WO2021147642A1

WO2021147642A1 - Methods, models and systems related to antibody immunogenicity

Info

Publication number: WO2021147642A1
Application number: PCT/CN2020/141999
Authority: WO
Inventors: Shide LIANG
Original assignee: Bio-Thera Solutions, Ltd.
Priority date: 2020-01-20
Filing date: 2020-12-31
Publication date: 2021-07-29

Abstract

Disclosed herein are methods of selecting a candidate therapeutic antibody with low immunogenicity, methods of producing an antibody with low immunogenicity, methods of reducing immunogenicity of an antibody, methods of identifying antibody immunogenicity, methods of predicting antibody immunogenicity, systems for identifying or predicting antibody immunogenicity, and antibodies designed or produced using the methods or systems.

Description

METHODS, MODELS AND SYSTEMS RELATED TO ANTIBODY IMMUNOGENICITY

FIELD

The present disclosure belongs to the field of biopharmaceutical and relates to methods of selecting a candidate therapeutic antibody with low immunogenicity, methods of producing an antibody with low immunogenicity, methods of reducing immunogenicity of an antibody, methods of identifying an antibody with low immunogenicity, methods of predicting antibody immunogenicity, systems for identifying or predicting antibody immunogenicity, and antibodies designed or produced using the methods or systems.

BACKGROUND

Since the first recombinant therapeutic protein, human insulin, was approved in 1982, more than 250 products have entered the marketplace with an estimated annual revenue of over 150 billion dollars. Therapeutic proteins including monoclonal antibodies, coagulation factors, replacement enzymes, fusion proteins, hormones, growth factors, and plasma proteins are now a fast growing segment of the pharmaceutical industry. These approved therapeutic proteins are indicated for a wide variety of areas such as cancers, autoimmunity/inflammation, exposure to infectious agents, and genetic disorders. The rapid advances in biomedical science and technology make it possible to address the unmet needs with new therapeutic proteins.

In comparison with small molecules that bind in a deep pocket, biopharmaceuticals can bind the flat surface of a protein with high specificity to interfere in vivo processes and restore previously untreatable conditions. However, when therapeutic proteins are administrated to patients, unwanted immune responses, such as generation of anti-drug antibody (ADA) , can cause a wide range of problems including altered pharmacokinetics, loss of efficacy, and even life-threatening complications. The immunogenicity against therapeutic proteins can be generated in both T cell dependent and T cell independent pathways. Antibodies generated from T cell dependent pathway have a higher affinity than those generated from T cell independent activation and appear to play a critical role in the development of antibody responses to biologic therapeutics. More specifically, T cells are activated by the recognition of linear antigenic peptides derived from the therapeutic proteins, called T cell epitope. The activated T cells then stimulate B cells to generate ADAs against the therapeutic protein.

So far computational prediction of T cell epitopes achieved significant progress. Numerous prediction algorithms have been developed and an AUC (area under curve) value of 0.786 was obtained for a large test set by consensus approach. Unlike molecules recognizing T cell epitopes, antibodies bind a conformational epitope on the protein surface, called B cell epitope. The existing tools predict the sequence-discontinuous B cell epitope based on the physiochemical properties of a protein structure and the performance is far from ideal with an accuracy slightly better than random. More frequently, T cell epitopes were predicted and deleted for biotherapeutic deimmunization, partly due to the difficulty of direct prediction of B cell epitopes.

Monoclonal antibodies contributed almost half of therapeutic proteins approved by the U.S. Food and Drug Administration (FDA) in the past several years. Immunogenicity is an important concern for therapeutic antibodies during drug development and regulation. For example, bococizumab, a humanized monoclonal antibody being developed to reduce the levels of low-density lipoprotein cholesterol, was recently discontinued after phase III clinical trials on 4300 patients citing decreased treatment efficacy due to high immunogenicity incidence rates.

There is a need to predict and reduce antibody immunogenicity.

SUMMARY

Disclosed herein include embodiments of a method for selecting a candidate therapeutic antibody with low immunogenicity. In some embodiments, the method comprises selecting an antibody having a large cavity volume at the complementarity-determining regions (CDR) of the antibody, having a significantly hydrophobic surface area at the heavy-chain complementarity-determining region 3 (CDRH3) loop of the antibody, or containing one or more glycine residues at the β turn of heavy-chain complementarity-determining region 2 (CDRH2) loop of the antibody.

In some embodiments, the candidate therapeutic antibody with low immunogenicity comprises two or more features selected from the list: having one or more large cavities at the CDR regions of the antibody, having a significantly hydrophobic surface area at the CDRH3 loop of the antibody, or containing one or more glycine residues at the β turn of CDRH2 loop of the antibody. In some embodiments, the candidate therapeutic antibody with low immunogenicity has one or more large cavities at the CDR regions, has a significantly hydrophobic CDRH3 loop, and contains one or more glycine residues at the β turn of CDRH2 loop.

In some embodiments, the sum of cavity volumes at the CDR regions is about

or more. In some embodiments, the sum of cavity volumes at the CDR regions is about

to about

For example, the sum of cavity volumes at the CDR regions is about

or a number or a range between any two of these values. In some embodiments, the sum of cavity volumes at the CDR regions is about

or more.

In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%or more. In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%to about 80%. For example, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 63%, 64%, 65%, 67%, 68%, 69%, 70%, 74%, 80%, or a number or a range between any two of these values. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 67%or more. In some embodiments, the CDRH3 hydrophobic surface area is about

or a number or a range between any two of these values.

In some embodiments, the β turn of CDRH2 loop (heavy chain (VH) 52-56) has at least one glycine residue. For example, the β turn of CDRH2 loop (VH 52-56) has one glycine residue, two glycine residues, three glycine residues, or four glycine residues. In some embodiments, the β turn of CDRH2 loop (VH 52-56) has at least two glycine residues.

In a further aspect, provided is a method of reducing immunogenicity of an antibody. In some embodiments, the method comprises: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, or introducing one or more glycine residues at the β turn of CDRH2 loop of the antibody.

In some embodiments, the method comprises two or more of the following: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, or introducing one or more glycine residues at the β turn of CDRH2 loop of the antibody. In some embodiments, the method comprises: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, and introducing one or more glycine residues at the β turn of CDRH2 loop of the antibody.

In some embodiments, the sum of cavity volumes at the CDR regions is increased to about

or more. In some embodiments, the sum of cavity volumes at the CDR regions is increased to about

to about

For example, the sum of cavity volumes at the CDR regions is increased about

or a number or a range between any two of these values. In some embodiments, the sum of cavity volumes at the CDR regions is increased to about

or more.

In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is increased to about 63%or more. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is increased to about 63%to about 80%. For example, the ratio of hydrophobic surface to total surface for the CDRH3 loop is increased to about 63%, 64%, 65%, 67%, 68%, 69%, 70%, 74%, 80%, or a number or a range between any two of these values. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is increased to about 67%or more. In some embodiments, the CDRH3 hydrophobic surface area is increased to about

or a number or a range between any two of these values.

In some embodiments, the β turn of CDRH2 loop (VH 52-56) has at least one glycine residue. For example, the β turn of CDRH2 loop (VH 52-56) has one glycine residue, two glycine residues, three glycine residues, or four glycine residues. In some embodiments, the β turn of CDRH2 loop (VH 52-56) has at least two glycine residues.

In a further aspect, provided is a method of producing an antibody with low immunogenicity. In some embodiments, the method comprises: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, or adding one or more glycine residues at the β turn of CDRH2 loop of the antibody.

In some embodiments, the method comprises two or more of the following: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, or adding one or more glycine residues at the β turn of CDRH2 loop of the antibody. In some embodiments, the method comprises: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, and adding one or more glycine residues at the β turn of CDRH2 loop of the antibody.

to about

For example, the sum of cavity volumes at the CDR regions is increased to about

or more.

or a number or a range between any two of these values.

In a further aspect, provided is an antibody produced by the method described herein. In some embodiments, the antibody comprises two or more features from the list: having one or more large cavities at the CDR regions, having a significantly hydrophobic CDRH3 loop, and containing one or more glycine residues at the β turn of CDRH2 loop. In some embodiments, the antibody is antibody that has not been previously made.

In a further aspect, provided is a method of producing an antibody with low immunogenicity, comprising producing an antibody having one or more large cavities at the CDR regions, a significantly hydrophobic surface area at the CDRH3 loop, and/or one or more glycine residues at the β turn of CDRH2 loop of the antibody. In some embodiments, the antibody is antibody that has not been previously made. In some embodiments, the method of producing an antibody with low immunogenicity comprises designing an antibody having one or more large cavities at the CDR regions, a significantly hydrophobic surface area at the CDRH3 loop, and/or one or more glycine residues at the β turn of CDRH2 loop of the antibody, and producing the antibody. In some embodiments, designing the antibody can be done by the methods or systems described herein. In some embodiments, the antibody has one or more large cavities at the CDR regions, has a significantly hydrophobic CDRH3 loop, and contains one or more glycine residues at the β turn of CDRH2 loop.

In some embodiments, the sum of cavity volumes at the CDR regions is about

to

For example, the sum of cavity volumes at the CDR regions is about

or more.

In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%or more. In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%to 80%. For example, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 63%, 64%, 65%, 67%, 68%, 69%, 70%, 74%, 80%, or a number or a range between any two of these values. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 67%or more. In some embodiments, the CDRH3 hydrophobic surface area is about

or a number or a range between any two of these values.

In a further aspect, provided is pharmaceutical compositions comprising the antibody, optionally together with a suitable carrier, excipient or diluent.

In a further aspect, provided is a method of identifying an antibody with low immunogenicity. In some embodiments, the method comprises analyzing two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified as having low immunogenicity or likely to have low immunogenicity.

In a further aspect, provided is a method of predicting antibody immunogenicity. In some embodiments, the method comprises analyzing two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the method comprises analyzing two features: cavity volume at the CDR regions of the antibody and hydrophobicity of CDRH3 loop of the antibody; wherein if the antibody has one or more large cavities at the CDR regions and a significantly hydrophobic CDRH3 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the method comprises analyzing two features: hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has a significantly hydrophobic CDRH3 loop and contains one or more glycine residues at the βturn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the method comprises analyzing three features: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop and contains one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the method described herein further comprises using a model which is able to analyze one or more of the following features: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody.

In some embodiments, the sum of cavity volumes at the CDR regions is calculated by the FPOCKET program. In some embodiments, the sum of cavity volumes at CDR regions is about

or more. In some embodiments, the sum of cavity volumes at CDR regions is about

to

For example, the sum of cavity volumes at CDR regions is about

or a number or a range between any two of these values. In some embodiments, the sum of cavity volumes at CDR regions is about

or more. In some embodiments, the sum of cavity volumes at CDR regions is about

or more.

In some embodiments, the hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area. In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%or more. In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%to 80%. For example, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 63%, 64%, 65%, 67%, 68%, 69%, 70%, 74%, 80%, or a number or a range between any two of these values. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 67%or more. In some embodiments, the CDRH3 hydrophobic surface area is about

or a number or a range between any two of these values.

In some embodiments, the model is called SVM model, which uses Support Vector Machine learning technology; features of antibodies with known immunogenicity are input into SVM model; parameters of SVM model are optimized in leave-one-out experiment; the trained SVM model is used for predicting immunogenicity.

In some embodiments, the model analyzes two features: the cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop. In some embodiments, the model analyzes two features: the hydrophobicity of CDRH3 loop and the information of the presence/absence of one or more glycine residues at CDRH2 β turn. In some embodiments, the model analyzes three features: the cavity volume at the CDR regions, the hydrophobicity of CDRH3 loop and the information of the presence/absence of one or more glycine residues at CDRH2 β turn.

In some embodiments, the crystal structural of the antibody is available, which is used to calculate the cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop.

In some embodiments, the crystal structure of the antibody is unavailable, the ABodyBuilder program is used to predict the antibody structure, and then the cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop are calculated.

In a further aspect, provided is a system for identifying or predicting antibody immunogenicity. In some embodiments, the system comprises a model described herein.

In some embodiments, the system comprises:

an input device for inputting features of an antibody;

and a module for calculating the sum of cavity volumes at the CDR regions of the antibody, the hydrophobicity of the CDRH3 loop of the antibody, or providing information of the presence/absence of glycine residues at CDRH2 β turn of the antibody, and analyzing the immunogenicity of the antibody; and

an output device for outputting immunogenicity of the antibody.

In some embodiment, the features of the antibody are calculated with the crystal structure of the antibody. In some embodiment, the features of the antibody are calculated with the predicted structure of the antibody. In some embodiment, the structure of the antibody is predicted by the ABodyBuilder program.

In some embodiments, the system comprises:

an input device for inputting features of antibodies with known immunogenicity and a test antibody;

at least one processor comprising a prediction module which predicts the structure of the test antibody and an analysis module which constructs a SVM model for analyzing immunogenicity using Support Vector Machine learning technology; wherein features described herein of antibodies with known immunogenicity are calculated and input into SVM model; parameters of SVM model are optimized in leave-one-out experiment; and the trained SVM model is used for identifying or predicting immunogenicity of the test antibody; and

an output device for outputting immunogenicity of the test antibody.

In some embodiments, the analysis module analyzes two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the analysis module analyzes two features: cavity volume at the CDR regions of the antibody and hydrophobicity of CDRH3 loop of the antibody; wherein if the antibody has a one or more large cavity at the CDR regions and a significantly hydrophobic CDRH3 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the analysis module analyzes two features: hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has a significantly hydrophobic CDRH3 loop and contains one or more glycine residues at the βturn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the analysis module analyzes three features: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop and contains one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the cavity volume at the CDR regions is calculated by the FPOCKET program. In some embodiments, the sum of cavity volumes at the CDR regions is about

to about

For example, the cavity volume at the CDR regions is about

or more.

In some embodiments, the hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 63%or more. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 63%to about 80%. For example, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 63%, 64%, 65%, 67%, 68%, 69%, 70%, 74%, 80%, or a number or a range between any two of these values. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is about 67%or more. In some embodiments, the CDRH3 hydrophobic surface area is about

or a number or a range between any two of these values.

In some embodiments, the crystal structure of the antibody is unavailable, the ABodyBuilder program is used to predict the antibody structure. The cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop are calculated with the modeled structure.

In some embodiments, the analysis module analyzes two features: the cavity volume at the CDR regions and hydrophobicity of CDRH3 loop. In some embodiments, the analysis module analyzes two features: the hydrophobicity of CDRH3 loop and the information of the presence/absence of glycine residues at CDRH2 βturn. In some embodiments, the analysis module analyzes three features: the cavity volume at the CDR regions, the hydrophobicity of CDRH3 loop and the information of the presence/absence of glycine residues at CDRH2 βturn.

The algorithm developed in the present disclosure could provide extra benefit to select candidate antibodies with low immunogenicity for clinical trials, especially, when the crystal structures are available. A humanized antibody, which is predicted as low immunogenicity by computational tools, could be a better choice for clinical development than the full human antibody generated from the expensive platform of transgenic mice.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1. Distribution pattern of calculated humanness scores for CDR regions; Fig 1a, Humanized antibodies; Fig 1b, Full human antibodies.

Figure 2. Crystal structures of idiotype-anti-idiotype Fv complex. Fig 2a, 1dvf; Fig 2b, 1iai; Fig 2c, 3bqu; Fig 2d, 1pg7; Fig 2e, 5jo4; Fig 2f, 5xaj. Grey, idiotypic antibody; orange, CDRH2 of idiotypic antibody; red, CDRH3 of idiotypic antibody; light blue, light chain of anti-idiotypic antibody; deep blue, heavy chain of anti-idiotypic antibody. The PDB code is shown for each complex structure.

Figure 3. Comparison of backbone structures for immunogenic and non-immunogenic therapeutic antibodies. Fig 3a, 15 immunogenic antibodies; Fig 3b, 14 non-immunogenic antibodies. The crystal structures of all the antibodies were obtained from PDB and superimposed to that of omalizumab (PDB code 4x7s) .

DETAILED DESCRIPTION

Definitions

As used herein, the following definitions shall apply unless otherwise indicated.

As used herein, unless otherwise stated, the singular forms “a, ” “an, ” and “the” include plural reference. Thus, for example, a reference to “an antibody” includes a plurality of antibodies.

As used herein, “about” will be understood by persons of ordinary skill in the art and will vary to some extent depending upon the context in which it is used. In some embodiments, “about” will mean up to plus or minus 10%or plus or minus 5%, or plus or minus 1%of the particular term. “About x” includes “x” .

As used herein, the term “comprising” is intended to mean that the compositions and method include the recited elements, but not excluding others. Embodiments defined by each of these transition features are within the scope of this disclosure.

CDRH2 is the second complementarity determining regions of the antibody heavy chains, CDRH3 is the third complementarity determining regions of the antibody heavy chains. VH 52-56 are from amino acid residues from the 52th to 56th of the heavy chain variable regions. Kabat numbering scheme was used in this study as indicated in example.

All publications are hereby incorporated by reference in their entirety for all purposes.

Low immunogenicity (which includes non-immunogenicity) of antibodies means that treatment-emergent anti-antibody response (AAR) occurs (e.g., is reported) or is predicted to occur in less than 2%of patients. Otherwise, high immunogenicity of antibodies means that treatment-emergent anti-antibody response (AAR) occurs (e.g., is reported) or is predicted to occur in at least 2%of patients. Detectable AAR means that AAR is detected at least once during treatment if patients in studies are tested at multiple time points. “Likely to have low immunogenicity” or a similar phrase means that an antibody has a more than 50%chance of having low or no immunogenicity. In some embodiment, the antibody has an at least 60%, at least 65%, at least 70%or at least 75%chance of having low or no immunogenicity.

A pocket located at the CDR regions means that at least of two third of the surrounding atoms of the pocket are the CDR residues.

The hydrophobicity of CDRH3 is defined as the sum of solvent accessible surface of all hydrophobic atoms thereof. The carbon, sulfur, and nitrogen atoms (excluding the non-protonated nitrogen atom of histidine side chain) are considered hydrophobic atoms.

Disclosed herein in one aspect, a method for selecting a candidate therapeutic antibody with low immunogenicity. In some embodiments, the method comprises selecting an antibody having one or more large cavities at the CDR regions of the antibody, having a significantly hydrophobic surface area at the CDRH3 loop of the antibody, or containing one or more glycine residues at the β turn of CDRH2 loop of the antibody.

In some embodiments, the candidate therapeutic antibody with low immunogenicity has one or more large cavities at the CDR regions, has a significantly hydrophobic CDRH3 loop, and contains one or more glycine residues at the β turn of CDRH2 loop.

In some embodiments, the sum of cavity volumes at CDR regions is about

or more. In some embodiments, the sum of cavity volumes at CDR regions is about

to about

In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%or more. In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%to about 80%. In some embodiments, the β turn of CDRH2 loop (VH 52-56) has at least one glycine residue.

In some embodiments, the sum of cavity volumes at CDR regions is about

or a number or a range between any two of these values. In some embodiments, the CDRH3 hydrophobic surface area is about

or a number or a range between any two of these values. In some embodiments, the β turn of CDRH2 loop (VH 52-56) has one glycine residue, two glycine residues, three glycine residues, or four glycine residues.

In some embodiments, the method comprises: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, and introducing Gly at the β turn of CDRH2 loop of the antibody.

In some embodiments, the sum of cavity volumes at CDR regions is increased to about

or more. In some embodiments, the sum of cavity volumes at CDR regions is increased to about

to about

In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is increased to about 63%or more. In some embodiments, the ratio of hydrophobic surface to total surface for the CDRH3 loop is increased to about 63%to about 80%. In some embodiments, the β turn of CDRH2 loop (VH 52-56) has at least one glycine residue.

or a number or a range between any two of these values. In some embodiments, the CDRH3 hydrophobic surface area is increased to about

or a number or a range between any two of these values. In some embodiments, the βturn of CDRH2 loop (VH 52-56) has one glycine residue, two glycine residues, three glycine residues, or four glycine residues.

In a further aspect, provided is a method of producing an antibody with low immunogenicity. In some embodiments, the method comprises: increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, or adding one or more glycine residues at the β turn of CDRH2 loop of the antibody. In some embodiments, the antibody is antibody that has not been previously made. In some embodiments, the method of producing an antibody with low immunogenicity comprises designing an antibody having one or more large cavities at the CDR regions, a significantly hydrophobic surface area at the CDRH3 loop, and/or one or more glycine residues at the β turn of CDRH2 loop of the antibody, and producing the antibody. In some embodiments, designing the antibody can be done by the methods or systems described herein.

In some embodiments, the antibody with low immunogenicity has one or more large cavities at the CDR regions, has a significantly hydrophobic CDRH3 loop, and contains one or more glycine residues at the β turn of CDRH2 loop.

to about

In a further aspect, provided is an antibody produced by the method. In some embodiments, the antibody has one or more large cavities at the CDR regions, has a significantly hydrophobic CDRH3 loop, and contains one or more glycine residues at the β turn of CDRH2 loop. In some embodiments, the antibody is antibody that has not been previously made.

In some embodiments, the sum of cavity volumes at CDR regions is about

or more. In some embodiments, the sum of cavity volumes at CDR regions is about

to about

In some embodiments, the sum of cavity volumes at CDR regions is about

In a further aspect, provided is a method of identifying an antibody with low immunogenicity. In some embodiments, the method comprises analyzing two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of one or more glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the methods described herein further comprises using a model which is able to analyze one or more of the following features: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody.

In some embodiments, the crystal structural of the antibody is available, which is used to calculate the cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop, and the method comprises constructing a model, which comprises two features: cavity volume at the CDR regions of the antibody and hydrophobicity of CDRH3 loop of the antibody; wherein if the antibody has one or more large cavities at the CDR regions and a significantly hydrophobic CDRH3 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity. In some embodiments, the cavity volume at the CDR regions is calculated by the FPOCKET program and hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area.

In some embodiments, the crystal structure of the antibody is unavailable, and the ABodyBuilder program is used to predict the antibody structure. The cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop are calculated with the modeled structure. The method comprises constructing a model, which analyzes two features: hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of one or more glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has a significantly hydrophobic CDRH3 loop and contains one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to having low immunogenicity. In some embodiments, the hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area.

In some embodiments, the model is called SVM model, which uses Support Vector Machine learning technology; features of antibodies with known immunogenicity are input into SVM model; parameters of SVM model are optimized in leave-one-out experiment; the trained SVM model is used for identifying or predicting immunogenicity.

In some embodiments, the sum of cavity volumes at CDR regions is about

or more. In some embodiments, the sum of cavity volumes at CDR regions is about

to about

In some embodiments, the sum of cavity volumes at CDR regions is about

In a further aspect, provided is a system for identifying or predicting antibody immunogenicity. In some embodiments, the system comprising a model described herein.

In some embodiments, the system comprises:

at least one processor comprising a prediction module which predicts the structure of the test antibody and an analysis module which constructs a SVM model for analyzing immunogenicity using Support Vector Machine learning technology; wherein features described herein of antibodies with known immunogenicity are calculated and input into SVM model; parameters of SVM model are optimized in leave-one-out experiment; the trained SVM model is used for identifying or predicting immunogenicity of the test antibody;

an output device for outputting immunogenicity of the test antibody.

In some embodiments, the analysis module analyzes two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of Gly at CDRH2 β turn of the antibody; wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity.

In some embodiments, the crystal structural of the antibody is available, which is used to calculate the cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop, and the analysis module analyzes two features: cavity volume at the CDR regions of the antibody and hydrophobicity of CDRH3 loop of the antibody; wherein if the antibody has one or more large cavities at the CDR regions and a significantly hydrophobic CDRH3 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity. In some embodiments, the cavity volume at the CDR regions is calculated by the FPOCKET program and hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area.

In some embodiments, the crystal structure of the antibody is unavailable, the ABodyBuilder program is used to predict the antibody structure, and then hydrophobicity of the CDRH3 loop is calculated with the predicted structure of the antibody. The analysis module analyzes two features: hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of one or more glycine residues at CDRH2 β turn of the antibody; wherein if the antibody has a significantly hydrophobic CDRH3 loop and contains one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified or predicted as having low immunogenicity or likely to have low immunogenicity. In some embodiments, the hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area.

In some embodiments, the sum of cavity volumes at CDR regions is about

or more. In some embodiments, the sum of cavity volumes at CDR regions is about

to about

In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%or more. In some embodiments, the significantly hydrophobic CDRH3 loop has a ratio of hydrophobic surface to total surface for the CDRH3 loop of about 63%to 80%. In some embodiments, the β turn of CDRH2 loop (VH 52-56) has at least one glycine residue.

In some embodiments, the sum of cavity volumes at CDR regions is about

Antibodies designed or predicted by the methods or systems described herein can be produced by conventional methods. For example, according to the amino acid sequence obtained by the methods or system described herein, the corresponding coding nucleic acid sequence can be designed by known methods, which can be introduced into host cells, such as CHO cells, HEK cells (e.g., HEK293F) , BHK cells, Cos1 cells, Cos7 cells, CV1 cells or mouse L cells, and the antibodies can be produced by culturing the cells in a suitable medium.

EXAMPLES

Sequences and structures of antibody variable regions

Sequences for the approved therapeutic antibodies were downloaded from the United States patent applications (patft. uspto. gov) and the KEGG drug database ( https: //www. genome. jp/kegg/drug) . A total of 52 humanized and full human antibodies (Table 3 and 4) approved by the FDA as of April 2018 excluding bispecific antibodies were collected.

was also excluded for having a similar chemical origin to another approved antibody

Only the Fv domains were considered since the other parts of the collected antibodies are innate and essentially non-immunogenic. A total of 52 antibodies were used when sequence-based features were investigated. The amino acid residues of the heavy chain (VH) and light chain (VL) were assigned Kabat numbering with the software tool ANARCI (Dunbar J, et al., Bioinformatics, 2016, 32: 298-300. ) . An expanded definition was used for the six loops at the CDR regions in comparison with the classical Kabat, including the additional VH positions 26-30 for CDRH1 and 49 for CDRH2.

Crystal structures were available for 29 out of the 52 antibodies in the Protein Data Bank (PDB, https: //www. rcsb. org) . Unless specifically indicated, statistical analysis of structural features for the 29 proteins (Table 3) was performed. Structural models for the other 23 antibodies (Table 4) were built with the online server ABodyBuilder to test the structure-based immunogenicity prediction algorithm developed in this study. Molecular graphics and analyses were performed with UCSF CHIMERA (Pettersen EF, et al., J Comput Chem, 2004, 25: 1605-12. ) for the protein structures.

Immunogenic and non-immunogenic antibodies

The level of reported immunogenicity was obtained in the prescribing information for each therapeutic antibody at the FDA website ( https: //www. accessdata. fda. gov/scripts/cder/daf) . The observed immunogenicity of therapeutic proteins is highly dependent on several factors including assay methodology, underling disease, concomitant medications. If possible, data generated at the same conditions were used to separate immunogenic antibodies from non-immunogenic antibodies. For example, when the test results of multiple doses were available, the rate of antibody development was considered for patients receiving 10 mg/kg or similar dose.

Antibodies were classified as having low immunogenicity or non-immunogenic when treatment-emergent anti-antibody response (AAR) was reported in less than 2%of patients. Otherwise, the antibodies were considered having high immunogenicity or immunogenic. AAR was considered detectable if patients in studies were tested at multiple time points and the AAR was detected at least once during treatment. If AAR data with and without concomitant immunosuppression were concurrently reported, results from patients not taking immunosuppressants were used.

The classification of antibody immunogenicity described in the present study was adopted in order to have a similar number of immunogenic and non-immunogenic antibodies for the convenience of statistical analysis. As a result, 28 out of the 52 antibodies are considered immunogenic and the other 24 are non-immunogenic (Table 3 and Table 4) . For the 29 antibodies with available crystal structures, 15 are immunogenic and 14 are non-immunogenic.

Definition of features

Residue humanness score is defined similarly to residue conservation score measured by the self-substitution score from the sequence profile as in the previous study. Sequence profiles were obtained by three rounds of PSI-BLAST searches against human sequences with the BLOSUM62 substitution matrix. The humanness score at position i is defined as

where M _ir is the self-substitution score in the position-specific substitution matrix generated from PSIBLAST for the residue type r at sequence position i, and B _rr is the diagonal element of BLOSUM62 for residue type r, and r was defined as a rare residue in case S _human is less than -6. The sum of humanness scores was calculated for residues at the CDR regions and framework, respectively.

The FPOCKET program was used to identify all pockets in the Fv domain with default parameters and calculate the volume for each pocket. The program was originally developed to identify cavities at the protein surface likely to bind small compounds. A pocket is considered to be located at the CDR region if two third of the surrounding atoms are of CDR residues. The sum of cavity volumes is calculated in case multiple pockets are identified at the CDR region.

The hydrophobicity of CDRH3 is defined as the sum of solvent accessible surface of all hydrophobic atoms thereof. The carbon, sulfur, and nitrogen atoms (excluding the non-protonated nitrogen atom of histidine side chain) are considered hydrophobic atoms. Hydrogen atoms, which were missing in the crystal structures, were added with the REDUCE program (Word JM, et al., J Mol Bio, 1999, 285: 1735-47) . The solvent probe was set to

The atomic radii of carbon, nitrogen, oxygen, sulfur, and polar hydrogen were set to 1.8, 1.65, 1.4, 1.85, and

respectively, and non-polar hydrogen atoms were ignored.

Training and prediction procedure

To predict the immunogenicity of a therapeutic antibody, three types of features were extracted: cavity volume at the CDR regions, hydrophobicity of CDRH3 loop, and the information of presence/absence of Gly (1 or 0) at CDRH2 turn. The library for support vector machines LIBSVM-3.22 was downloaded from http: //www. csie. ntu. edu. tw/～cjlin/libsvm and used for combining features and machine learning classification. The optimal parameters for model training were derived by the recommended method through cross-validation. The prediction accuracy is defined as the number of correctly predicted immunogenic and non-immunogenic antibodies divided by the number of total predictions.

A “grid-search” on C and γ using cross-validation is recommended for the SVM model. Various pairs of (C, γ) values were tried and the one with the best cross-validation accuracy was picked. It was found that trying exponentially growing sequences of C and γ was a practical method to identify good parameters (for example, C=2 ^-5, 2 ^-3, ……. 2 ¹⁵, γ=2 ^-15, 2 ^-13, ….. 2 ³) .

Example 1. Irrelevance of humanness and immunogenicity

The study started from the investigation of two frequently used features in antibody immunogenicity research: humanness score and T cell epitope prediction. The humanness score, i.e., the distance to the human consensus sequence, was defined in a different way from other methods (Gao SH, et al., BMC Biotechnol, 2013, 13: 55; Clavero-Alvarez A et al., Sci Rep, 2018, 8: 14820. ) . Specifically, the study focused on the CDR regions to maximize the gap between the calculated values for different types of antibodies. Indeed, the humanness scores of full human antibodies are significantly higher than those of humanized antibodies at the CDR regions (p value < 10 ^-15) but the difference is minimal for the framework (Table 1) . Unfortunately, when the 52 antibodies were divided into humanized antibodies and full human antibodies, the mean value of the calculated humanness scores of immunogenic antibodies was equivalent to that of non-immunogenic antibodies and the distribution ranges were similar within either group (Figure 1) . That is, the immunogenicity of therapeutic antibodies could not be predicted based on the humanness score.

See Table 1 for details.

Table 1. Similar humanness scores for immunogenic and non-immunogenic therapeutic antibodies

The humanness score was calculated as a negative number close to zero for human-like sequences and large in magnitude for non-human-like sequences.

Example 2. Irrelevance of T cell epitope prediction results and immunogenicity

T cell epitope is one of factors contributing to immune responses. The activation of helper T cell is essential for B cell proliferation, antibody class switching, and an increase in antibody production. Recognition of linear epitopes bound by MHC class II molecules on the surface of antigen presenting cells is a critical step for T cell activation. The study predicted T cell epitopes for the Fv domains of immunogenic and non-immunogenic antibodies using the state-of-the-art consensus method (Wang P et al., BMC bioinformatics, 2011, 11: 568) and the online server ( http: //tools. iedb. org/mhcii/) in April 2018. On average, immunogenic antibodies do not contain more fragments as a good binder against the full reference set of MHC class II alleles than non-immunogenic antibodies (Table 2) . The results were essentially indistinguishable for the two types of antibodies even when the study changed the cut-off value for the predicted T cell epitopes. Therefore, the immunogenicity of therapeutic antibodies could not be inferred by T cell epitope prediction results.

See Table 2 for details.

Table 2. T cell epitope prediction results averaged for immunogenic and non-immunogenic antibodies

^aThe lowest rank (the best binder) calculated for a reference panel of 27 alleles against all fragments of the Fv domain of an antibody. ^bNo. of alleles if they predicted as a percentile rank < 3%against anyone of the Fv fragments. ^cNo. of total allele-fragment pairs with a percentile rank < 3%.

Example 3. Idiotype-anti-idiotype complex structures

The crystal structure of an idiotype-anti-idiotype complex precisely shows how the anti-idiotypic antibody binds the idiotypic antibody. The study searched the PDB database with the key words “antibody” and “complex” and found 6 idiotype-anti-idiotype complexes from 1883 results in April 2018. As shown in Figure 2, the anti-idiotypic antibody binds exclusively to the CDR regions of the idiotypic antibody, especially, the CDRH2 loop and the CDRH3 loop. Only a few atoms at the framework of the idiotypic antibody, i.e., B cell epitopes residing in the CDR regions, make direct contact with the anti-idiotypic antibody. Motivated by the crystal structures of the 6 complexes, the study focused on the CDR regions in the following studies in order to identify features that distinguish immunogenic antibodies from non-immunogenic antibodies.

Example 4. Cavity at CDR regions

By visual analysis of crystal structures (Figure 3) , the study found that the central cavities between the heavy chains and the light chains of immunogenic antibodies were smaller than those of non-immunogenic antibodies. In addition, the length of CDRH3 loops of the 15 immunogenic antibodies (9.7 residues in average) is slightly shorter than that of the 14 non-immunogenic antibodies (11.4) . The CDRH3 loop of immunogenic antibodies tend to form a curve to fill the cavity at the CDR regions, providing a smooth surface at the CDR regions of an idiotypic antibody that allows the binding of anti-idiotypic antibodies. On the contrary, the relatively long and rigid CDRH3 loop of non-immunogenic antibodies protrude directly into solvent, which makes it difficult for the anti-idiotypic antibodies to bind the CDR regions (Figure 3) .

For quantitative evaluation, the study used the FPOCKET program (Le Guilloux V, et al., BMC bioinformatics, 2019, 10: 168. ) to identify the cavities at the CDR regions for an all-atom antibody structure and calculate the total cavity volume (Table 3) . Consistent with visual analysis, the cavity volume of immunogenic antibodies

was found to be smaller than that of non-immunogenic antibodies

The difference is statistically significant according to t student test (p value <0.05) . Due to the large size of CDR regions, the study assumes that anti-idiotypic antibodies are not likely to bind the surface patch with a deep cavity, which is ideal for the binding of a small molecule on the other hand.

See Table 3 for details

Table 3. Features distinguishing immunogenic antibodies from non-immunogenic antibodies

Crystal structures were used for statistical analysis.

Example 5. Hydrophobicity of CDRH3 loop

Hydrophobic interactions play a critical role for the tight binding of a protein complex. However, no hydrophobicity difference was observed for the whole CDR regions between immunogenic antibodies (58.5 ±2.8%) and non-immunogenic antibodies (59.2 ± 2.4%) . The percentage of hydrophobic surface area in the total surface area actually varies within a small range for protein stability reason. Since hydrophobic interactions are favorable for protein association, theoretically, it is easy for the anti-idiotypic antibodies to bind the therapeutic antibodies with a hydrophobic CDRH3 loop, which usually is located at the center of the binding site (Figure 2) . Unexpectedly, the CDRH3 loops of immunogenic antibodies have a significantly smaller hydrophobic surface area (p value <0.05) than that of non-immunogenic antibodies (Table 3) . This is not caused by the slight difference of loop length. In fact, the ratio of hydrophobic surface to total surface is also smaller for the CDRH3 loops of the 15 immunogenic antibodies (62.5 ± 10.2%) than that of the 14 non-immunogenic antibodies (67.4 ±4.3%) . The study assumes that the antigen receptors, which potentially bind the therapeutic antibody with a hydrophobic CDRH3 loop, also bind similar self antibodies. As a result, the immature B cells with the cross-reactive antigen receptors on the surface are eliminated or inactivated during the early development and the foreign antibodies with a hydrophobic CDRH3 show low immunogenicity.

Example 6. Number of Gly at CDRH2 turn

Besides the CDRH3 loop, the CDRH2 loop is frequently located at the center of the binding site of an idiotypic antibody (Figure 2) . In general, the β turn of CDRH2 loop (VH 52-56) is glycine rich for antibodies from various species for structural reasons but contains no glycine in 7 out of the 52 therapeutic antibodies. Interestingly, 6 of the 7 antibodies are immunogenic. Despite the small size of the analyzed data set, the study infers that antibodies containing no glycine at the CDRH2 turn are immunogenic. Actually, the study found the other humanized antibody, huBrE-3, with such an unusual CDRH2 loop. The anti-drug antibodies were detectable in 1 out of 7 patient’s serum in the initial clinical evaluation (Kramer EL, et al., Clinical cancer research: an official journal of the American Association for Cancer Research 1998, 4: 1679-88. ) .

Example 7. Prediction of immunogenicity

Support vector machine (SVM) learning technology was used to integrate the features discussed above for immunogenicity prediction. The study achieved an impressive accuracy of 83%for the 29 therapeutic antibodies when the two features, cavity volume at the CDR regions and hydrophobicity of CDRH3 loop calculated from the crystal structures, were used in the leave-one-out experiment. The accuracy, however, was decreased to 76%by combining the information of presence/absence glycine at CDRH2 turn additionally due to over fit resulting from the small data set. Moreover, the SVM model trained with the two effective features of the 29 antibodies shows no predictive ability (48%accuracy) for the 23 test antibodies, of which the crystal structures are unavailable and the modeled structures have to be used for prediction.

The study found that the cavity volume calculated by FPOCKET could be significantly affected by the coordinate errors for the modeled structures resulting in low prediction accuracy. To investigate the effect of inaccurate structures, the study used ABodyBuilder (Leem J, et al., MAbs, 2016, 8: 1259-1268. ) to predict the structures for 5 immunogenic antibodies, of which no cavity at the CDR regions were identified in the crystal structures, adalimumab, panitumumab, ustekinumab, daclizumab, and efalizumab (Table 3) . The observed structure was excluded from the template library for each prediction. Using the modeled structures, the study found cavities at the CDR regions for all of 5 antibodies with an average volume of

which makes them indistinguishable from non-immunogenic antibodies. On the other hand, the exposed hydrophobic surface area of CDRH3 was fairly consistent with that of the observed structures for the 5 antibodies despite a small increase in most cases (339, 310, 245, 168, and

versus 340, 176, 310, 112, and

respectively) . The study thus utilized another set of features, hydrophobicity of CDRH3 and the information of the presence/absence of glycine at CDRH2 turn, for SVM classification and achieved an accuracy of 79%in leave-one-out experiment for the 29 antibodies calculated with crystal structures. When the trained SVM model was used for the modeled structures of 23 test antibodies (Table 4) , a meaningful accuracy of 65%was obtained compared to no predictive ability of the above-mentioned SVM model, for which the training features contain the cavity volume at the CDR regions. Therefore, dependent on availability of the crystal structure, different SVM models trained with appropriate features should be used to maximize the prediction accuracy.

As shown in Table 3 and Table 4, the exposed hydrophobic surface area of CDRH3 of immunogenic antibodies was smaller than that of non-immunogenic antibodies for both crystal structures and modeled structures. However, the surface area of CDRH3 could be systematically overestimated for the structures modeled by ABodyBuilder. Since the crystal structures usually were unavailable for the predicted antibodies, it was reasonable to use the modeled structures for SVM model training and prediction consistently. When hydrophobicity of CDRH3 loops calculated from the modeled structures and the information of the presence/absence of Gly at CDRH2 turn were used for SVM classification, the study achieved an accuracy of 78%in leave-one-out experiment for the 23 test antibodies. The study then used the trained SVM model to predict the immunogenicity for 11 therapeutic antibodies approved after April 2018 including cemiplimab, emapalumab, erenumab, fremanezumab, galcanezumab, lanadelumab, mogamulizumab, polatuzumab, ravulizumab, risankizumab, and romosozumab. The prediction was made with modeled structures and the result was correct for 7 out of the 11 antibodies. Similarly, the present disclosure successfully predicted the high immunogenicity of bococizumab, which was discontinued after phase 3 clinical trial (Ridker PM, et al., N Engl J Med, 2017, 376: 1517-1526. ) .

See Table 4 for details.

Table 4. Prediction results for the modeled structures of therapeutic antibodies

^aPredicted immunogenic and non-immunogenic antibodies were indicated by “1” and “0” , respectively

Claims

A method of selecting a candidate therapeutic antibody with low immunogenicity comprising:

selecting an antibody having one or more large cavities at the CDR regions of the antibody, having a significantly hydrophobic surface area at the CDRH3 loop of the antibody, or containing one or more glycine residues at the β turn of CDRH2 loop of the antibody.
A method of reducing immunogenicity of an antibody comprising:

increasing the cavity volume at the CDR regions of the antibody, increasing the hydrophobic surface area at the CDRH3 loop of the antibody, or introducing one or more glycine residues at the β turn of CDRH2 loop of the antibody.
An antibody produced by the method of claim 2.
The antibody of claim 3, wherein the antibody has one or more large cavities at the CDR regions, has a significantly hydrophobic CDRH3 loop, and contains one or more glycine residues at the β turn of CDRH2 loop.
A pharmaceutical composition comprising the antibody of claim 3 or claim 4, optionally together with a suitable carrier, excipient or diluent.
A method of identifying an antibody with low immunogenicity comprising:

analyzing two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of one or more glycine residues at CDRH2 β turn of the antibody;

wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is identified as having low immunogenicity or likely to have low immunogenicity.
A method of predicting antibody immunogenicity comprising:

analyzing two or more features selected from the list: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody, or information of the presence/absence of one or more glycine residues at CDRH2 β turn of the antibody;

wherein if the antibody has two or more of one or more large cavities at the CDR regions, a significantly hydrophobic CDRH3 loop, or one or more glycine residues at the β turn of CDRH2 loop, the antibody is predicted as having low immunogenicity or likely to have low immunogenicity.
The method of any one of claims 6-7, wherein the cavity volume at the CDR regions is calculated by the FPOCKET program and hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area.
The method of any one of claims 6-8, the method comprises using a model which is able to analyze one or more of the following features: cavity volume at the CDR regions of the antibody, hydrophobicity of CDRH3 loop of the antibody and information of the presence/absence of glycine residues at CDRH2 β turn of the antibody;

wherein the model is called SVM model, which uses Support Vector Machine learning technology; features of antibodies with known immunogenicity are calculated and input into SVM model; parameters of SVM model are optimized in leave-one-out experiment; the trained SVM model is used for predicting immunogenicity.
The method of any one of claims 6-9, wherein the crystal structural of the antibody is available, the sum of cavity volumes at the CDR regions and hydrophobicity of CDRH3 loop are directly calculated.
The method of any one of claims 6-9, wherein the crystal structure of the antibody is unavailable, the ABodyBuilder program is used to predict the antibody structure, and then the sum of cavity volume at the CDR regions and the hydrophobicity of CDRH3 loop are calculated with the modeled structure.
The method of any one of claims 1-2 and 6-11, wherein the sum of cavity volumes at the CDR regions is
or more.
The method of any one of claims 1-2 and 6-11, wherein the sum of cavity volumes at the CDR regions is
or more.
The method of any one of claims 1-2 and 6-11, wherein the sum of cavity volumes at the CDR regions is
or more.
The method of any one of claims 1-2 and 6-14, wherein the ratio of hydrophobic surface to total surface for the CDRH3 loop is 65%or more.
The method of any one of claims 1-2 and 6-15, wherein the β turn of CDRH2 loop (VH 52-56) has at least one glycine residue.
The method of any one of claims 1-2 and 6-15, wherein the β turn of CDRH2 loop (VH 52-56) has at least two glycine residues.
A system for identifying or predicting antibody immunogenicity comprising:

an input device for inputting features of an antibody;

and a module for calculating the sum of cavity volumes at the CDR regions of the antibody, the hydrophobicity of the CDRH3 loop of the antibody, or providing information of the presence/absence of glycine residues at CDRH2 β turn of the antibody, and analyzing the immunogenicity of the antibody; and

an output device for outputting immunogenicity of the antibody.
The system of claim 18, wherein the system comprises at least one processor comprising a prediction module which predicts the structure of a test antibody and an analysis module which constructs a SVM model for analyzing immunogenicity using Support Vector Machine learning technology; wherein features described of antibodies with known immunogenicity are calculated and input into SVM model; parameters of SVM model are optimized in leave-one-out experiment; the trained SVM model is used for identifying or predicting immunogenicity of the test antibody.
The system of any one of claim 18-19, wherein the cavity volume at the CDR regions is calculated by the FPOCKET program and hydrophobicity of CDRH3 loop is evaluated by exposed hydrophobic surface area.
The system of any one of claims 18-20, wherein the crystal structural of the antibody is available, the sum of cavity volumes at the CDR regions and hydrophobicity of CDRH3 loop are directly calculated.
The system of any one of claims 18-20, wherein the crystal structure of the antibody is unavailable, the ABodyBuilder program is used to predict the antibody structure, and then the sum of cavity volumes at the CDR regions and hydrophobicity of the CDRH3 loop are calculated with the modeled structure.
Use of the method of any one of claims 1-2 and 6-17, or the system of any one of claims 18-22 in producing an antibody with low immunogenicity.