unbound fraction containing the unbound therapeutic protein or target, and the bound fraction containing the therapeutic protein-target complex; and (c) quantifying the abundance of the known structure (e.g., attribute) for each of the unbound fraction and the bound fraction. In an exemplary embodiment, if the abundance of the known structure (e.g., attribute) in the unbound fraction is greater than the abundance of the known structure in the bound fraction, the known structure (e.g., attribute) has a negative effect on the interaction between the therapeutic protein and the target. In an illustrative example, if the abundance of a known structure (e.g., an attribute) in the unbound fraction is less than or equal to the abundance of the known structure in the bound fraction, the known structure does not have a negative effect on the interaction between the therapeutic protein and the target.

[0012] In various aspects, the method is a method for identifying attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target. In an exemplary embodiment, the method includes: (a) applying stress to a first sample containing a therapeutic protein; (b) contacting the first sample with a second sample containing a target to form a mixture containing (i) a therapeutic protein-target complex, (ii) an unbound therapeutic protein, and (iii) an unbound target; (c) separating the mixture into at least two fractions, the unbound fraction containing the unbound therapeutic protein, and the bound fraction containing the therapeutic protein-target complex; and (d) identifying and quantifying the abundance of the attributes present in the species of the therapeutic protein for each of the unbound fraction and the bound fraction. For example, each attribute in the species of the therapeutic protein can be identified and quantified. In an exemplary embodiment, if the abundance of the attribute in the unbound fraction is greater than the abundance of the attribute in the bound fraction, the attribute negatively affects the interaction between the therapeutic protein and the target. In an illustrative example, if the abundance of the attribute in the unbound fraction is less than or equal to the abundance of the attribute in the bound fraction, the attribute does not negatively affect the interaction between the therapeutic protein and the target.

[0013] In various examples, the method is a method of determining the effect of a known attribute present in a species of a therapeutic protein on the interaction between the therapeutic protein and a target. In an exemplary embodiment, the method includes: (a) contacting a first sample containing a therapeutic protein having a known attribute with a second sample containing the target to form a mixture containing (i) a therapeutic protein-target complex, (ii) an unbound therapeutic protein, and (iii) an unbound target; (b) separating the mixture into at least two fractions, the unbound fraction containing the unbound therapeutic protein and the bound fraction containing the therapeutic protein-target complex; and (c) quantifying the abundance of the known attribute for each of the unbound fraction and the bound fraction. In an exemplary embodiment, if the abundance of the known attribute in the unbound fraction is greater than the abundance of the known attribute in the bound fraction, the known attribute has a negative effect on the interaction between the therapeutic protein and the target. In an illustrative example, if the abundance of the known attribute in the unbound fraction is less than or equal to the abundance of the known attribute in the bound fraction, the known attribute does not have a negative effect on the interaction between the therapeutic protein and the target.

[0014] As used herein, the term "therapeutic protein" refers to any molecule, which may be naturally occurring or engineered or synthetic, that includes at least one polypeptide chain, intended to achieve a therapeutic effect when administered to a subject for the treatment of a disease or condition. In various embodiments, a therapeutic protein interacts with a target to form an interaction, e.g., a binding interaction. In various examples, the formation of an interaction between a therapeutic protein and a target is important for the mechanism of action of the therapeutic protein. In various embodiments, this interaction is related to the therapeutic effect of the therapeutic protein in terms of the intended treatment. In an exemplary example, a therapeutic protein interacts (e.g., binds) with a target with the required affinity and specificity. Affinity is a measure of the strength of the interaction (e.g., binding interaction) between a therapeutic protein and a target. In an exemplary embodiment, a therapeutic protein has a high affinity for a target and therefore binds to more targets in a shorter time than a therapeutic protein with a lower affinity. In an exemplary embodiment, the affinity of a therapeutic protein for a target is measured by the equilibrium association constant K A In an exemplary embodiment, the therapeutic protein has an equilibrium association constant for the target, K A At least 10 5 Mol -1 , at least 10 6 Mol -1 , at least 10 7 Mol -1 , at least 10 8 Mol -1 , at least 10 9 Mol -1 Or at least 10 10 Mol -1 As will be understood by those skilled in the art, K A can be affected by factors including pH, temperature and buffer composition.

[0015] In an exemplary embodiment, the strength of the interaction between a therapeutic protein and a target can be described in terms of its sensitivity. D is the equilibrium dissociation constant, i.e., the k off / k onIt is the ratio of K D and K. A is inversely proportional to D The value is related to the concentration of the therapeutic protein (the amount of therapeutic protein required for a particular experiment) and therefore the K D The smaller the value (the lower the concentration), the higher the affinity of the therapeutic protein. In an exemplary embodiment, the binding strength of a therapeutic protein to a target is expressed as K D In an exemplary embodiment, the K of a therapeutic protein for a target is D is about 10 -1 Below, about 10 -2 Below, about 10 -3 Below, about 10 -4 Below, about 10 -5 Below, about 10 -6 In an exemplary embodiment, the K of a therapeutic protein against a target is D is micromolar, nanomolar, picomolar, or femtomolar. In an exemplary embodiment, the K D is about 10 -4 ~10 -6 or 10 -7 ~10 -9 or 10 -10 ~10 -12 or 10 -13 ~10 -15 In an exemplary embodiment, the K of a therapeutic protein is in the range D is about 10 -12 M~about 10 -6 M, about 10 -11 M~about 10 -6 M, about 10 -10 M~about 10 -6 M, about 10 -9 M~about 10 -6 M, about 10 -8 M~about 10 -6 M, about 10 -7 M~about 10 -6 M, about 10 -12 M~about 10 -7 M, about 10 -12 M~about 10 -8 M, about 10 -12 M~about 10 -9 M, about 10 -12 M~about 10 -10M or about 10 -12 M~about 10 -11 In an exemplary embodiment, the K of a therapeutic protein is in the range D is about 10 -11 M~about 10 -9 In some such embodiments, the interaction is considered a strong interaction (e.g., a strong binding interaction). In some embodiments, the K D is about 10 -8 ~about 10 -6 M, and in some such instances, the interaction is considered a weak interaction (e.g., a weak binding interaction).

[0016] In various embodiments, a therapeutic protein is subjected to conditions that lead to a change in its structure (e.g., a change in the structure of an amino acid of the therapeutic protein), leading to the formation of a species of therapeutic protein. In exemplary embodiments, the altered structure of an amino acid is referred to as an "attribute" and can be characterized with respect to its chemical identity or attribute type and position within the amino acid sequence of a therapeutic protein (e.g., the position of the amino acid where the attribute is present). For example, asparagine and glutamine residues are susceptible to deamidation. A deamidated asparagine at position 10 of a therapeutic protein amino acid sequence is an example of an attribute. A list of exemplary attribute types for specific amino acids is provided in Table A. Thus, a "structure" as used herein can comprise, consist essentially of, or consist of an attribute type listed in Table A or a combination of two or more attribute types listed in Table A. An attribute is an example of a structure, and it will be understood that whenever "structure" is referred to herein, an attribute is contemplated as an example of a structure, unless otherwise indicated.

[0017] [Table 1]

[0018] Because a therapeutic protein contains multiple amino acids, the therapeutic protein may have multiple attributes (e.g., multiple amino acids with altered structures) and may be described in terms of its attribute profile. As used herein, the term "attribute profile" refers to a list of attributes of a therapeutic protein. In various examples, the attribute profile optionally provides a chemical identity or attribute type, e.g., deamidation, relative to the native structure of the therapeutic protein. In various examples, the attribute profile provides the location of the attribute, e.g., the position of the amino acid at which the attribute is present. Exemplary attribute profiles of therapeutic proteins (e.g., therapeutic protein attribute profiles) are shown in Figures 3 and 4. In some embodiments, the attribute profile provides a description of all attributes present in a therapeutic protein. In other embodiments, the attribute profile provides a description of some attributes present in a therapeutic protein. For example, the attribute profile may provide only attributes present in a particular portion of a therapeutic protein, e.g., the extracellular domain, variable region, CDRs. A species of therapeutic protein is characterized by the attributes present in the therapeutic protein. A species of therapeutic protein may differ from another species of the same therapeutic protein by having a different attribute profile. If two therapeutic proteins have different attribute profiles, then the therapeutic proteins represent two different species of therapeutic proteins. If two therapeutic proteins have identical attribute profiles, then the therapeutic proteins are considered to be the same species of therapeutic proteins.

[0019] In various examples, a therapeutic protein is placed in conditions that cause its structure to change (e.g., the formation of one or more attributes), which change in structure alters the affinity of the therapeutic protein for its target. In various embodiments, a therapeutic protein is placed in conditions that cause its structure to change (e.g., the formation of one or more attributes), which change in structure reduces the affinity of the therapeutic protein for its target. In some embodiments, the reduction in affinity leads to a partial or total loss of the ability of the therapeutic protein to interact (e.g., bind) with the target. In various examples, the partial or total loss of the ability of the therapeutic protein to interact (e.g., bind) with the target ultimately reduces the efficacy of the therapeutic protein. In alternative examples, a therapeutic protein is placed in conditions that cause its structure to change (e.g., the formation of one or more attributes), which change in structure does not alter the affinity of the therapeutic protein for its target. In various embodiments, the change in structure does not reduce the affinity of the therapeutic protein for its target. Without being bound to any particular theory, the methods of the present disclosure advantageously accurately and precisely distinguish attributes of a therapeutic protein that affect the interaction between the therapeutic protein and a target from attributes that do not affect the interaction.

[0020] In various embodiments, a therapeutic protein is subjected to conditions that lead to a change in its structure (e.g., a change in the structure of the amino acids of the therapeutic protein), leading to the formation of one or more attributes, thereby creating a species of therapeutic protein. In various embodiments, the change in the structure of a therapeutic protein can be caused by a post-translational modification involving one or more intracellular enzymes, and thus the change in the structure of a therapeutic protein can be caused by an enzymatic modification. Enzymatic modifications are known in the art and include, for example, lipidation (e.g., palmitoylation, myristoylation), phosphorylation, glycosylation, ubiquitination, nitrosylation, methylation, acetylation, and proteolysis. In various embodiments, the amino acid sequence of a therapeutic protein having a post-translational modification is altered by one or more amino acid residues including a moiety (e.g., lipid, sugar, peptide, phosphate group, ubiquitin tag, methyl, acetyl group), but is considered to be the same as the unmodified therapeutic protein. In various examples, the change in the structure of a therapeutic protein can be caused by a change in the protein's environment. For example, the environment of a protein may change in pH, salinity, osmolarity, pressure, temperature, exposure to light, UV light, etc., exposure to air or oxygen, agitation / shaking, exposure to chemicals or materials (e.g., metals, plastics), or long-term storage, and such environmental changes cause changes in the structure of one or more amino acids. The change in environment may be, for example, a change in culture medium or a change in formulation components, a change in storage conditions, etc. In some embodiments, the change in environment occurs during one or more of the process steps leading up to administration to a patient (e.g., production (e.g., recombinant production), harvesting, purification, formulation, filling, packaging, storage, delivery, and final preparation of the protein just prior to administration to a patient). Exemplary modifications that may occur during manufacturing include, for example, removal of residues from the polypeptide of a therapeutic protein and / or protease-mediated cleavage. Exemplary changes that may occur after production of a therapeutic protein but prior to administration (e.g., during packaging and / or filling, storage and / or transportation / delivery) include, for example, oxidation, reduction, deamidation, deamination, aggregation, denaturation, precipitation, hydrolysis, aspartate isomerization, N-terminal and C-terminal modifications.Thus, in various aspects, the change in the structure of a Therapeutic protein can be caused by a non-enzymatic modification (e.g., chemical modification) that does not involve intracellular enzymes, and thus the change in the structure of a Therapeutic protein can be caused by a non-enzymatic modification (e.g., chemical modification). In various embodiments, the modification occurs after administration of the Therapeutic protein and occurs in vivo (relative to the subject to which the Therapeutic protein is administered).

[0021] In various embodiments, a target is analyzed by the methods of the present disclosure. Thus, in various embodiments, the target is subjected to conditions that lead to a change in its structure (e.g., a change in the structure of the amino acid of the target), leading to the formation of a target species. In exemplary embodiments, the changed structure of an amino acid is referred to as an "attribute" and can be characterized with respect to its chemical identity or attribute type and its position in the amino acid sequence of the target (e.g., the position of the amino acid where the attribute is present). For example, asparagine and glutamine residues are susceptible to deamidation. A deamidated asparagine at position 10 of the amino acid sequence of the target is an example of an attribute. A list of exemplary attribute types for certain amino acids is provided in Table A. Thus, in the methods described herein, an attribute can comprise, consist essentially of, or consist of an attribute type provided in Table A or a combination of two or more attribute types provided in Table A.

[0022] Since a target contains multiple amino acids, the target may have multiple attributes (e.g., multiple amino acids with altered structures) and may be described in terms of its attribute profile. As used herein, the term "attribute profile" may refer to a list of attributes of a target. In various examples, the attribute profile optionally provides a chemical identity or attribute type, e.g., deamidation, relative to the native structure of the target. In various examples, the attribute profile provides the location of the attribute, e.g., the location of the amino acid where the attribute is present. In some embodiments, the attribute profile provides a description of all attributes present in the target. In other embodiments, the attribute profile provides a description of some attributes present in the target. For example, the attribute profile may only provide attributes present in a particular portion of the target, e.g., the extracellular domain. A species of a target is characterized by the attributes present in the target. A species of a target may differ from another species of the same target by having different attribute profiles. If two targets have different attribute profiles, then the targets represent two different species of targets. If two targets have identical attribute profiles, then the targets are considered to be the same species of target.

[0023] In various examples, the target is placed in conditions that cause a change in its structure (e.g., the formation of one or more attributes), which changes the affinity between the therapeutic protein and the target. In various embodiments, the target is placed in conditions that cause a change in its structure (e.g., the formation of one or more attributes), which changes the affinity of the therapeutic protein for the target. In some embodiments, the decrease in affinity leads to a partial or total loss of the target's ability to interact (e.g., bind) with the therapeutic protein. In various examples, the partial or total loss of the target's ability to interact (e.g., bind) with the therapeutic protein ultimately reduces the potency of the target. In alternative examples, the target is placed in conditions that cause a change in its structure (e.g., the formation of one or more attributes), which does not change the affinity between the therapeutic protein and the target. In various embodiments, the change in structure does not reduce the affinity between the therapeutic protein and the target. Without being bound to any particular theory, the methods of the present disclosure advantageously accurately and precisely distinguish attributes of a target that affect the interaction between the target and a therapeutic protein from attributes that do not affect the interaction.

[0024] In various embodiments, the target is placed in conditions that lead to a change in its structure (e.g., a change in the structure of the amino acid of the target), leading to the formation of one or more attributes, thereby creating a species of the target. In various embodiments, the change in the structure of the target can be caused by a post-translational modification involving one or more intracellular enzymes, and thus the change in the structure of the target can be caused by an enzymatic modification. Enzymatic modifications are known in the art and include, for example, lipidation (e.g., palmitoylation, myristoylation), phosphorylation, glycosylation, ubiquitination, nitrosylation, methylation, acetylation, and proteolysis. In various embodiments, the amino acid sequence of a target having a post-translational modification is considered to be the same as that of an unmodified target, although it is altered by one or more amino acid residues including a moiety (e.g., lipid, sugar, peptide, phosphate group, ubiquitin tag, methyl, acetyl group). In various examples, the change in the structure of the target can be caused by a change in the environment of the protein. For example, the environment of a protein may change in pH, salinity, osmolality, pressure, temperature, exposure to light, UV light, etc., exposure to air or oxygen, agitation / shaking, exposure to chemicals or materials (e.g., metals, plastics), or long-term storage, and such environmental changes cause changes in the structure of one or more amino acids. The environmental changes may be, for example, changes in culture medium or formulation components, changes in storage conditions, etc. In some embodiments, the environmental changes occur during any one or more of the process steps leading up to administration to a patient (e.g., production (e.g., recombinant production), harvesting, purification, formulation, filling, packaging, storage, delivery, and final preparation of the protein just prior to administration to a patient). Exemplary modifications that may occur during manufacturing include, for example, removal of residues from the target polypeptide and / or cleavage mediated by proteases. Exemplary changes that may occur after production of the target but prior to administration (e.g., during packaging and / or filling, storage and / or transportation / delivery) include, for example, oxidation, reduction, deamidation, deamination, aggregation, denaturation, precipitation, hydrolysis, aspartate isomerization, N-terminal and C-terminal modifications.Thus, in various embodiments, the change in the structure of the target can be caused by a non-enzymatic modification (e.g., a chemical modification) that does not involve an intracellular enzyme; thus, the change in the structure of the target can be caused by a non-enzymatic modification (e.g., a chemical modification).

[0025] Samples and Stress With respect to the disclosed methods of identifying attributes of a therapeutic protein or target that affect an interaction between a therapeutic protein and a target, the method in an exemplary embodiment includes applying a stress to a first sample comprising the therapeutic protein or target, and contacting the first sample with a second sample comprising the target or therapeutic protein to form a mixture comprising: (i) the therapeutic protein-target complex, (ii) unbound therapeutic protein, and (iii) unbound target.

[0026] In various embodiments, the first sample comprises a population of species of therapeutic proteins. In various examples, the population is a homogenous population of therapeutic proteins, and optionally, each of the therapeutic proteins present in the first sample is the same species. In various examples, the population is a heterogeneous population comprising at least two different species of therapeutic proteins. In various embodiments, the heterogeneous population comprises at least three, at least four, at least five, at least six or more different species of therapeutic proteins. Optionally, the heterogeneous population comprises more than seven, more than eight, more than nine, more than ten, more than twenty, more than thirty, more than forty, more than fifty different species of therapeutic proteins. In some embodiments, each species of the population has a unique attribute profile. In illustrative examples, the species of therapeutic proteins are the only proteins present in the first sample. In some embodiments, the first sample comprises (i) a population of therapeutic proteins, and (ii) a pharma- ceutically acceptable carrier, diluent, excipient, or combination thereof. In various embodiments, the first sample preferably contains acceptable formulation components that are non-toxic to patients. The first sample optionally contains agents to modify, maintain or preserve, for example, pH, osmolality, viscosity, clarity, color, isotonicity, odor, sterility, stability, rate of dissolution or release, adsorption or permeability of the therapeutic protein in the first sample. In general, excipients can be classified based on their mechanism of stabilizing proteins against various chemical and physical stresses. Some excipients mitigate the effects of specific stresses or modulate the specific susceptibility of a particular polypeptide. Other excipients have a more general effect on the physical and covalent stability of proteins. Common excipients for liquid and lyophilized therapeutic protein formulations are shown in Table B (see also (Kamerzell, Esfandiary, Joshi, Middaugh, & Volkin, Adv Drug Deliv Rev 63(13):1118-1159 (2011))).

[0027] [Table 2]

[0028] [Table 3]

[0029] [Table 4]

[0030] Other excipients are known in the art (see, e.g., (Powell, Nguyen, & Baloian, PDA J Pharm Sci Technol 52(5):238-311 (1998))).

[0031] In various embodiments, the first sample comprises a population of therapeutic proteins and a buffer. In exemplary embodiments, the buffer is a buffer listed in Table C.

[0032] [Table 5]

[0033] In various examples, the first sample comprises one or more species of therapeutic proteins and serum or a serum fraction. The term "serum" as used herein refers to a fraction of blood remaining after clotting proteins and blood cells have been removed. "Serum fraction" refers to a portion of serum remaining after one or more components have been removed from the serum. In some embodiments, the serum fraction is a depleted fraction of serum. "Depleted fraction of serum" or "depleted fraction" as used herein means a fraction of serum from which one or more components have been removed. The term "whole serum" is serum from which components have not been removed. In exemplary embodiments, the first sample comprises whole serum. In exemplary embodiments, the whole serum is human serum, bovine serum (e.g., fetal bovine serum), rabbit serum, mouse serum, rat serum, cynomolgus monkey serum, horse serum, or pig serum. In a preferred embodiment, the whole serum is human serum. In exemplary embodiments, the depleted fraction of serum is an IgG-depleted serum fraction or a molecular weight range-depleted serum ("depleted fraction serum" or "depleted fraction of serum"). In exemplary embodiments, the IgG-depleted serum fraction is obtained by removing IgG from serum using Protein A, such as Protein A affinity chromatography. In exemplary embodiments, the depleted fraction of serum is a fraction depleted of molecules having a preselected molecular weight range. In exemplary examples, this preselected molecular weight range is about 30 kDa to about 300 kDa or higher. In various embodiments, the depleted fraction is obtained by size-based filtration, or centrifugation, or ultrafiltration (see, e.g., Kornilov et al., J Extracell Vesicles 7(1):1422674 (2018)). In various aspects, depleted serum is obtained from commercial suppliers (e.g., Thermo Fisher Scientific (Waltham, MA), CalBiochem® (Millipore Sigma, Burlington, MA), Quidel (San Diego, CA), and Complement Technologies (Tyler, TX). In exemplary aspects, the depleted fraction is a double-depleted fraction, optionally double-depleted of IgG or double-depleted of molecules having a preselected molecular weight."Doubly depleted" refers to a fraction of serum that has been subjected to a depletion or removal technique twice. In an illustrative example, a first sample containing serum or a serum fraction contains one or more serum proteins.

[0034] In various embodiments, the first sample comprises a population of target species. In various examples, the population is a homogenous population of targets, and optionally, each of the targets present in the first sample is of the same species. In various examples, the population is a heterogeneous population comprising at least two different species of targets. In various embodiments, the heterogeneous population comprises at least three, at least four, at least five, at least six or more different species of targets (including ranges between any two of the enumerated values, e.g., 2-4, 2-6, or 4-6 different species of targets). Optionally, the heterogeneous population comprises more than seven, more than eight, more than nine, more than ten, more than twenty, more than thirty, more than forty, more than fifty different species of targets (including ranges between any two of the enumerated values). In some embodiments, each species of the population has a unique attribute profile. In an illustrative example, the target species is the only protein present in the first sample.

[0035] In an exemplary embodiment, the method includes applying a stress to the first sample. In various examples, the stress can be any condition that leads to at least one change in the structure of an amino acid of the therapeutic protein or target, e.g., the stress can be any condition that leads to the formation of at least one attribute in an amino acid of the therapeutic protein or target. Optionally, the stress leads to a change in the structure of multiple amino acids of the therapeutic protein or target, e.g., the stress leads to multiple attributes (e.g., at least or about 2, at least or about 3, at least or about 4, at least or about 5, at least or about 6, at least or about 7, at least or about 8, at least or about 9, at least or about 10 or more attributes). In various examples, the stress results in the formation of one or more attributes that were not present in the therapeutic protein or target prior to the addition of the stress. Thus, in some aspects, the addition of the stress results in the formation of a species of the therapeutic protein or target that was not present in the first sample prior to the addition of the stress.

[0036] In various embodiments, the stress applied is (A) exposure to visible light, ultraviolet (UV) light, heat, air / oxygen, freeze / thaw cycles, shaking / agitation, chemicals and materials (e.g., metals, metal ions, chaotropic salts, surfactants, preservatives, organic solvents, plastics), molecules and cells (e.g., immune cells), or (B) a change in pH (e.g., a change of more than 1.0, 1.5, or 2.0), pressure, temperature, osmolality, salinity, or (C) long-term storage. In some embodiments, the change in temperature is at least or about 1° C., at least or about 2° C., at least or about 3° C., at least or about 4° C., at least or about 5° C., or more. The methods of the present disclosure are not limited to any particular type of stress. In exemplary embodiments, the stress is exposure to elevated temperatures, e.g., 25° C., 40° C., 50° C., optionally in the formulation. In exemplary examples, such exposure to elevated temperatures mimics an accelerated stress program. In some embodiments, the stress is exposure to visible and / or ultraviolet light, oxidizing reagents (e.g., hydrogen peroxide), air / oxygen, freeze / thaw cycles, shaking, long-term storage as a formulation under intended product storage conditions; mildly acidic pH (e.g., pH 3-4) or high pH (e.g., pH 8-9) simulating exposure to some purification conditions / steps. In some embodiments, the stress is a change in pH greater than 1.0, 1.5, 2.0, or 3.0. In exemplary embodiments, the stress is exposure to ultraviolet light, heat, air, freeze / thaw cycles, shaking, long-term storage, a change in pH, or a change in temperature, optionally with a change in pH greater than about 1.0 or greater than about 2.0, and optionally with a change in temperature greater than about 2 degrees Celsius or greater than about 5 degrees Celsius. In various embodiments, the stress applied to the first sample results in the formation of at least one species of therapeutic protein or target having a unique attribute profile relative to the attribute profile of the therapeutic protein or target prior to application of the stress. In various embodiments, the attribute is formed as a result of a chemical modification, which optionally alters the mass-to-charge ratio (m / z) of a charged ion of an amino acid of a therapeutic protein or target.Optionally, the chemical modification is, for example, glycosylation, hydroxylation, glycation, deamidation, oxidation, reduction, isomerization, aggregation, degradation, acetylation, or clipping by hydrolytic proteolysis.

[0037] Optionally, the stress degrades or dissociates about 5% to about 30%, about 10% to about 30%, about 15% to about 30%, about 20% to about 30%, about 25% to about 30%, about 5% to about 25%, about 5% to about 20%, about 5% to about 15%, or about 5% to about 10% of the complex formed between the therapeutic protein and the target. In various embodiments, the stress reduces the level of interaction between the therapeutic protein and its target. In some embodiments, the stress reduces the interaction by about 10% to about 50% (e.g., about 10% to about 45%, about 10% to about 40%, about 10% to about 35%, about 10% to about 30%, about 10% to about 25%, about 10% to about 20%, about 10% to about 15%) compared to the interaction in corresponding conditions without stress. In some embodiments, the stress reduces the K D Increase this K D is associated with weaker binding. In some embodiments, the stress increases the amount of unbound therapeutic protein by 10% to about 50% (e.g., about 10% to about 45%, about 10% to about 40%, about 10% to about 35%, about 10% to about 30%, about 10% to about 25%, about 10% to about 20%, about 10% to about 15%). Without being bound to a particular theory, the stress applied in the methods disclosed herein results in a faster and more robust and reproducible generation of therapeutic protein species resulting in abundant and diverse species for enhanced detection of species that may be produced during manufacturing, storage, and in human circulation (intravenous or subcutaneous space in a human subject).

[0038] With respect to the disclosed method of determining the effect of a known attribute present in a species of a therapeutic protein or target on an interaction between a therapeutic protein and a target, the method in an exemplary embodiment includes contacting a first sample containing a therapeutic protein with a known attribute or a target with a second sample containing the target or therapeutic protein (e.g., the first sample can contain a therapeutic protein and the second sample can contain a target, or the first sample can contain a target and the second sample can contain a therapeutic protein) to form a mixture containing (i) a therapeutic protein-target complex, (ii) unbound therapeutic protein, and (iii) unbound target. Optionally, the therapeutic protein with a known attribute is a therapeutic protein that has been engineered to have a particular attribute at a particular amino acid at a particular position within the amino acid sequence of the therapeutic protein. In various aspects, the known attribute has been engineered at a known amino acid position of the therapeutic protein prior to contacting with the second sample. Optionally, the therapeutic protein has been subjected to the addition of a stress known to cause the known attribute. Optionally, the target has been subjected to the addition of a stress known to cause the known attribute. The stress can be any of those described herein. In some embodiments, the therapeutic protein or target has an engineered attribute (e.g., oxidation (for M, W, H), deamidation (for N, Q), isomerization (for D)) at a site that is believed to be part of the binding site of the therapeutic protein to which the target binds, or the binding site of the target to which the therapeutic protein binds.

[0039] In various embodiments, a therapeutic protein includes two or more known attributes. In exemplary embodiments, a therapeutic protein includes two, three, four, five, six, seven, eight, nine, ten or more (including ranges between any two of the enumerated values, e.g., 2-10) known attributes. Optionally, a therapeutic protein includes one or more known attributes at a location hypothesized to be part of a paratope or binding site located within the therapeutic protein, the paratope or binding site being the site at which a target interacts with or binds to the therapeutic protein. In various embodiments, a therapeutic protein includes only one known attribute, but the method of determining the effect of a known attribute present in a species of therapeutic protein on an interaction between a therapeutic protein and a target is performed multiple times, each time using the same therapeutic protein with different known attributes. In various embodiments, the method of determining the effect of a known attribute present in a species of therapeutic protein on an interaction between a therapeutic protein and a target allows for mapping of a paratope or binding site of a therapeutic protein to a target. Thus, the present disclosure further provides a method of mapping a paratope for an interaction between a Therapeutic protein and a target, the paratope being in a Therapeutic protein, the method comprising determining the influence of a known attribute of the Therapeutic protein on the interaction between the Therapeutic protein and the target, according to either identifying an attribute of the Therapeutic protein that influences the interaction between the Therapeutic protein and the target or according to the presently disclosed method of determining the influence of the known attribute. If a structure (e.g., an attribute) of the Therapeutic protein negatively influences the interaction between the Therapeutic protein and the target, the structure is considered to be a candidate for being part of the paratope of the Therapeutic protein (e.g., at its residues). In some embodiments, the methods described herein further comprise generating a paratope map of the Therapeutic protein. The paratope map may include a graphical representation and / or a coordinate map, and may optionally be electronic.

[0040] In various embodiments, the target includes two or more known attributes. In exemplary embodiments, the target includes two, three, four, five, six, seven, eight, nine, ten or more (including ranges between any two of the enumerated values, e.g., 2-10) known attributes. Optionally, the target includes one or more known attributes at a location hypothesized to be part of an epitope or binding site located within the target, the epitope or binding site being the site at which a therapeutic protein interacts with or binds to the target. In various embodiments, the target includes only one known attribute, but the method of determining the effect of a known attribute present in a species of the target on the interaction between the target and a therapeutic protein is performed multiple times, each time using the same target with different known attributes. In various embodiments, the method of determining the effect of a known attribute present in a species of the therapeutic protein on the interaction between a therapeutic protein and the target allows for mapping of the epitope or binding site of the target to the therapeutic protein. Thus, the present disclosure further provides a method of mapping an epitope for an interaction between a Therapeutic protein and a target, the epitope being in the target, the method comprising determining the influence of a known attribute of the Therapeutic protein on the interaction between the Therapeutic protein and the target, according to either the steps of identifying an attribute of the Therapeutic protein that influences the interaction between the Therapeutic protein and the target or the methods of the present disclosure determining the influence of the known attribute.

[0041] In accordance with the methods of the present disclosure, a first sample is contacted with a second sample that includes a binding partner of a protein present in the first sample. In various embodiments, the first sample includes a therapeutic protein, and the second sample includes a target. Optionally, the second sample includes a population of targets. In illustrative examples, the target is a protein or a polypeptide or peptide. In various examples, the population of targets is a homogenous population, and optionally, each of the targets present in the second sample is the same (e.g., if the target is a protein, is chemically identical and includes the same amino acid sequence). In some embodiments, each of the targets of the second sample has the same amino acid sequence. In various examples, the population is a heterogeneous population that includes at least two different targets. In various embodiments, the heterogeneous population includes at least three, at least four, at least five, at least six or more different targets. Optionally, the heterogeneous population comprises more than 7, more than 8, more than 9, more than 10, more than 20, more than 30, more than 40, more than 50 different targets. In an exemplary example, the target is the only protein present in the second sample. In various embodiments, the target is the only protein present in the second sample. Optionally, the second sample comprises a pharma- ceutically acceptable carrier, diluent, excipient, or combinations thereof, such as one or more substances in Table B and / or Table C. In some embodiments, the second sample comprises a target and serum or a serum fraction. Optionally, the second sample comprising serum or a serum fraction comprises one or more serum proteins. In various examples, the first sample comprises a target, and the second sample comprises a therapeutic protein. In an exemplary example, the target is a protein or is a polypeptide or peptide. In various examples, the population of therapeutic proteins is a homogenous population, and optionally, each of the therapeutic proteins present in the second sample is the same (e.g., if the target is a protein, is chemically identical and comprises the same amino acid sequence). In some embodiments, each of the therapeutic proteins of the second sample has the same amino acid sequence. In various examples, the population is a heterogeneous population that includes at least two different targets.In various embodiments, the heterogeneous population comprises at least three, at least four, at least five, at least six or more different therapeutic proteins. Optionally, the heterogeneous population comprises more than seven, more than eight, more than nine, more than ten, more than twenty, more than thirty, more than forty, more than fifty different therapeutic proteins. In illustrative examples, the therapeutic protein is the only protein present in the second sample. In various embodiments, the therapeutic protein is the only protein present in the second sample.

[0042] separation In exemplary embodiments, the disclosed methods include separating the mixture into at least two fractions. In some aspects, the mixture is separated into several (e.g., 5, 6, 7, 8, 9, 10 or more) fractions. In some aspects, the separation step of the methods disclosed herein preserves the native folding, conformation and binding ability of the therapeutic protein and its target. In various aspects, the mixture is separated into an unbound fraction, which contains unbound therapeutic protein or target, and a bound fraction, which contains therapeutic protein-target complexes.

[0043] Suitable methods and techniques for separating mixtures into fractions are known in the art, see, e.g., Coskun, North Clin Istanb 3(2):156-160(2016); Snyder et al., Practical HPLC Method Development, 2nd ed., John Wiley&Sons, Inc. 1997; Snyder et al., Introduction to Modern Liquid Chromatography, John Wiley&Sons, Inc., Hoboken, NJ, 2010; Heftmann, Chromatography: Fundamentals and applications of chromatography and related differential migration methods, 6 thed., Volume 69A, Elsevier, Amsterdam, Netherlands, 2004; Mori and Barth, Size Exclusion Chromatography, Springer-Verlag, Berlin, 1999. In some embodiments, separation is based on charge, e.g., ion exchange chromatography, capillary isoelectric focusing (cIEF) and / or capillary zone electrophoresis (CZE), or based on hydrophobicity, e.g., reversed phase (RP; e.g., RP-HPLC) and hydrophobic interaction chromatography (HIC-HPLC). In various embodiments, separation is based on size, e.g., size exclusion chromatography (SEC; e.g., SE-HPLC), sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), capillary electrophoresis with sodium dodecyl sulfate (CE-SDS). The methods described herein are used to detect product oxidation of Met or Trp residues, fragmentation / clipping, isomerization of Asp, deamidation, formation of pyroglutamic acid at the N-terminus. In an exemplary embodiment, the mixture is separated into at least two fractions using a technique that separates the components of the mixture based on size, charge, hydrophobicity, affinity for a capture molecule, or a combination thereof. In various examples, the technique is size exclusion chromatography (SEC), affinity chromatography, precipitation using beads or cells, free-flow fractionation (FFF), ion exchange chromatography (IEX), hydrophobic interaction chromatography (HIC) or ultracentrifugation (UC). Optionally, the mixture is separated into at least two fractions using a technique that separates the components of the mixture based on size, and optionally, the technique is size exclusion chromatography (SEC).

[0044] In various embodiments, the mixture is separated into at least two fractions using a technique that separates components of the mixture based on their affinity to a capture molecule bound to a solid support, optionally beads or cells. In various examples, the mixture is separated by (i) adding the mixture to a container, e.g., a tube, containing beads or cells bound to capture molecules, the cells expressing the capture molecules on their surface, (ii) centrifuging the container (e.g., tube) to obtain a supernatant and a pellet, (iii) recovering the supernatant from the pellet to obtain an unbound fraction, (iv) releasing the bound fraction from the pellet with a solution, (v) centrifuging the container (e.g., tube) containing the pellet and solution to obtain a second supernatant containing the bound fraction and a second pellet containing the beads or cells, and (vi) recovering the second supernatant to obtain the bound fraction. In some embodiments, the mixture is separated by (i) adding the mixture containing the beads bound to the capture molecule to a column to obtain a flow-through and a bound fraction, (ii) collecting the flow-through to obtain an unbound fraction, and (iii) releasing the bound fraction from the beads with a solution and collecting a solution containing the bound fraction. Suitable solid supports include, for example, beads, resins, and papers, optionally made of cellulose, silica, alumina, glass, plastic, or combinations thereof. In an exemplary embodiment, the capture molecule bound to the solid support is a protein. The capture molecule can be the same as the target. Advantageously, the capture molecule is not limited to a specific molecule.

[0045] In an exemplary embodiment, the method includes contacting a first sample with a second sample that includes a target, where the target in the second sample is bound to a solid support, such as a resin, beads, the wall of a multi-well plate, or the like, or the target is expressed on the cell surface of cells in the second sample. Contacting the first sample with the second sample results in the formation of a mixture that includes a therapeutic protein-target complex (which then binds to the solid support) and unbound therapeutic protein or target. In an exemplary embodiment, the mixture is separated into at least two fractions by centrifuging the mixture (e.g., in a container, e.g., a tube) and collecting the unbound fraction when collecting the supernatant. In various embodiments, the bound fraction is collected by dissociating the therapeutic protein-target complex from the solid support, centrifuging the solid support and the complex, and collecting the bound fraction when collecting the supernatant. A schematic of this method is shown in FIG. 6A.

[0046] In an illustrative example, the method includes applying a first sample to a column (e.g., a chromatography column) that includes a target bound to a solid support (e.g., a resin, beads). In various embodiments, the first sample is applied to the column, and the therapeutic protein of the first sample binds to the target bound to the solid support to form a therapeutic protein-target complex (which is then bound to the solid support). In some embodiments, the flow-through represents the unbound fraction. In various examples, the bound fraction is collected by dissociating the therapeutic protein-target complex from the solid support and collecting the complex. A schematic of this method is shown in FIG. 6B. In an illustrative example, the method includes applying a first sample to a column (e.g., a chromatography column) that includes a therapeutic protein bound to a solid support (e.g., a resin, beads). In various embodiments, the first sample is applied to the column, and the target of the first sample binds to the therapeutic protein bound to the solid support to form a therapeutic protein-target complex (which is then bound to the solid support). In some embodiments, the flow-through represents the unbound fraction. In various instances, the bound fraction is collected by dissociating the therapeutic protein-target complex from the solid support and collecting the complex.

[0047] In an exemplary embodiment, the therapeutic protein comprises an antibody Fc domain and the target is an Fc receptor, see Figures 7A and 7B.

[0048] Identifying and Quantifying Attributes In an exemplary embodiment of a method for identifying attributes of a therapeutic protein or target that affect an interaction between a therapeutic protein and a target, for each of the unbound and bound fractions, the method comprises identifying and quantifying the abundance of each attribute present in the therapeutic protein or target species, where if the abundance of the attribute in the unbound fraction is greater than the abundance of the attribute in the bound fraction, the attribute negatively affects the interaction between the therapeutic protein and the target. In various aspects, the method comprises using a mass spectrometer to identify and quantify the abundance of each attribute of the therapeutic protein or target species in each of the unbound and bound fractions.

[0049] In an exemplary embodiment of a method for determining the effect of a known attribute present in a species of a therapeutic protein or target on an interaction between a therapeutic protein and a target, the method comprises quantifying the abundance of the known attribute for each of the unbound and bound fractions, where if the abundance of the known attribute in the unbound fraction is greater than the abundance of the known attribute in the bound fraction, the known attribute has a negative effect on the interaction between the therapeutic protein and the target. In various aspects, the method comprises quantifying the abundance of the known attribute in each of the unbound and bound fractions using a mass spectrometer.

[0050] In an exemplary embodiment, each of the unbound and bound fractions is subjected to a peptide digestion step, optionally a trypsin digestion step, prior to the quantification step or the identification and quantification step. Optionally, the method includes performing liquid chromatography / mass spectrometry (LC / MS). Protocols for LC-MS peptide mapping are known in the art. See, for example, Ren et al., Anal. Biochem. 392:12-21 (2009), Mouchahoir and Schiel, Anal Bioanl Chem 410(8):2111-2126 (2018). Briefly, a therapeutic protein undergoes a digestion process that includes reduction, alkylation, and digestion to generate peptides with attributes. In some examples, reduction is performed with dithiothreitol (DTT), alkylation is performed with 2-iodoacetamide (IAA), and digestion is performed with trypsin, such that all peptides generated terminate with Lys or Arg or the C-terminal residue of the therapeutic protein. Once generated, the peptides are loaded onto a column for the next chromatography step. In an exemplary embodiment, LC-MS is performed with a labeling reagent to quantify the abundance of the attribute. In an exemplary embodiment, the method does not utilize a labeling reagent in the LC-MS peptide mapping (e.g., the LC-MS peptide mapping is performed in the absence of a labeling agent). In an exemplary example, the measurement of abundance by LC-MS peptide mapping is performed as two analyses: one on the unbound fraction and one on the bound fraction. In an exemplary example, the two analyses are performed in parallel or simultaneously. Such an embodiment differs from prior art methods that include labeling the unbound fraction with one labeling reagent and labeling the bound fraction with a second reagent of a different mass, mixing the two samples, and performing a single sample preparation procedure and a single LC-MS analysis.

[0051] In exemplary embodiments, data obtained from LC-MS peptide mapping provides identification and quantification of the abundance of each species in each fraction. In exemplary examples, the mass of the peptides is accurately measured with an online mass spectrometer (MS). Optionally, the peptides are fragmented with a mass spectrometer and the masses of the fragments are also measured (MS / MS) to confirm the amino acid sequence and to identify and quantify the amino acid attributes of the peptides. In various embodiments, fragmentation occurs at all residues, providing a single amino acid sequence coverage. When different residues are modified on the same peptide, they have, in various embodiments, different elution times or masses or both. Thus, in some embodiments, they are selected for fragmentation (MS / MS) and the modifications of different residues on the same peptide are identified and quantified. In exemplary examples, when two attributes occur on the same peptide, modern instrumentation and data processing software can identify and quantify them. In some examples, the attributes are statistically independent. In some embodiments, attribute pairing effects have been reported, but they are less common (non-stochastic, non-random distribution of modifications). In some embodiments, the attributes are random.

[0052] In an exemplary embodiment, no cross-linking agents are used in any step of the method. Prior art methods use cross-linking as well as HDX, fast photochemical oxidation of proteins (FPOP) to assess protein conformation and detect the presence or absence of protein-bone or amino acid side chain hydrogen bonds or other non-covalent interactions.

[0053] Abundance ratio and statistical significance In an exemplary embodiment of the method of the present disclosure, for each attribute, a ratio is determined between the abundance of the attribute in the unbound fraction and the abundance of the attribute in the bound fraction. In various examples, if the ratio is greater than about 1, the attribute negatively affects the interaction between the therapeutic protein and the target. In various examples, if the ratio is greater than about 1.5, the attribute negatively affects the interaction between the therapeutic protein and the target. In various examples, if the ratio is greater than about 2.0, the attribute negatively affects the interaction between the therapeutic protein and the target. Optionally, the ratio is expressed as log2. In certain embodiments, if the log2 ratio is greater than about 0.58, the attribute affects the interaction between the therapeutic protein and the target. Optionally, if the log2 ratio is greater than about 1.0, the attribute affects the interaction between the therapeutic protein and the target.

[0054] In an exemplary embodiment, the method includes repeating the steps of the method, optionally repeating the separating and quantifying steps, where each time the steps are repeated, a ratio is determined for each attribute to obtain a plurality of ratios, and the method includes determining the statistical significance of the plurality of ratios. In some aspects, the method includes repeating the steps at least two, three or four times, optionally at least five or more times. In an exemplary aspect, the method including determining the statistical significance of the plurality of ratios includes calculating a p-value, optionally the p-value is calculated by a t-test. Methods for calculating a p-value by a t-test are known in the art. See, e.g., Greenland et al., Eur J Epidemiol 31:37-350 (2016). In various examples, a p-value less than 0.05 or a negative log 10 If the p-value is greater than 1.3, the attribute negatively influences the interaction (e.g., binding interaction) between the therapeutic protein and the target. In various examples, if the p-value is less than 0.0005 or has a negative log 10If the p-value is greater than 3.3, the attribute negatively impacts the interaction (e.g., binding interaction) between the therapeutic protein and the target. In some embodiments, the method comprising determining the statistical significance of a plurality of ratios comprises calculating a corrected p-value (p*) of these ratios, where p*=p / N, where p is the p-value and N is the total number of attributes tested. In an illustrative example, the method further comprises plotting the average or mean ratio of each attribute in a volcano plot, where the statistical significance of the ratio is plotted as a function of the ratio. In various embodiments, the abundance is expressed as a ratio and then plotted on a graph (e.g., a scatter graph). Volcano plots are known in the art. See, e.g., . In some embodiments, the abundance ratio is plotted for each attribute, and the graph includes all abundance ratios. In an illustrative embodiment, the abundance ratio is expressed on a log scale, e.g., a log2 scale, and the abundance ratio is expressed as the ratio of log2 abundance. FIG. 1B compares examples where abundance ratios are expressed as absolute values ​​versus log2. As shown in FIG. 1B, the log2 scale is more convenient and commonly used in statistics because it provides a symmetrical scale for ratios above and below 1. In exemplary embodiments, a species is identified as associated with an interaction between a therapeutic protein and a target if the ratio is greater than 1.5 (or greater than about 0.58 on the log2 scale). In exemplary embodiments, a species is identified as associated with an interaction if the ratio is greater than 2.0 (or greater than about 1 on the log2 scale). In some embodiments, the threshold or cutoff for determining associated attributes is done by selecting a cutoff where there are no false positive results located in the upper left quadrant of the volcano plot. See the area bounded by the dashed lines in FIG. 1B. For example, in FIG. 1B, the two vertical dashed lines of the fold change threshold are symmetrically moved outward until the upper left quadrant is clear. (-log 10 The horizontal threshold lines for the p-values ​​are moved upward until the top left quadrant is clear. In some instances, these thresholds define the ratios and p-values ​​in the top right quadrant where the attributes / modifications reduce binding statistically significantly.

[0055] Therapeutic protein type and affinity Advantageously, the methods of the present disclosure are not limited to any particular type of therapeutic protein involved in an interaction, e.g., a binding interaction, with a target. In an illustrative example, the therapeutic protein interacts with a target that is a protein. In various aspects, the therapeutic protein includes a protein that is native or naturally endogenously produced, e.g., by a mammal (e.g., a human). Alternatively, in some aspects, the therapeutic protein is a protein that is not found in nature or naturally endogenously produced by an organism, but is synthetic or engineered. In some aspects, the therapeutic protein is a hormone, growth factor, cytokine, cell surface receptor, or any ligand thereof. In some embodiments, the therapeutic protein is a fusion protein, a chimeric protein, a protein fragment, or a conjugate that includes a protein.

[0056] In an exemplary embodiment, the therapeutic protein comprises an antibody, an antigen-binding antibody fragment, or an antibody protein product. Collectively, antibodies form a family of plasma proteins known as immunoglobulins and contain immunoglobulin domains. (Janeway et al., Immunobiology: The Immune System in Health and Disease, 4 thed., Elsevier Science Ltd. / Garland Publishing, 1999). As used herein, the term "antibody" refers to a protein having the conventional immunoglobulin type, including heavy and light chains, and including variable and constant regions. For example, an antibody may be an IgG, which is a "Y-shaped" structure with two pairs of identical polypeptide chains, each pair having one "light" chain (usually having a molecular weight of about 25 kDa) and one "heavy" chain (usually having a molecular weight of about 50-70 kDa). An antibody has a variable region and a constant region. In the IgG structure, the variable region is generally about 100-110 or more amino acids, includes three complementarity determining regions (CDRs), and is primarily responsible for antigen recognition and differs substantially among other antibodies that bind different antigens. The constant region allows the antibody to recruit cells and molecules of the immune system. The variable region is made up of the N-terminal region of each light chain and heavy chain, while the constant region is made up of the C-terminal portion of each heavy chain and light chain (Janeway et al., "Structure of the Antibody Molecule and the Immunoglobulin Genes", Immunobiology: The Immune System in Health and Disease, 4 th ed.Elsevier Science Ltd. / Garland Publishing, (1999)).

[0057] The general structure and properties of antibody CDRs have been described in the art. Briefly, in the antibody framework, the CDRs are embedded within the framework in the variable regions of the heavy and light chains, where they constitute the regions that play a major role in antigen binding and antigen recognition. A variable region typically contains at least three heavy chain or light chain CDRs (Kabat et al., 1991, Sequences of Proteins of Immunological Interest, Public Health Service NIH, Bethesda, Md.; see also Chothia and Lesk, 1987, J. Mol. Biol. 196:901-917; Chothia et al., 1989, Nature 342:877-883), which are located within framework regions (termed framework regions 1-4, FR1, FR2, FR3 and FR4 by Kabat et al., 1991; see also Chothia and Lesk, 1987, supra).

[0058] The antibody may include any constant region known in the art. Human light chains are classified as kappa and lambda light chains. Heavy chains are classified as mu, delta, gamma, alpha, or epsilon, defining the antibody's isotype as IgM, IgD, IgG, IgA, and IgE, respectively. IgG has several subclasses, including, but not limited to, IgG1, IgG2, IgG3, and IgG4. IgM has subclasses, including, but not limited to, IgM1 and IgM2. The embodiments of the present disclosure include all such classes or isotypes of antibodies. The light chain constant region may be, for example, a kappa- or lambda-type light chain constant region, such as a human kappa-type or human lambda-type light chain constant region. The heavy chain constant region may be, for example, an alpha-type, delta-type, epsilon-type, gamma-type, or mu-type heavy chain constant region, such as a human alpha-type, human delta-type, human epsilon-type, human gamma-type, or human mu-type heavy chain constant region. Thus, in exemplary embodiments, the antibody is of the isotype IgA, IgD, IgE, IgG or IgM, including any one of IgG1, IgG2, IgG3 and IgG4.

[0059] The antibody may be a monoclonal or polyclonal antibody. In some embodiments, the antibody comprises a sequence substantially similar to a naturally occurring antibody produced by a mammal (e.g., mouse, rabbit, goat, horse, chicken, hamster, human, etc.). In this respect, the antibody may be considered a mammalian antibody (e.g., mouse antibody, rabbit antibody, goat antibody, horse antibody, chicken antibody, hamster antibody, human antibody, etc.). In certain aspects, the antibody is a human antibody. In certain aspects, the antibody is a chimeric antibody or a humanized antibody. The term "chimeric antibody" refers to an antibody that comprises domains from two or more different antibodies. A chimeric antibody may, for example, comprise a constant domain from one species and a variable domain from another species, or more commonly, may comprise stretches of amino acid sequences from at least two species. A chimeric antibody may also comprise domains from two or more different antibodies within the same species. The term "humanized" when used in reference to an antibody refers to an antibody having at least the CDR regions from a non-human source that has been engineered to have a structure and immune function more similar to a true human antibody than the original source antibody. For example, humanization can include grafting CDRs from a non-human antibody, such as a mouse antibody, onto a human antibody. Humanization can also include selected amino acid substitutions to make the non-human sequence more similar to a human sequence.

[0060] Antibodies can be cleaved into fragments by enzymes such as, for example, papain and pepsin. Papain cleaves antibodies to produce two Fab fragments and an Fc fragment. Pepsin cleaves antibodies to produce an F(ab')2 fragment and a pFc' fragment. In an exemplary embodiment of the present disclosure, the therapeutic protein comprises an antigen-binding antibody fragment. As used herein, the term "antigen-binding antibody fragment" refers to a portion of an antibody that can bind to the antigen of the antibody, also known as an "antigen-binding fragment" or "antigen-binding portion." In an exemplary example, the antigen-binding antibody fragment is a Fab fragment or an F(ab')2 fragment.

[0061] The structure of antibodies has been exploited to create a wide range of alternative forms, spanning a molecular weight range of at least about 12-150 kDa, from monomers (n=1) to dimers (n=2), trimers (n=3), tetramers (n=4) and potentially even higher valency (n) ranges, referred to herein as "antibody protein products." Antibody protein products include those based on the complete antibody structure and those that mimic antibody fragments that retain complete antigen-binding capacity, such as scFv, Fab and VHH / VH (discussed below). The smallest antigen-binding antibody fragment that retains an intact antigen-binding site is the Fv fragment, consisting entirely of the variable (V) region. Soluble, flexible amino acid peptide linkers are used to link the V region to the scFv (single chain variable fragment) fragment to stabilize the molecule, or constant (C) domains are added to the V region to generate Fab fragments [fragment, antigen-binding]. Both scFv and Fab fragments can be easily produced in host cells (e.g., prokaryotic host cells). Other antibody protein products include dimeric and multimeric antibody formats such as diabodies, triabodies and tetrabodies or minibodies (miniAbs), including disulfide bond stabilized scFv (ds-scFv), single chain Fab (scFab) and different types consisting of scFv linked to oligomerization domains. The smallest fragments are the VHH / VH of camelid heavy chain Abs and single domain Abs (sdAbs). The building blocks most frequently used to generate new antibody types are single chain variable (V)-domain antibody fragments (scFv), which contain V domains (VH and VL domains) from heavy and light chains linked by a peptide linker of about 15 amino acid residues. Peptibodies or peptide-Fc fusions are yet another antibody protein product. The structure of a peptibody consists of a biologically active peptide grafted to an Fc domain. Peptibodies are well described in the art, see, e.g., Shimamoto et al., mAbs 4(5):586-591 (2012).

[0062] Other antibody protein products include single chain antibodies (SCAs), diabodies, triabodies, tetrabodies, and bispecific or triabodies. Bispecific antibodies can be classified into five major classes: BsIgG, appended IgG, BsAb fragments, bispecific fusion proteins, and BsAb conjugates. See, e.g., Spiess et al., Molecular Immunology 67(2)Part A:97-106(2015).

[0063] In exemplary embodiments, the therapeutic protein comprises any one of these antibody protein products. In exemplary embodiments, the therapeutic protein is any one of: scFv, Fab VHH / VH, Fv fragment, ds-scFv, scFab, dimeric antibody, multimeric antibody (e.g., diabody, triabody, tetrabody), miniAb, camelid heavy chain antibody peptibody VHH / VH, sdAb, diabody; triabody; tetrabody; bispecific or trispecific antibody, BsIgG, appended IgG, BsAb fragment, bispecific fusion protein, and BsAb conjugate.

[0064] In an exemplary embodiment, the therapeutic protein is a bispecific T cell engager (BiTE) molecule. A BiTE molecule is a bispecific antibody construct or bispecific fusion protein that contains two antibody binding domains (or targeting regions) linked to each other. BiTEs are constructed from two single chain variable fragments (scFv) linked in tandem by a flexible linker (scFc-scFc, Huehls et al, Immunol Cell Biol 93(3):290-296(2015)). One arm of the molecule is engineered to bind to a protein found on the surface of cytotoxic T cells, while the other arm is designed to bind to a specific protein found primarily on tumor cells. When the BiTE molecule binds to both targets, it forms a bridge between the cytotoxic T cells and the tumor cells, allowing the T cells to recognize the tumor cells and fight them off by injecting harmful molecules. For example, the tumor-binding arm of the molecule can be modified to create different BiTE antibody constructs that target different types of cancer. The term "binding domain" in reference to a BiTE molecule refers to a domain that (specifically) binds / interacts / recognizes a given target epitope or a given target site on a target molecule (antigen). The structure and function of the first binding domain (tumor cell antigen recognition) and preferably also the structure and / or function of the second binding domain (cytotoxic T cell antigen) are based on the structure and / or function of an antibody (e.g., a full-length immunoglobulin molecule or a whole immunoglobulin molecule). For example, a BiTE molecule comprises a first binding domain characterized by the presence of three light chain CDRs (i.e., CDR1, CDR2 and CDR3 of the VL region) and three heavy chain CDRs (i.e., CDR1, CDR2 and CDR3 of the VH region). Preferably, the second binding domain also comprises the minimal structural requirements of an antibody to allow target binding. More preferably, the second binding domain comprises at least three light chain CDRs (i.e. CDR1, CDR2 and CDR3 of the VL region) and / or three heavy chain CDRs (i.e. CDR1, CDR2 and CDR3 of the VH region).It is envisaged that the first and / or second binding domains are generated or obtained by phage display or library screening techniques, rather than grafting onto a scaffold of CDR sequences from an existing (monoclonal) antibody. A binding domain may typically comprise an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH), but need not comprise both. An Fd fragment, for example, has two VH regions and often retains some antigen-binding function of an intact antigen-binding domain. Examples of (modified) antigen-binding antibody fragments include: (1) a Fab fragment, which is a monovalent fragment having the VL, VH, CL, and CH1 domains; (2) an F(ab')2 fragment, which is a bivalent fragment having two Fab fragments linked by a disulfide bridge at the hinge region; (3) an Fd fragment having two VH and CH1 domains; (4) an Fv fragment having the VL and VH domains of a single arm of an antibody, (5) a dAb fragment having a VH domain (Ward et al., (1989) Nature 341:544-546); (6) isolated complementarity determining regions (CDRs), and (7) single-chain Fv (scFv), the latter of which is preferred (e.g., from an scFV library).

[0065] In exemplary embodiments, the therapeutic protein is a bispecific antibody, a trispecific antibody, a molecule, or an Fc fusion protein. In some embodiments, the therapeutic protein comprises a single chain variable fragment (scFv) that binds to a first target, and optionally, the therapeutic protein comprises a second scFv that binds to a second target (optionally different from the first target).

[0066] In an illustrative example, the Kd of a therapeutic protein-target interaction (e.g., a binding interaction) is about 10 -11 M~about 10 -9 and affinity purification techniques are SEC, precipitation using bound target-labeled beads or precipitation with cells with surface-expressed targets. Fc-containing therapeutic proteins also interact with Fc receptors, and this interaction occurs at approximately 10 -8 ~about 10 -6FcRn and Fc gamma are probably the most important Fc receptors. The FcRn and Fc gamma receptors are approximately -8 ~about 10 -6 It binds weakly to Fc-containing proteins with a Kd of M. FcRn receptors are present on the surface of many cells, allowing internalization and recycling of Fc-containing proteins to the surface, extending their lifespan in the circulation. Fc gamma receptors are present on cells of the immune system and are involved in antibody-dependent cellular cytotoxicity (ADCC), target recognition and killing by immune cells. Precipitation by FcRn-labeled beads, Fc gamma receptor-labeled beads, or by cells with surface-expressed FcRn and Fc gamma receptors. Complexes of Fc-containing proteins and Fc receptors bind weakly and typically dissociate on the column after injection into the SEC column, are diluted on the column, do not associate, and elute separately. When dissociated after injection into the SEC column, the mixture is diluted and the possibility of association is reduced. On the other hand, interaction with Fc receptors immobilized on beads in tubes (containers) or affinity columns maintains the original high concentration of protein and Fc receptor mixture, allowing both dissociation and association.

[0067] In exemplary embodiments, the therapeutic protein comprises an antibody, an antigen-binding antibody fragment or an antibody protein product, optionally a bispecific T cell engager (BiTE), a bispecific antibody, a trispecific antibody, an Fc fusion protein. In some embodiments, the therapeutic protein is a BiTE® molecule. BiTE® molecules are engineered bispecific monoclonal antibodies that direct the cytotoxic activity of T cells against cancer cells. They are a fusion of two single chain variable fragments (scFv) of different antibodies or amino acid sequences from four different genes on a single peptide chain of about 55 kilodaltons. One of the scFvs binds to T cells via the CD3 receptor and the other binds to tumor cells via a tumor-specific molecule. Blinatumomab (BLINCYTO®) is an example of a BiTE® molecule that is specific for CD19. Modified BiTE® molecules (e.g., those modified to extend half-life) can also be used in the methods of the disclosure.

[0068] In various embodiments, the therapeutic protein is Activase® (alteplase); alirocumab, Aranesp® (darbepoetin-alpha), Epogen® (epoetin alpha or erythropoietin); Avonex® (interferon beta-la); Bexxar® (tositumomab); Betaseron® (interferon beta); bococizumab (anti-PCSK9 monoclonal antibody designated L1L3, see U.S. Pat. No. 8,080,243); Campath® (alemtuzumab); Dynepo® (epoetin delta); Velcade® (bortezomib); MLN0002 (anti-alpha4beta7 mAb); MLN1202 (anti-CCR2 chemokine receptor mAb); Enbrel® (etanercept); Eprex® (epoetin alfa); Erbitux® (cetuximab); evolocumab; Genotropin® (somatropin); Herceptin® (trastuzumab); Humatrope® (somatropin [rDNA-derived] injection); Humira® (adalimumab); Infergen® (interferon alfacon-1); Natrecor® (nesiritide); Kineret® (anakinra), Leukine® (sargamostim); Lymph oCide® (epratuzumab); Benlysta™ (belimumab); Metalyse® (tenecteplase); Mircera® (methoxypolyethylene glycol epoetin beta); Mylotarg® (gemtuzumab ozogamicin); Raptiva® (efalizumab); Cimzia® (certolizumab pegol); Soliris™ (eculizumab); pexelizumab (anti-C5 complement); MEDI-524 (Numax®); Lucentis® (ranibizumab); edrecolomab (Panorex®); Trabio® (lerdelimumab); TheraCim hR3 (nimotuzumab); Omnitarg (pertuzumab,2C4; Osidem® (IDM-1); OvaRex® (B43.13); Nuvion® (Vizilizumab); Cantuzumab mertansine (huC242-DMl); NeoRecormon® (Epoetin beta); Neumega® (Oprelvekin); Neulasta® (PEGylated filgastrim, PEGylated G-CSF, PEGylated hu-Met-G-CSF); Neupogen® (filgrastim); Orthoclone OKT3® (muromonab-CD3), Procrit® (epoetin alfa); Remicade® (infliximab), Reopro® (abciximab), Actemra® (anti-IL6 receptor mAb), Avastin® (bevacizumab), HuMax-CD4 (zanolimumab), Rituxan® (rituximab); Tarceva® (erlotinib); Roferon-A® (interferon alfa-2a); Simulect® (basilisulfamethasone); ximab); Stelara™ (ustekinumab); Prexige® (lumiracoxib); Synagis® (palivizumab); 146B7-CHO (anti-IL15 antibody, see U.S. Pat. No. 7,153,507), Tysabri® (natalizumab); Valortim® (MDX-1303, an anti-B. anthracis protective antigen mAb); ABthrax™; Vectibix® (panitumumab); Xolair® (omalizumab), ETI211 (anti-MRSA mAb), IL-1 Trap (the Fc portion of human IgG1 and the extracellular domains of both IL-I receptor components (type 1 receptor and receptor accessory protein)), VEGF Trap (the Ig domain of VEGFRI fused to IgG1 Fc), Zenapax® (daclizumab); Zevalin® (ibritumomab tiuxetan), Zetia (ezetimibe), atacicept (TACI-Ig), anti-α4β7 mAb (vedolizumab); galiximab (anti-CD80 monoclonal antibody),Anti-CD23 mAb (lumiliximab); BR2-Fc (huBR3 / huFc fusion protein, soluble BAFF antagonist); Simponi™ (golimumab); mapatuzumab (human anti-TRAIL receptor-1 mAb); ocrelizumab (anti-CD20 human mAb); HuMax-EGFR (zalutumumab); M200 (volociximab, anti-α5β1 integrin mAb); MDX-010 (ipilimumab, anti-CTLA-4 mAb and VEGFR-I (IMC-18F1); anti-BR3 mAb; anti-C. difficile toxin A and toxin BC mAbs MDX-066 (CDA-I) and MDX-1388); anti-CD22 dsFv-PE38 conjugates (CAT-3888 and CAT-8015); anti-CD25 mAb (HuMax-TAC); anti-TSLP antibody; anti-TSLP receptor antibody (U.S. Pat. No. 8,101,182); anti-TSLP antibody designated A5 (U.S. Pat. No. 7,982,016); (anti-CD3 mAb (NI-0401)); adecatumumab (MT201, anti-EpCAM-CD326 mAb); MDX-060, SGN-30, SGN-35 (anti-CD30 mAb); MDX-1333 (anti-IFNAR); HuMax CD38 (anti-CD38 mAb); anti-CD40L mAb; anti-Cripto mAb; anti-CTGF idiopathic pulmonary fibrosis stage I fibrogen (FG-3019); anti-CTLA4 mAb; anti-eotaxin 1 mAb (CAT-213); anti-FGF8 mAb; anti-ganglioside GD2 mAbs; anti-sclerostin antibodies (see U.S. Pat. Nos. 8,715,663 or 7,592,429), anti-sclerostin antibodies designated Ab-5 (U.S. Pat. Nos. 8,715,663 or 7,592,429); anti-ganglioside GM2 mAb; anti-GDF-8 human mAb (MYO-029); anti-GM-CSF receptor mAb (CAM-3001); anti-HepC mAb (HuMax HepC); MEDI-545,MDX-1103 (anti-IFNα mAb); anti-IGFIR mAb; anti-IGF-IR mAb (HuMax-Inflam); anti-IL12 / IL23p40 mAb (briakinumab); anti-IL-23p19 mAb (LY2525623); anti-IL13 mAb (CAT-354); anti-IL-17 mAb (AIN457); anti-IL2Ra mAb (HuMax-TAC); anti-IL5 receptor mAb; anti-integrin receptor mAb (MDX-Ol8, CNTO95); anti-IPIO ulcerative colitis mAb (MDX-1100); anti-LLY antibody; BMS-66513; anti-mannose receptor / hCGβ mAb (MDX-1307); anti-mesothelin dsFv-PE38 conjugate (CAT-5001); anti-PDlmAb (MDX-1 106 (ONO-4538); anti-PDGFRα antibody (IMC-3G3); anti-TGFβ mAb (GC-1008); anti-TRAIL receptor-2 human mAb (HGS-ETR2); anti-TWEAK mAb; anti-VEGFR / Flt-1 mAb; anti-ZP3 mAb (HuMax-ZP3); NVS antibody #1; NVS antibody #2 or amyloid beta monoclonal antibody or an antibody listed in Table D. Optionally, the therapeutic protein is selected from the group consisting of infliximab, bevacizumab, cetuximab, ranibizumab, palivizumab, abagovomab, actoxumab, adalimumab, afelimomab, afutuzumab, alacizumab, alacizumab pegol, ald518, alemtuzumab, alirocumab, altumomab, amatuximab, anatumomab mafenatox, anrukinzumab, apolizumab, arcitumomab, acelizumab, altinumab, atlizumab, atolimumab, tocilizumab, bapineuzumab, basiliximab, bavituximab, Bectumomab, belimumab, benralizumab, bertilimumab, besilesomab, bevacizumab, bezlotoxumab, biciromab, bivatuzumab, bivatuzumab mertansine, blinatumomab, brosozumab, brentuximab vedotin, briakinumab, brodalumab, canakinumab, cantuzumab mertansine, caplacizumab, capromab pendetide, carlumab, catumaxomab, cc49, cedelizumab, certolizumab pegol, cetuximab, sitatuzumab bogatox, cixutumumab, clazakizumab, clenoliximab,clivatuzumab tetraxetan, conatumumab, crenezumab, cr6261, dacetuzumab, daclizumab, darotuzumab, daratumumab, demcizumab, denosumab, detumomab, dorlimomab aritox, drozitumab, durigotumab, dupilumab, ecromeximab, eculizumab, edovacomab, edrecolomab, efalizumab, efungumab, elotuzumab, elsilimomab, enabatuzumab, enlimomab pegol, enokizumab, enoticumab, ensituximab, epitumomab situxetan, epratuzumab, erenumab, er Lisumab, Ertumaxomab, Etaracizumab, Etrolizumab, Evolocumab, Exibirumab, Fanolesomab, Faralimomab, Farletuzumab, Fasinumab, Fbta05, Felvizumab, Fezakinumab, Ficlatuzumab, Figitumumab, Franvotumab, Fontolizumab, Foralumab, Foravirumab, Fresolimumab, Furlanumab, Futuximab, Galiximab, Ganitumab, Gantenerumab, Gavilimomab, Gemtuzumab ozogamicin, Gevokizumab, Girentuximab, Glembatumumab vedotin, Golimumab, Go miriximab, gs6624, ibalizumab, ibritumomab tiuxetan, icrucumab, igovomab, imciromab, imgatuzumab, inlacumab, indatuximab ravtansine, infliximab, intetumumab, inolimomab, inotuzumab ozogamicin, ipilimumab, iratumumab, itolizumab, ixekizumab, keliximab, labetuzumab, lebrikizumab, remaresomab, lerdelimumab, lexatumumab, ribivirumab, ligelizumab, lintuzumab, lirilumab, lorvotuzumab mertansine, lucatumumab, rumili Ximab, mapatuzumab, maslimomab, mavrilimumab, matuzumab, mepolizumab, metelimuamab, milatuzumab, minretumomab, mitumomab, mogamulizumab, morolimumab, motavizumab, moxetumomab passudotox, muromonab-cd3, nacolomab butafenatox, namilumab, naptumomab estafenatox, narutumumab, natalizumab, nebacumab, necitumumab, nerelimomab, nesbacumab, nimotuzumab, nivolumab, nofetumomab merpentane, ocaratuzumab, ocrelizumab, odulimomab, ofatumumab,Olaratumab, Olokizumab, Omalizumab, Onartuzumab, Oportuzumab monatox, Oregovomab, Olticumab, Otelixizumab, Oxelumab, Ozanezumab, Ozolarizumab, Pazibaximab, Palivizumab, Panitumumab, Panobacumab, Palsatuzumab, Pascolizumab, Pateclizumab, Patritumab, Panitumumab, Panobacumab , palsatuzumab, pascolizumab, pateclizumab, patritumab, pemtumomab, perakizumab, pertuzumab, pexelizumab, pidilizumab, pintumomab, placumab, ponezumab, priliximab, pritumumab, PRO140, quilizumab, racotumomab, radletumab, rafivirumab, ramucirumab, ranibizumab, raxivac, Mab, regavirumab, reslizumab, rilotumumab, rituximab, lobatumumab, loredumab, romosozumab, rontalizumab, rovelizumab, ruplizumab, samalizumab, sarilumab, satumomab pendetide, secukinumab, sevirumab, sibrotuzumab, sifalimumab, siltuximab, simtuzumab, siplizumab sirukumab, solanezumab, solitomab, sonepcizumab, sontuzumab, stamulumab, sulesomab, subizumab, tabalumab, tacatuzumab tetraxetan, tadocizumab, talizumab, tanezumab, taplitumomab paptox, tefibazumab, terimomab alitox, tenatumomab, tefibazumab, tenelixima , teplizumab, teprotumumab, tezepelumab, TGN1412, tremelimumab, ticilimumab, tildrakizumab, tigatuzumab, TNX-650, tocilizumab, toralizumab, tositumomab, tralokinumab, trastuzumab, TRBS07, tregalizumab, tucotuzumab-celmoleukin, tuvilumab, ublitu These are: ximab, urelumab, urtoxazumab, ustekinumab, bapaliximab, batelizumab, vedolizumab, veltuzumab, beparimomab, besencumab, visilizumab, volociximab, borsetuzumab mafodotin, votumumab, zalutumumab, zanolimumab, zatuximab, diralimumab, and zolimomab-alitox.

[0069] [Table 6]

[0070] [Table 7]

[0071] [Table 8]

[0072] In various embodiments, the therapeutic protein comprises an Fc domain. In various embodiments, the therapeutic protein has a relatively weak affinity for the target. In various embodiments, the therapeutic protein has a relatively weak affinity for the target. -8 ~about 10 -6 M is the equilibrium dissociation constant for the target (K D Optionally, the mixture is separated into at least two fractions by precipitation using beads or cells bound to a capture molecule. In some embodiments, the capture molecule is FcRn or an Fc gamma receptor. Fc-containing therapeutic proteins also interact with Fc receptors, and this interaction occurs at about 10 -8 ~about 10 -6 M. FcRn and Fc gamma are in some embodiments very important Fc receptors. FcRn and Fc gamma receptors are about 10 -8 ~about 10 -6It binds weakly to Fc-containing proteins with a Kd of 10 M. FcRn receptors are present on the surface of many cells, allowing internalization and recycling of Fc-containing proteins to the surface, extending their lifespan in the circulation. Fc gamma receptors are present on cells of the immune system and are involved in antibody-dependent cellular cytotoxicity (ADCC), the recognition and killing of targets by immune cells. Precipitation can also be performed using, for example, FcRn-labeled beads, Fc gamma receptor-labeled beads, or cells with surface-expressed FcRn and Fc gamma receptors. Complexes of Fc-containing proteins and Fc receptors bind weakly and typically dissociate on the column after injection into the SEC column, are diluted on the column, do not associate, and elute separately. When dissociated after injection into the SEC column, the mixture is diluted and the probability of association is reduced. On the other hand, interaction with Fc receptors immobilized on beads in tubes (containers) or affinity columns maintains the original high concentration of protein and Fc receptor mixture, allowing both dissociation and association. Approximately 10 -8 ~about 10 -6 K of M D In an illustrative example of an Fc-containing therapeutic protein that weakly interacts with an Fc receptor having a nucleotide sequence similar to that of the Fc receptor, the method includes (i) adding the mixture to a container, e.g., a tube, containing beads or cells bound to a capture molecule, the cells expressing the capture molecule on their surface, (ii) centrifuging the tube containing the mixture and the beads or cells, and (iii) removing the supernatant from the pellet to obtain a first fraction and a second fraction, where optionally the capture molecule is an Fc receptor (FcRn or Fcγ).

[0073] In various embodiments, the therapeutic protein comprises a single chain variable fragment (scFv) that binds to a target, and in some embodiments, the therapeutic protein comprises a K D About 10 -11 ~about 10 -9 M. Optionally, the mixture is separated into at least two fractions by SEC or precipitation using beads or cells. In an illustrative example, the K of a therapeutic protein-target interaction (e.g., a binding interaction) is D is about 10 -11 M~about 10-9 and the separation technique is SEC, precipitation using bound target-labeled beads or precipitation with cells with surface-expressed target.

[0074] Epitope mapping procedure The described method can also be used for epitope mapping. "Epitope mapping" refers to the identification of binding residues on a target protein. To do so, the target protein is stressed under the same or similar stress conditions as the therapeutic antibody. The remaining experimental flow can be the same as that described for the paratope mapping method, except that bound target (in the antibody-target complex) fraction and unbound target fraction are collected and compared. If the structure (e.g., attribute) of the target negatively affects the interaction between the target and the therapeutic protein, the structure is considered to be a candidate for being part of the epitope of the target (e.g., at that residue).

[0075] Alternative methods for epitope / paratope mapping Epitope / paratope mapping results in the identification of residues of an antigen (epitope) that are bound by residues of an antibody (paratope). Mapping is a useful tool for understanding antibody-target interactions, elucidation of the stoichiometry of the interaction, and mechanisms of drug action. This knowledge will facilitate design optimization for the next generation of therapeutic antibodies, including optimization of binding strength, stoichiometry, mechanisms of blocking or internalization with cell surface receptors. Several mapping techniques have traditionally been utilized for epitope / paratope mapping, including X-ray crystallography (Luo et al., 2015); nuclear magnetic resonance (NMR) spectroscopy (Blech et al., 2013); scanning mutagenesis followed by binding assays (Blech et al., 2013); mass spectrometry-based techniques including deuterium exchange (HDX) (Englander, 2006), (Sevy et al., 2013); fast photochemical oxidation of proteins (FPOP) (Jones et al., 2011); and differential chemical modification (Dhungana et al., 2009) (Abbott et al., 2014), (Opuni et al., 2018), (Goswami et al., 2018).

[0076] X-ray crystallography provides the most accurate determination of interactions between the side chains and main chain atoms of individual amino acids in both the epitope and the paratope. Amino acids that are within 5 Å of each other are generally considered to be contact residues. However, proximity does not guarantee binding, and alternative techniques such as NMR (Blech et al., 2013) or scanning mutagenesis (Englander, 2006) are required to identify binding residues with more confidence. Also, obtaining diffraction quality crystals is often a laborious and lengthy process, and sometimes impossible for some antibody-antigen pairs. NMR is another technique for high-resolution epitope mapping (Blech et al., 2013). However, it is a complex method, requiring prior crystal structure determination of the free antigen and expression in mammalian or microbial systems with isotopically labeled nutrients. 15 N, 13 C and potentially 2A complete resonance assignment of the H-labeled protein is required.

[0077] Scanning mutagenesis followed by binding assays of either antigen or antibody is an analytically and technically simpler technique. It analyzes binding to mutant forms of antigen or antibody. Loss of binding to a particular mutant is interpreted as the mutated amino acid constituting the epitope or paratope. Mutagenesis also offers the advantage of screening hundreds of proteins by using alanine scanning or shotgun mutagenesis. This technique can potentially provide epitope / paratope information at the amino acid level, but has the problem of uncertainty in not knowing whether the mutation really affected a key interacting residue or merely disrupted the folding of the mutant protein (Abbott et al., 2014). Without being limited by theory, in the case of paratope mapping of antibodies, scanning mutagenesis (or chemical modification of amino acids) must not change the folding and must be more precise than in the case of antigens. This is because the paratope is usually located in the CDRs, which are unfolded, unstructured, flexible loops. Therefore, amino acid substitutions or chemical modifications of side chains are expected to have a low impact on folding. Described herein is paratope / epitope mapping by SEC of stressed antibody-target complexes followed by characterization by mass spectrometry, a technique somewhat similar to mutagenesis, with the following important differences: instead of mutations, the side chains of amino acid residues are altered by chemical modifications, and instead of one-by-one binding assay measurements, this method allows for the characterization of molecules with chemical modifications on different residues in parallel.

[0078] Epitope / paratope mapping by HDX-MS compares the hydrogen-deuterium exchange rate of each backbone amide hydrogen in the antibody-antigen complex with that in the free antigen or antibody. The slower exchanging residues in the complex are due to the added protection by the conjugation. In the FPOP method, hydrogen peroxide, added at low concentration to the protein solution, is photolyzed with a laser to form hydroxyl radicals, which chemically modify the side chains of surface-exposed residues. The slower exchanging residues in the complex are due to the added protection by the conjugation. The use of a pulsed laser and radical scavengers limits the radical lifetime and ensures labeling on a microsecond time scale, which is faster than most protein unfolding. This is to ensure that the loss of interaction is not caused by unfolding. The principle of the differential chemical modification method is the differential surface exposure and reagent accessibility of the protein antigen in the presence and absence of the antibody, which determines the rate and extent of chemical modification of amino acid residues on the antigen (Dhungana et al., 2009). In all mass spectrometry based techniques, amino acid modifications are performed after conjugation and not before, as in the newly proposed method.

[0079] The following examples are presented merely to illustrate the present invention and are not intended to limit the scope of the invention in any way. EXAMPLES

[0080] Example 1 This example is an exemplary method of the disclosure and describes a method in which the therapeutic protein is an IgG antibody specific for antigen 1 and the target is a soluble form of the carbohydrate binding domain (CBD) of antigen 1.

[0081] Materials and Methods material Antibody 1 is an IgG antibody with a molecular weight of 149 kDa. Antibody 1 was formulated at a concentration of 70 mg / ml in a solution containing 10 mM acetic acid (pH 5.2) and 9% sucrose. The target of Antibody 1 used in the experiments was a 17,043.5 Da soluble portion of the carbohydrate binding domain (CBD) of Antigen 1 tagged with six His residues. The target was formulated at a concentration of 0.81 mg / mL in a solution containing 30 mM HEPES (pH 7.6), 0.15 M NaCl and 3 mM CaCl2 and stored at -70°C before use.

[0082] Light stress protocol (P2) Antibody 1 was photostressed by exposure to cool white light and ultraviolet (UV) light to induce instability leading to the formation of species and therefore different species of Antibody 1. Samples were subjected to photostress protocol P2 according to the International Council for Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) guideline ICH Q1B “Stability testing, photostability testing of new drug substances and products” (available at http: / / academy.gmp-compliance.org / guidemgr / files / MEDIA412.PDF). Briefly, Antibody 1 samples in glass vials were exposed to approximately 1.2 million lux hours of illumination consisting of both cool white fluorescent and at least 200 watt-hours / square meter of UV irradiance. All samples were maintained at 25°C. The cool white fluorescent lamps were set at 8 klux and the incubation time was 1.2 million lux hours / 8 klux=150 hours. For UV irradiation, the UV lamps were set at 10 watts per square meter and the incubation time was 200 watt-hours per square meter / 10 watts per square meter = 20 hours. The light stress protocol P2, which includes UV irradiation, was harsh and more than the typical stress experienced by therapeutic proteins during production, purification, formulation, storage and human circulation. However, it can generate light-induced oxidation at solvent-exposed tryptophan, methionine and histidine residues and is therefore useful.

[0083] Mixture of Antibody 1 with its target Antibody 1 (38 μL of 0.25 mg / mL of Antibody 1 solution) and target (2 μL of 0.81 mg / mL of AGSR1 solution) were mixed in a 1:2 molar ratio. It has been previously demonstrated that an unstressed antibody molecule binds two target molecules, which is typical for a bivalent antibody (such as one with two Fab arms). This mixture was incubated in an HPLC vial at 37° C. for 10 min. A 20 μL aliquot of the mixture was then injected onto an HPLC column for SEC separation, as described below.

[0084] Size Exclusion Chromatography (SEC) and Fraction Collection After incubation, the antibody 1-target mixture was separated by SEC using a G3000SWxl TOSOH Bioscience, 7.8 mm ID x 30 cm column (cat. no. 08541, TOSOH Bioscience, San Francisco, CA), the mobile phase included 100 mM sodium phosphate and 250 mM NaCl, pH 6.8. The flow rate was set at 0.5 ml / min, the column temperature was set at 37°C, the run time was set at 35 min, and the autosampler was set at 4°C. Ultraviolet / visible spectroscopy (UV / VIS) detection was performed at 214 nm and 280 nm. The elution fractions were collected using a filter with a molecular weight cutoff of >10 kDa and eluted with 7.5 M guanidine elution buffer. The eluted fractions were subjected to sample preparation for peptide mapping as described below.

[0085] Peptide mapping Peptide mapping of collected fractions was performed using a sample preparation procedure including refolding with guanidine, reduction and alkylation of disulfide bonds, buffer exchange and digestion with trypsin to peptides suitable for LC-MS analysis as described in (Ren et al., Anal. Biochem. 392:12-21 (2009)). Briefly, samples containing antibody 1 were diluted to approximately 1 mg / ml in 0.5 ml of denaturing buffer (7.5 M guanidine hydrochloride (GdnHCl) and 0.25 M Tris) at pH 7.5. Reduction was achieved by adding 3 μl of 0.5 M dithiothreitol (DTT) followed by incubation at room temperature for 30 min. Carboxymethylation was achieved by adding 7 μl of 0.5 M iodoacetic acid (IAA). The reaction was carried out at room temperature for 15 min in the dark. Excess IAA was quenched by adding 4 μl of 0.5 M DTT. Reduced and alkylated antibody 1 samples were buffer exchanged into digestion buffer (0.1 M Tris or 0.1 M ammonium bicarbonate) at pH 7.5 using a NAP-5 column (GE Healthcare, Piscataway, NJ, USA). Lyophilized trypsin was dissolved in water to a final concentration of 1 mg / ml. Digestion was initiated by adding 1 mg / ml trypsin solution to the reduced, alkylated and buffer exchanged antibody 1 samples to achieve an enzyme / substrate ratio of 1:25. Digestion was carried out for 30 min at 37°C. The final digest was quenched by the addition of 5 tl of 20% FA. LC-MS / MS peptide mapping analysis of digested antibody 1 samples was performed as described (Ren et al., 2009, supra) on an Agilent 1290 UHPLC system coupled to a Thermo Scientific Q-Exactive Biopharma mass spectrometer. The resulting LC-MS / MS raw data and the sequences of Antibody 1 and the target were used to identify and quantitate the modifications using MassAnalyzer software (Zhang, Anal. Chem. 81:8354-8364 (2009)).

[0086] result To evaluate the relevance of the formed attributes to the interaction between Antibody 1 and its target, an Antibody 1 sample was mixed with its target (Figure 2). The first eluting SEC peak (28-29 min) contains two target molecules bound to Antibody 1 (bound fraction), whereas the second eluting SEC peak (29-30 min) contains Antibody 1 with one arm not bound to the target and nullified by attributes (Figure 2A) (partially unbound fraction). The third SEC peak (30.5-31.5 min) was taken as completely unbound antibody since its elution time was the same as that of the control-Antibody 1 reference material (completely unbound fraction). The amount of material in the third fraction (completely unbound antibody) was much less than the first fraction (bound fraction), therefore the first (bound) and second (partially unbound) fractions containing comparable amounts of antibody were compared for better quantification. Equal sample amounts and sample preparation and analysis conditions improve the quality of the label-free quantification method used in the present method. The unconsumed target (target only) was the smallest species and eluted at approximately 33 min. Because the target is smaller in size and mass compared to Antibody 1 (approximately 17 kDa versus approximately 150 kDa), the complex of Antibody 1 bound to the target showed a relatively small shift in elution time and partially coeluted with the SEC peak (Figure 2A).

[0087] The first (bound) and second (partially unbound) fractions separated by SEC were subjected to trypsin digestion and LC-MS peptide mapping to identify and quantify amino acid residues and modifications affecting binding. The statistical significance obtained from multiple LC-MS peptide mapping analyses and the fold change (FC) values ​​of the modifications in the unbound vs. bound antibody fractions are shown as a volcano plot (Figure 2B). The plot contained 188 dots for a total of 188 modifications identified in Antibody 1. Thirty-nine modifications were statistically different in the unbound vs. bound fractions with p values ​​less than 0.05. They were located above the horizontal lines in Figures 2B and 3. The remaining modifications were at similar levels, indicating that they did not affect binding or were caused by sample preparation. The criteria for unbound / bound FC were set as follows: Since the most negative Log2FC value was −0.83 at the bottom of FIG. 2B, the positive Log2FC threshold should be set at 0.83, and all modifications in between were considered not to affect binding. Modifications with log2FC>0.83 (FC>1.78) were considered to be associated with binding. This criterion, together with the requirement of a p-value <0.05, identified 15 modifications with statistically significant changes (FIG. 3). Grouping the modifications by residue, 7 residues in the variable regions of the heavy and light chains of Antibody 1 were identified that had a statistically significant effect on the binding between Antibody 1 and the target (FIG. 4). The majority of the identified residues are in CDR1 (around residue 30) and CDR2 (around residue 50) of the light chain and CDR2 (around residue 50) of the heavy chain (FIG. 2), which are in close proximity to the target (data not shown), consistent with the crystal structure (data not shown). The single residue that most affected antibody 1-target binding was cysteine ​​23 outside the CDR (C23, FIG. 4). This finding suggests that long-range (allosteric) effects may play a role caused by modification of C23, suggesting that reduction of the disulfide bond involving C23 and modification of this residue reduced the mass by 15.9941. Reduction of the disulfide bond may decrease the energy of unfolding (melting temperature), weakening the protein structure and resulting in loss of binding.Alternatively, a weaker structure without disulfide bonds could make the surrounding CDR residues more susceptible to modification, with subsequent modification disrupting binding. Either way, modification of C23 correlates with loss of binding, despite not being in close proximity to the target. Only very small (but statistically significant) levels of modification of C23 were observed in photostressed samples (<0.12%).

[0088] In an alternative embodiment, to make the attribution more reliable, a corrected p-value (p*) may be considered, which includes the total number of modifications tested as p*=p / N, where p is the p-value for the test of a single modification and N is the number of modifications tested (see

[0073] ). For example, if there are 188 modifications in the list identified and quantified by LC-MS peptide mapping, the corrected p-value (p*) should be set to 0.05 / 188=0.000266, -log10(0.000266)=3.6. In practice, the significance threshold (horizontal dashed line in Figure 2) is set at -log10(p)=1.3 to -log10(p*)=3.6. Also, there should be no dots (false positive results) located in the top left quadrant of the volcano plot bounded by the dashed line (Figure 2). The two vertical dashed lines for the fold change thresholds should be moved symmetrically outwards and the horizontal significance line should be moved upwards until the top left quadrant is clear. These thresholds define that the modification affects binding.

[0089] Even after the severe light stress of the P2 protocol, modification levels were relatively low, less than 5% at residues that affect binding ("Initial P2" in Figures 3, 4). Modifications in unstressed standards were less than 1% (data not shown). In general, the relevance of attributes is defined not only by their impact on binding, but also by whether the modifications survive at any significant level during normal manufacturing processes, storage, and / or human circulation. Other stress conditions for accelerated stability testing generally include: 1) high temperature stress in formulation buffer (e.g., 37°C for several weeks), and 2) increasing the pH from a weakly acidic formulation buffer to a physiological buffer (PBS, pH 7) and incubation at 37°C for several weeks.

[0090] Example 2 This example describes an exemplary method of the disclosure in which the therapeutic protein is an Fc fusion protein, antibody protein product 1 (APP1), that binds to antigen 2.

[0091] Materials and Methods APP1 is an Fc fusion protein that binds two different antigens, one of which is antigen 2. Antigen 2 is expressed by T cells. The Fc portion of APP1 is non-glycosylated. The molecular weight of APP1 is 106 kDa. APP1 was formulated at a concentration of 1.98 mg / ml in 10 mM glutamic acid and 9% sucrose, with a pH of 4.2. The target of APP1 used in this study was a conjugate containing two molecules of antigen 2 fused to human IgG1 Fc. The Fc portion of the target was glycosylated. The target was formulated at a concentration of 3 mg / ml (0.38 EU / mg. MW=61,058.5 Da) in 10 mM acetate (pH 5.2), 100 mM NaCl (containing 9% sucrose).

[0092] APP1 was stressed by changing the pH of the formulation from 4.2 to 7.0 by dilution with PBS and incubated for 2 weeks at 37° C. APP1 was stressed formulated in G42SuT at a concentration of 1.1 mg / mL in 0.5 mL vials. APP1 and target were mixed at a 1:1 molar ratio. The mixture was incubated in an HPLC vial at 37° C. for 10 minutes.

[0093] result To assess the relevance of the formed attributes to the interaction between APP1 and its target, an APP1 sample was mixed with its target and then fractions were separated by SEC. In SEC, proteins elute in order of size and molecular weight from larger to smaller molecules. The first eluting SEC peak (13-14 min) contained the bound fraction containing APP1 bound to the target (Figure 5A). The second eluting SEC peak (16 min) was the bound target species (not bound to APP1), which eluted before APP1 (with a MW of 106 kDa). This suggests that the bound target species forms a non-covalent dimer with a molecular weight of 122 kDa and a structure of 2× the target.

[0094] A number of modifications were identified and plotted in a statistical significance vs. fold change volcano plot (Figure 5B). pH stress followed by incubation at 37°C, pH 7 for 2 weeks induced the formation of several attributes, especially deamidation of asparagine residues. Criteria for non-binding / binding fold change were set as described in Example 1, and FC>2 (log2FC>1) and corrected p-value<0.05 / 100 (-log10p-value>3.3) were identified as statistically significant in reducing binding. Three modifications on different residues met the criteria and appeared as dots in the upper right quadrant (Figure 5B). Of the two target binding domains of APP1, only the antigen 2 binding domain was evaluated in this study. The three residues with the highest combined attributes of statistical significance and fold change were in the CDR3 region of the antigen 2 binding domain of APP1: (1) HC CDR3 N357, (2) HC CDR3 N360, and (3) LC CDR1 W431. The fact that these changes occurred in the antigen 2 binding domain validates the accuracy of this method. From their position along the fold change axis, these three residues should have the following binding strength: N357>W431>N360. The fold change values ​​FC N357=142 and FC N360=3 indicate that N357 plays a much larger role in binding. Some dots (attributes) with relatively high fold changes were below the significance threshold of -log10(p*)=3.3. Many of them were from the target binding and Fc regions of the proteins, suggesting that they were induced during SEC fraction collection and sample preparation. This is probably because the concentration of the protein in the unbound fraction was lower than that of the bound one, and the unbound protein was subjected to oxidative species and other impurities. Since these modifications vary from one sample preparation to another, the statistical significance was low.

[0095] Example 3 This example describes an exemplary method of the present disclosure that uses affinity chromatography-precipitation techniques, including the use of target-labeled beads or precipitation with cells that express the target on their surface.

[0096] In this example, tightly bound therapeutic protein-target complexes (K D =10 -11 -10 -9 Affinity separation (affinity enrichment) of the bound (M) and unbound modified therapeutic proteins is performed using the bound target immobilized on resin beads or the bound target expressed as a cell surface protein (Figure 6A). In the first group, the bound target protein is immobilized on Actigel ALD Superflow resin (Sterogene Bioseparations, Inc., Carlsbad, CA) according to the manufacturer's protocol. The activated resin contains a free carbonyl group that binds to the primary amine on the target protein. In the first subgroup, the beads are placed in an affinity chromatography column (Figure 6B). In the second subgroup, the beads are placed in a centrifuge tube with gentle rocking. In the second group, cells expressing the bound target as a cell surface protein are used instead of the bound target protein immobilized on the beads.

[0097] The therapeutic protein is stressed and then mixed with beads or cells covered with the binding target. In the first subgroup, an aliquot of the mixture is injected into an affinity chromatography column. In the second subgroup, the centrifuge tube is gently shaken overnight. In the second group, cells expressing the binding target on the cell surface are mixed with the stressed therapeutic protein and incubated in the centrifuge tube with gentle rocking. After mixing, the centrifuge tube is centrifuged to separate the beads or cells containing the bound fraction containing the therapeutic protein-binding target complex (located in the pellet) from the unbound fraction containing the therapeutic protein that is not bound to the binding target residue (located in the supernatant) (Figure 6A). In the first subgroup of the first group, the fractions are eluted from the column with a suitable elution buffer such that the bound fraction containing the therapeutic protein-binding target complex is retained on the column while the unbound fraction containing the therapeutic protein that is not bound to the binding target is eluted from the column. In the first group involving the resin, the conditions for these steps are similar to those described for the purification of the antibody of interest from blood, e.g., as described in Liu et al., J. Biol. Chem. 283:29266-29272 (2008) and / or Goetze et al., Glycobiology 21:949-959 (2011).

[0098] The separated bound and unbound fractions are subjected to LC-MS peptide mapping to identify amino acid residues and their attributes that affect binding, and results are obtained according to FIG. 1 and substantially as described in Example 1.

[0099] Antibodies, BITEs, proteins, and Fc fusion proteins can be evaluated by the method described in this example. This method is expected to fractionate a larger amount of bound and unbound therapeutic protein compared to the method of Example 1 using SEC. The method of this example is also advantageous when the extracellular soluble domain of the target is difficult to generate and the target is only produced on the cell surface.

[0100] Example 4 This example describes an exemplary method of the disclosure in which the affinity purification technique includes affinity chromatography, precipitation with binding target-labeled beads, or precipitation with a cell surface binding target, in this example, the therapeutic protein comprises the Fc of an antibody and the binding target is an Fc receptor.

[0101] In this example, weakly bound Fc-containing therapeutic protein-Fc receptor complexes (K D =10 -8 ~10 -6 Affinity separation (affinity enrichment) of modified Fc-containing proteins (Fc-M) and unbound proteins is performed using Fc receptors immobilized on resin beads or Fc receptors expressed by cells. In the first group, Fc receptors are immobilized on Actigel ALD Superflow resin (Sterogene Bioseparations, Inc., Carlsbad, CA) according to the manufacturer's protocol. The activated resin contains a free carbonyl group that binds to the primary amines on the Fc receptor. In the first subgroup, the Fc receptor-bound beads are placed in an affinity chromatography column (Figure 7B). In the second subgroup, the beads are placed in a centrifuge tube with gentle rocking (Figure 7A). In the second group, cells expressing Fc receptors on their cell surface are used instead of beads.

[0102] The Fc-containing therapeutic protein is stressed and subsequently mixed with beads or cells covered with Fc receptors. In a first subgroup, an aliquot of the mixture is injected into an affinity chromatography column. In a second subgroup, the centrifuge tube is gently shaken overnight. In a second group, cells expressing Fc receptors are mixed with the stressed Fc-containing therapeutic protein and incubated in the centrifuge tube with gentle rocking. After mixing, the centrifuge tube is centrifuged to separate the beads or cells containing the bound fraction comprising the Fc-containing therapeutic protein-Fc receptor complex (located in the pellet) from the unbound fraction comprising the Fc-containing therapeutic protein not bound to Fc receptor residues (located in the supernatant). In a first subgroup of the first group, the fractions are eluted from the column with a suitable elution buffer such that the bound fraction comprising the Fc-containing therapeutic protein-Fc receptor complex is retained on the column, while the unbound fraction comprising the Fc-containing therapeutic protein not bound to the Fc receptor is eluted from the column. In the first group involving the resin, the conditions for these steps are similar to those described for the purification of the antibody of interest from blood, e.g., as described in Liu et al., J. Biol. Chem. 283:29266-29272 (2008) and / or Goetze et al., Glycobiology 21:949-959 (2011).

[0103] The separated bound and unbound fractions are subjected to LC-MS peptide mapping to identify amino acid residues and their modifications that affect binding, and the results are obtained according to FIG. 1 and substantially as described in Example 1.

[0104] Example 5 This example describes an exemplary method for determining the importance of an attribute of a Therapeutic protein.

[0105] Fc-containing proteins are approximately 10 -8 ~about 10 -6 K of M DIt is known that Fc receptors bind to Fc receptors with an affinity characterized by an affinity of 1. In this experiment, the procedure described in Example 3 is carried out, except that a library of therapeutic fusion protein species containing Fc regions with known attributes is used. Each member of the library is the same therapeutic fusion protein, but with a unique attribute (relative to other members of the library). Affinity purification is performed using beads coupled to the Fc receptor. In some examples, the beads are coated with protein A, which binds to the Fc receptor. Fractions containing unbound Fc fusions (found in the supernatant) are analyzed by LC-MS peptide mapping alongside fractions containing bound fusion protein-binding target complexes. For each species in each fraction, the abundance of the species in the fraction containing unbound Fc fusions is compared to the abundance of the species in the fraction containing bound fusion protein-target complexes. For species with an abundance ratio greater than 1, the species is considered to contain an attribute associated with the binding interaction between the therapeutic protein and the target.

[0106] Example 6 This example demonstrates the identification of key chemical modifications and paratope mapping by size-exclusion chromatography (SEC) of stressed antibody-target complexes.

[0107] Introduction Therapeutic proteins, including antibodies and Fc fusion proteins, undergo numerous chemical modifications during production and storage, including deamidation, oxidation, isomerization, hydroxylation, glycation, glycosylation, clips, etc. They are also exposed to harsh conditions during stress testing, including forced oxidation, extreme pH, UV light to study possible degradation pathways and the suitability of methods to detect them. Some of these modifications are located at residues in the binding region, leading to loss of binding / potency and are classified as critical quality attributes. Traditionally, the importance of modifications can be assessed by a laborious process of fractionating and characterizing intact protein variants one by one. We now describe a method that allows for automated, massively parallel assessment of the importance of all possible modifications and for mapping the paratope-epitope of the antibody-target binding site.

[0108] method Therapeutic antibody samples were treated for several hours with UV light designed to generate multiple chemical modifications in the antibody population. The stressed antibody was then incubated with its target protein and the mixture was fractionated by SEC on the earlier eluting antibody-target complexes followed by elution of unbound antibody. Fractions were collected and antibody modifications in each fraction were identified and quantified by LC-MS / MS peptide mapping using an Agilent 1290 HPLC and Thermo QE-Biopharma MS. For statistical significance, this procedure was repeated several times (n=3). To distinguish important modifications affecting binding from non-important modifications and artifactual modifications caused by sample preparation, volcano plots of statistical significance versus the ratio of modifications in unbound / bound antibody species (fold change) were used (Figures 8A-8D).

[0109] Preliminary data Typically powerful K D =10 -11 ~10 -9 In the therapeutic antibody-target binding affinity range of M, the antibody and target elute together from the SEC as a complex. Chemical modifications of the stressed antibody increase the affinity by K D =10 -8When the binding affinity is weaker than M, the unbound antibody elutes away from its target. In this study, 275 chemical modifications were identified and quantified by peptide mapping of bound and unbound antibody molecules, which are shown as volcano plots (Figure 9). 27 modifications were shown with high statistical significance (p-value < 0.006) and fold change (> 1.7) of unbound / bound antibody species. These modifications occurred in 11 residues identified as the paratope of the antibody (Figures 10A-13I). The paratope map was plotted as modification score versus residue number (Figures 10A-10B). The modification score was chosen as the sum of the logarithm of the fold change and the logarithm of the p-value (Log2 fold change + -log10 p-value). This is because the scales have similar values ​​and both are important to separate the signal from the noise. Ten of the 11 residues with the highest scores were located in the complementarity determining regions (CDRs), which are regions that are usually involved in binding. Residues in conserved regions that are not typically involved in binding had significantly lower scores. Correlation of the antibody-target complex with crystallography data revealed that 8 of 11 important antibody residues were within 5 Å of the target, indicating that these antibody residues are close to and may interact with the target. These results supported the validity of the method. The only significant modification detected outside the CDRs was on a cysteine ​​residue (C23) close to CDR1, suggesting that long-range (allosteric) effects caused by reduced disulfide bonds may play a role. This method utilizes modifications on amino acid side chains to map protein-protein interactions, similar to fast photochemical oxidation of proteins (FPOP), but does not require fast photochemical oxidation using lasers. This method is particularly important in situations where other structural data (from crystallography or NMR) are not available during early product development of therapeutic antibodies.

[0110] Thus, a method is provided for SEC of stressed antibody-target complexes followed by identification of key chemical modifications and paratope mapping by peptide mapping of bound and unbound antibody species.

[0111] Example 7 This example demonstrates competitive SEC affinity separation to identify antibody modifications that affect binding to a target protein.

[0112] Introduction Typically, for therapeutic proteins (e.g., antibodies, BiTE® molecules, therapeutic proteins, Fc fusion proteins, etc.), a number of attributes are identified, including glycosylation, hydroxylation, glycation, deamidation, oxidation, isomerization, clips, and other chemical modifications. The attributes include post-translational modifications of therapeutic proteins that occur during cell culture as well as chemical modifications / degradation that occur during production, purification, formulation, and storage. Some of these modifications occur in the binding region and may affect the binding to the target and efficacy. The efficacy of the therapeutic protein is one of the main criteria for the assessment of the importance of the attributes. This method describes the experimental assessment of the importance of the attributes by competitive affinity binding between an antibody and its target Her2, followed by characterization of the bound and unbound therapeutic proteins by SEC fractionation and LC-MS / MS peptide mapping (Figure 14).

[0113] method The soluble part of human recombinant Her2 receptor was used for binding to the stressed antibody. A Tosoh TSK Gel G3000SWXL column was used for SEC separation of bound antibody-antigen complexes and unbound antibody. The stressed antibody (45C10d)-receptor samples were also analyzed using a SEC-UV-MALS system to determine the binding stoichiometry (Figures 16A-16C). Corresponding fractions from each SEC run were collected, digested with trypsin, and then analyzed using LC-MS / MS. Mass spectrometry data were searched for mAb heavy and light chains, receptor, and trypsin using MassAnalyzer. A volcano plot, a type of scatter plot, was used to represent peptide mapping results and identify changes in large data sets consisting of replicate data.

[0114] Preliminary data In this method, characterization involves the identification and quantification of modifications in the bound and unbound therapeutic protein fractions with the goal of identifying key quality attributes responsible for loss of binding. Equilibrium dissociation constant Kd<10 -8 In the case of M, the protein-ligand complex containing the antibody-target remains bound during SEC separation. This method was applied to several antibody:target complexes by mixing the antibody and target in a ratio defined by the stoichiometry of the interaction, typically 1:2. However, in the case of the antibody-Her2 complex, the post-stress modifications did not lead to a loss of binding, and even the stressed and degraded antibody molecule retained a Kd<10. -8 It was suggested that M. When less acceptor was provided (1:1), antibody species with modifications that decrease binding elute as unbound from SEC (Figures 15A-15G). Fractions were collected for two experiments (Figure 17): 1) CEX and 2) reduced peptide mapping. For peptide mapping, protein fractions were collected, guanidine-rich denatured, reduced, alkylated, and digested.

[0115] CEX data showed that the bound fraction contained the major species, while the unbound fraction contained a high proportion of acidic and basic variants (Figures 18A-18D). After displaying triplicate measurements (n=3) as volcano plots, peptide mapping revealed the proportions of modifications and their p-values ​​(Figures 19, 20A-20E). The top right region should contain the modifications that most affect binding. Several modifications with ratios above 1.5 were summarized in an abundance plot (Figure 21). HC D102 isomerization and LC N30 deamidation were 4-fold higher in the unbound fraction with high statistical significance. HC N55 deamidation and M107 oxidation were also present, but they were not statistically different between unbound and bound.

[0116] Our findings are consistent with previously published potency measurements (Figures 22A-22B). Some modifications at the N-terminus of the HC and LC were associated with non-binding. A thorough review of the antibody-Her2 crystal structure by Cho et al. 2003 showed that the N-terminus of the HC and LC are closer to the target, and thus modifications at the termini can indeed result in loss of binding (Figures 22A-22B).

[0117] Thus, it can be seen that competitive SEC of stressed antibody / target mixtures, followed by peptide mapping, revealed residues and modifications that reduced but did not eliminate binding.

[0118] References The following references are cited in Example 6: Abbott WM, Damschroder MM, Lowe DC.2014.Current approaches to fine mapping of antigen-antibody interactions.Immunology 142:526-535. Blech M,Peter D,Fischer P,Bauer MM,Hafner M,Zeeb M,Nar H.2013.One target-two different binding modes:structural insights into gevokizumab and canakinumab interactions to interleukin-1beta.J Mol.Biol.425:94-111. Dhungana S,Fessler MB,Tomer KB.2009.Epitope mapping by differential chemical modification of antigens.Methods Mol.Biol.524:119-134. Englander SW.2006.Hydrogen exchange and mass spectrometry:A historical perspective.J Vasc Surg 17;1481(-1489):Goswami D,Zhang J,Bondarenko PV,Zhang Z.2018.MS-based conformation analysis of recombinant proteins in design,optimization and development of biopharmaceuticals,https: / / doi.org / 10.1016 / j.ymeth.2018.04.011.Methods 144:134-151. Jones LM,Sperry B,Carroll A,Gross ML.2011.Fast photochemical oxidation of proteins for epitope mapping.Anal Chem 83:7657-7661. Luo J,Liu Z,Guo Y,Li M.2015.A structural dissection of large protein-protein crystal packing contacts.Sci.Rep.5:14214. Opuni KFM,Al-Majdoub M,Yefremova Y,El-Kased RF,Koy C,Glocker MO.2018.Mass spectrometric epitope mapping.Mass Spectrom Rev.37:229-241. Sevy AM,Healey JF,Deng W,Spiegel PC,Meeks SL,Li R.2013.Epitope mapping of inhibitory antibodies targeting the C2 domain of coagulation factor VIII by hydrogen-deuterium exchange mass spectrometry.J Thromb.Haemost.11:2128-2136.

[0119] All references cited in this specification (e.g., publications, patent applications, and patents) are herein incorporated by reference to the same extent as if each reference was individually and specifically indicated to be incorporated by reference and was set forth in its entirety herein.

[0120] In the context of describing this disclosure (and particularly in the context of the claims which follow), use of the terms "a," "an," and "the" and similar referents should be construed to encompass both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. The terms "including," "having," "including," and "containing" should be construed as open-ended terms (i.e., meaning "including, but not limited to"), unless otherwise specified.

[0121] The recitation of ranges of values ​​herein, unless otherwise indicated herein, is merely intended to serve as a shorthand method of referring individually to each of the separate values ​​falling within the range, and each of the endpoints, and each separate value and endpoint is incorporated herein as if it were individually recited herein.

[0122] All of the methods described herein may be performed in any suitable order unless otherwise indicated herein or clearly contradicted by context. The use of any examples or exemplary language (e.g., "etc.") provided herein is intended merely to further clarify the disclosure and does not impose limitations on the scope of the disclosure unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the disclosure.

[0123] Preferred embodiments of the present disclosure are described herein, including the best mode known to the inventors for carrying out the present disclosure. Variations of those preferred embodiments will become apparent to those skilled in the art upon reading the above description. The inventors expect that those skilled in the art will adopt such variations as necessary, and the inventors intend for the present disclosure to be practiced in other forms than those specifically described herein. Accordingly, the present disclosure includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements is encompassed in the present disclosure in all possible variations thereof unless otherwise indicated herein or clearly contradicted by context.

Claims

1. 1. A method for identifying structural features of a therapeutic protein that affect an interaction between the therapeutic protein and a target, wherein the therapeutic protein is an antigen binding protein and the target is an antigen, the method comprising: a. applying a stress to a first sample containing a therapeutic protein; b. contacting the first sample with a second sample containing the target to form a mixture containing (i) a therapeutic protein-target complex, (ii) unbound therapeutic protein, and (iii) unbound target; c. separating the mixture into (i) an unbound fraction comprising unbound therapeutic protein and target, and (ii) a bound fraction comprising therapeutic protein-target complexes; d. digesting proteins in the unbound and bound fractions; e. identifying and quantifying the abundance of structures present in the therapeutic protein by liquid chromatography / mass spectrometry (LC-MS) peptide mapping for each of the unbound and bound fractions; f. determining that if the abundance of a structure in the unbound fraction is greater than the abundance of the structure in the bound fraction, the structure negatively affects the interaction between the therapeutic protein and the target; The method includes:

2. 2. The method of claim 1, wherein the structure is determined to be a candidate for becoming part of the paratope of the Therapeutic protein if the structure has a negative effect on the interaction between the Therapeutic protein and the target.

3. 1. A method for determining the effect of a known structure present in a therapeutic protein on an interaction between said therapeutic protein and a target, said therapeutic protein being an antigen binding protein and said target being an antigen, said method comprising: a. contacting a first sample containing a therapeutic protein having a known structure with a second sample containing a target to form a mixture containing (i) the therapeutic protein-target complex, (ii) unbound therapeutic protein, and (iii) unbound target; b. separating the mixture into (i) an unbound fraction comprising unbound therapeutic protein and target, and (ii) a bound fraction comprising therapeutic protein-target complexes; c. digesting proteins in the unbound and bound fractions; d. Quantifying the abundance of the known structure in the therapeutic protein by LC-MS peptide mapping for each of the unbound and bound fractions; e. determining that the known structure has a negative effect on the interaction between the Therapeutic protein and the target if the abundance of the known structure in the unbound fraction is greater than the abundance of the known structure in the bound fraction; The method includes:

4. 4. The method of claim 3, wherein if the known structure has a negative effect on the interaction between the Therapeutic protein and the target, the structure is determined to be a candidate for being included in the paratope of the Therapeutic protein.

5. 4. The method of claim 1, wherein the first sample comprises one or more species of the therapeutic protein, each species comprising a different structural profile.

6. The method of claim 5 , wherein the first sample further comprises serum or a serum fraction.

7. 4. The method of claim 1 or 3, wherein the mixture is separated into fractions using a technique selected from size exclusion chromatography (SEC), affinity chromatography, precipitation using beads or cells, free-flow fractionation (FFF), ion exchange chromatography (IEX), hydrophobic interaction chromatography (HIC) and ultracentrifugation (UC).

8. 4. The method of claim 1 or 3, comprising determining, for each structure, a ratio of the abundance of the structure in the unbound fraction to the abundance of the structure in the bound fraction, and determining that if the ratio is greater than 1, the structure negatively affects the interaction between the therapeutic protein and the target.

9. 9. The method of claim 8, comprising repeating the steps of separating, digesting, identifying and quantifying, wherein each time the steps are repeated, a ratio is determined for each structure to obtain a plurality of ratios of the abundance of the structure in the unbound fraction to the abundance of the structure in the bound fraction, and the method further comprises determining the statistical significance of the plurality of ratios based on calculating a p-value.

10. The method of claim 1 or 3, wherein the structure is formed as a result of chemical modification.

11. 11. The method of claim 10, wherein the chemical modification alters the mass-to-charge ratio (m / z) of a charged ion of an amino acid of the therapeutic protein.

12. The method of claim 10, wherein the chemical modification is glycosylation, hydroxylation, glycation, deamidation, oxidation, reduction, isomerization, aggregation, degradation, acetylation or hydrolytic proteolytic clipping.

13. 2. The method of claim 1, wherein the stress is exposure to ultraviolet light, heat, air, freeze / thaw cycles, shaking, long-term storage, a change in pH or a change in temperature.

14. 14. The method of claim 13, wherein the change in pH is greater than 1.0 or greater than 2.

0.

15. 14. The method of claim 13, wherein the change in temperature is 2 degrees Celsius or more, or 5 degrees Celsius or more.

16. 4. The method of claim 1 or 3, wherein the antigen binding protein comprises an antibody, an antigen-binding antibody fragment, a bispecific T cell engager molecule, a bispecific antibody, a trispecific antibody, or an Fc fusion protein.

Images