Treatment of autoimmune diseases

Abstract

Antibodies for binding to PADs and their use in therapy.

Description

[Technical Field]

[0001] Rheumatoid arthritis (RA) is a common autoimmune disease with a chronic progressive phenotype. It is a chronic systemic inflammatory disease affecting small and large joints, resulting in progressive joint destruction, loss of function, and chronic pain / fatigue with increasing disability. [Background technology]

[0002] Despite the development of newer therapies for the treatment of RA [1-4], particularly the advent of biologic therapies that block tumor necrosis factor alpha (TNF-alpha), interleukin-6 receptors, or deplete B cells, many patients still suffer from poorly controlled active disease, and response rates remain low. Current targeted therapies typically begin with disease-modifying antirheumatic drugs (DMARDs) such as methotrexate, followed by targeted therapies such as newer biologic therapies, which are cycled based on a "treat to target" principle, attempting to achieve remission, low disease activity, and / or ACR70 (70% improvement in disease activity). However, these have met with limited success, with only 20-30% of patients achieving ACR70. There is a clear need for more effective therapies to treat RA, ideally those that target causative disease pathways.

[0003] Histologically, RA is characterized by synovitis accompanied by T and B cell infiltrates, often organized into germinal center-like structures, and macrophages, dendritic cells, and neutrophils, the latter of which are particularly abundant in synovial fluid in the early stages of the disease.[5-7] Although these histological features bear little resemblance to other autoimmune or inflammatory conditions, it has become clear that RA possesses unique features that cannot be easily explained by the traditional T cell-centered pathogenesis hypothesis: an early event is a loss of tolerance.

[0004] Peptidyl arginine deiminase (PAD) is a family of five isozymes (PAD1, 2, 3, 4, and 6) encoded by different genes in the human genome [8]. PADs are calcium-dependent enzymes that catalyze a post-transcriptional modification known as citrullination, which is the conversion of the basic charged amino acid residue arginine to the neutral residue citrulline. Citrullinated proteins induce the production of anti-citrullinated protein antibodies (ACPAs) and cyclic citrullinated peptides (CCPs). ACPAs and CPPs may contribute to the perpetuation of autoimmune responses.

[0005] WO2012026309(A9) [9] describes anti-PAD4 antibodies.

[0006] WO 2014 / 086365(A1)

[10] describes antibodies that bind to rabbit PAD2 (rPAD2).

[0007] WO 2016 / 155745(A1)

[11] suggests cross-reactive mouse monoclonal antibodies against PAD2, PAD4 and PAD3.

[0008] WO2016143753(A1)

[12] describes anti-PAD4 antibodies.

[0009] Aosasa et al. 2021

[13] described a chimeric anti-PAD2 antibody.

[0010] Thus, there remains a need for effective compositions for treating autoimmune diseases such as RA. Summary of the Invention

[0011] The present invention relates to an anti-PAD4 antibody having high affinity and specificity for human and cynomolgus monkey PAD4. The present invention relates to an anti-PAD2 antibody having high affinity and specificity for human, cynomolgus monkey, and mouse PAD2.

[0012] The present invention further relates to bispecific antibodies with high affinity and specificity for PAD2 (human, mouse and cynomolgus monkey) and PAD4 (human and cynomolgus monkey, or mouse).

[0013] The present invention also relates to anti-PAD4 and anti-PAD2 antibodies and anti-PAD2 / PAD4 bispecific antibodies that are highly potent in inhibiting PAD4 and / or PAD2 activity and that are potent in inhibiting PAD activity in the synovial fluid of rheumatoid arthritis (RA) patients.

[0014] The present invention relates to the treatment of autoimmune diseases through targeting both PAD2 and PAD4. In particular, the present invention relates to the use of a combination of anti-PAD2 and anti-PAD4 antibodies, or an anti-PAD2 / 4 bispecific antibody, in the treatment of autoimmune diseases characterized by elevated PAD activity, such as RA. The present invention is supported by data provided herein for the first time showing that the activities of PAD2 and PAD4 in RA disease are non-redundant. Surprisingly, targeting both PAD2 and PAD4 has been shown to be necessary and sufficient to neutralize PAD activity in the whole blood, serum, and synovial fluid of RA patients. [Brief explanation of the drawings]

[0015] [Figure 1] PAD2- and PAD4-dependent production of citrullinated RA antigens. ELISA measuring the production of citrullinated antigens (fibrinogen beta chain and alpha enolase) by PAD2 and PAD4. [Figure 2A]The potency of prior art humanized anti-PAD4 antibodies, evaluated by the H3 histone citrullination assay (Figure 2A) or the BAEE PAD activity assay (Cayman Chemical) (Figure 2B). [Figure 2B] The potency of prior art humanized anti-PAD4 antibodies, evaluated by the H3 histone citrullination assay (Figure 2A) or the BAEE PAD activity assay (Cayman Chemical) (Figure 2B). [Figure 3A] PAD2 and PAD4 expression Figure 3A: PAD4 and PAD2 protein levels in sera from rheumatoid arthritis (RA) patients compared to healthy donors (HD). (RA) n = 90, healthy donors (HD) n = 24. Bars indicate the median. The median of PAD2 in HD is <LLOD. Mann-Whitney test. ****p < 0.0001. PAD measured with Cayman ELISA kit. Figure 3B: PAD2 and PAD4 protein levels are high in RA synovial fluid. RA synovial fluid (SF) n = 5. Bars indicate the median. Synovial fluid was diluted and PAD levels were measured with Cayman ELISA kit. Figure 3C: RNA expression profiles of PAD4 in immune cells from human and cynomolgus macaque (cyno). Figure 3D: RNA expression profiles of PAD4 in immune cells from human and cynomolgus macaque (cyno). Expression of PAD2 and PAD4 is shown compared to expression in human B cells. [Figure 3B]PAD2 and PAD4 Expression Figure 3A: PAD4 and PAD2 protein levels in sera from rheumatoid arthritis (RA) patients compared to healthy donors (HD). (RA) n = 90, healthy donors (HD) n = 24. Bars indicate the median. The median of PAD2 in HD is <LLOD. Mann-Whitney test. ****p < 0.0001. PAD measured with Cayman ELISA kit. Figure 3B: PAD2 and PAD4 protein levels are high in RA synovial fluid. RA synovial fluid (SF) n = 5. Bars indicate the median. Synovial fluid was diluted and PAD levels were measured with Cayman ELISA kit. Figure 3C: RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkeys (Cynomolgus, cyno). Figure 3D: RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkeys (cyno). Expression of PAD2 and PAD4 is shown compared to expression in human B cells. [Figure 3C] PAD2 and PAD4 Expression Figure 3A: PAD4 and PAD2 protein levels in sera from rheumatoid arthritis (RA) patients compared to healthy donors (HD). (RA) n = 90, healthy donors (HD) n = 24. Bars indicate the median. The median of PAD2 in HD is <LLOD. Mann-Whitney test. ****p < 0.0001. PAD measured with Cayman ELISA kit. Figure 3B: PAD2 and PAD4 protein levels are high in RA synovial fluid. RA synovial fluid (SF) n = 5. Bars indicate the median. Synovial fluid was diluted and PAD levels were measured with Cayman ELISA kit. Figure 3C: RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkeys (Cynomolgus, cyno). Figure 3D: RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkeys (cyno). Expression of PAD2 and PAD4 is shown compared to expression in human B cells. [Figure 3D]PAD2 and PAD4 Expression Figure 3A: PAD4 and PAD2 protein levels in sera from rheumatoid arthritis (RA) patients compared to healthy donors (HD). (RA) n = 90, healthy donors (HD) n = 24. Bars indicate the median. The median of PAD2 in HD is <LLOD. Mann-Whitney test. ****p < 0.0001. PAD measured by Cayman ELISA kit. Figure 3B: PAD2 and PAD4 protein levels are high in RA synovial fluid. RA synovial fluid (SF) n = 5. Bars indicate the median. Synovial fluid was diluted and PAD levels were measured by Cayman ELISA kit. Figure 3C: RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkeys. Figure 3D: RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkeys (cyno). Expression of PAD2 and PAD4 is shown compared to expression in human B cells. [Figure 4A] Cell Surface Quantification of PAD2 and PAD4 Quantification of PAD2 (Figure 4B) and PAD4 (Figure 4A) on the surface of immune cells. [Figure 4B] Cell Surface Quantification of PAD2 and PAD4 Quantification of PAD2 (Figure 4B) and PAD4 (Figure 4A) on the surface of immune cells. [Figure 5] Interferometric Scattering (iSCAT) of Individual Protein Molecules in the Vicinity of the PAD4 and Anti-PAD4 Fab Surfaces Anti-PAD4 = clone 42 (48LO0063, IgG or Fab). [Figure 6] Interferometric Scattering (iSCAT) of Individual Protein Molecules in the Vicinity of the PAD4 and Anti-PAD4 IgG Surfaces Anti-PAD4 = clone 42 (48LO0063, IgG or Fab). [Figure 7] iSCAT of PAD2 and Anti-PAD2 Fab / IgG iSCAT of PAD2 and anti-PAD2 Fab and IgG. Anti-PAD2 = clone 22 (IgG or Fab). [Figure 8A] Overview of iSCAT Figure 8A: PAD4, Figure 8B: PAD2. [Figure 8B] Overview of iSCAT. Figure 8A: PAD4, Figure 8B: PAD2. [Figure 9A] Potency assay optimization Figure 9A: Optimization of potency assay parameters. Figure 9B: PAD4 Fab Histone H3 assay. 3h 45min PAD4 incubation. [PAD4]=15pg.ml. [Figure 9B] Potency assay optimization Figure 9A: Optimization of potency assay parameters. Figure 9B: PAD4 Fab Histone H3 assay. 3h 45min PAD4 incubation. [PAD4]=15pg.ml. [Figure 10A] Bispecific antibody formats Figure 10A: Monovalent Duet mAb. Figure 10B: Bivalent bispecific Bis3. [Figure 10B] Bispecific antibody formats Figure 10A: Monovalent Duet mAb. Figure 10B: Bivalent bispecific Bis3. [Figure 11] Histone-H3 Activity Assay. A histone H3 substrate is coated onto a plate, and active PAD in the sample deaminates the arginine residue to form citrulline. These citrullinated epitopes are then detected by standard immunoassay methods. This assay can be used to demonstrate target association in both the circulatory and synovial compartments (where PAD2 activity is higher) without sample dilution. [Figure 12A] Inhibition of PAD activity Figure 12A: Inhibition of PAD2 in synovial fluid (1:1000 dilution). Figure 12B: Inhibition in neutrophil supernatant (1:250 dilution). PAD activity in diluted synovial fluid samples was determined using a histone-H3 PAD activity assay. EDTA sequesters calcium, inhibiting PAD activity and serving as a background control. [Figure 12B] Inhibition of PAD activity Figure 12A: Inhibition of PAD2 in synovial fluid (1:1000 dilution). Figure 12B: Inhibition in neutrophil supernatant (1:250 dilution). PAD activity in diluted synovial fluid samples was determined using a histone-H3 PAD activity assay. EDTA sequesters calcium, inhibiting PAD activity and serving as a background control. [Figure 13]PAD2 Antibody Affinity Binding affinity KD (nM) for anti-PAD2 monoclonal antibodies (as Fab). Biotinylated PAD2 was captured on a CM5 / C1-streptavidin surface. Affinity: KD (nM). Data shown are the average, n = 2-9 experiments (Table 72). [Figure 14] PAD4 antibody affinity. Binding affinity KD (nM) for anti-PAD4 monoclonal antibodies (as Fab). Biotinylated PAD4 was captured on a CM5 / C1-streptavidin surface. Affinity: KD (nM). Data shown are the average of n=1 to 6 experiments (Table 74). [Figure 15] DuetMab PAD2 / PAD4 binding behavior. iSCAT of the PAD2 / PAD4 bispecific antibody DuetMab. [Figure 16] Bis3 PAD2 / PAD4 binding behavior. iSCAT of Bis3 PAD2 / PAD4 bispecific antibodies. [Figure 17] DuetMab vs. Bis3 Binding Schematic showing the differences between Bis3 and DuetMab formats binding to PAD2 and PAD4 dimers. [Figure 18] Effect of Fc modifications on thermal stability The thermal stability of bispecific antibody formats in the context of different Fc modifications was assessed by Nano-DSF. Tonset values ​​for clones 01-12 are shown (°C) (Table 92). [Figure 19A] Accelerated Stability The tendency of DuetMab and Bis3 formats to aggregate at 40°C (Figure 19B) and 45°C (Figure 19A) compared to 4°C was assessed by HP-SEC. [Figure 19B] Accelerated Stability The tendency of DuetMab and Bis3 formats to aggregate at 40°C (Figure 19B) and 45°C (Figure 19A) compared to 4°C was assessed by HP-SEC. [Figure 20] Cynomolgus monkey research immunohistochemistry [Figure 21]In vitro cytokine release, plate-bound antibodies. Cytokine expression after exposure to Bis3 (clone 12) or DuetMab (clone 06) format antibodies (HD = high dose). [Figure 22A] Efficacy of the Bis3 format. Figure 22A: Study protocol. Figure 22B: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22C: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22D: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. Figure 22E: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. LD: low dose. HD: high dose. Abatacept: abatacept. Format: Bis3 (scFv PAD4, Fab: PAD2, name: clone 12). Numbers refer to individual experiments. [Figure 22B] Efficacy of the Bis3 format. Figure 22A: Study protocol. Figure 22B: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22C: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22D: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. Figure 22E: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. LD: low dose. HD: high dose. Abatacept: abatacept. Format: Bis3 (scFv PAD4, Fab: PAD2, name: clone 12). Numbers refer to individual experiments. [Figure 22C]Efficacy of the Bis3 format. Figure 22A: Study protocol. Figure 22B: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22C: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22D: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. Figure 22E: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. LD: low dose. HD: high dose. Abatacept: abatacept. Format: Bis3 (scFv PAD4, Fab: PAD2, name: clone 12). Numbers refer to individual experiments. [Figure 22D] Efficacy of the Bis3 format. Figure 22A: Study protocol. Figure 22B: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22C: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22D: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. Figure 22E: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. LD: low dose. HD: high dose. Abatacept: abatacept. Format: Bis3 (scFv PAD4, Fab: PAD2, name: clone 12). Numbers refer to individual experiments. [Figure 22E]Efficacy of the Bis3 format. Figure 22A: Study protocol. Figure 22B: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22C: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 36. Figure 22D: PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. Figure 22E: Endogenous PAD activity measured in plasma using a histone H3 citrullination assay from days 0 to 106. LD: low dose. HD: high dose. Abatacept: abatacept. Format: Bis3 (scFv PAD4, Fab: PAD2, name: clone 12). Numbers refer to individual experiments. [Figure 23A] Potency of anti-PAD2 and anti-PAD4 antibodies against recombinant PAD The potency of clones 12, 22, and 42 was directly compared to that of prior art anti-PAD2 and anti-PAD4 antibodies using an optimized histone-H3 citrullination ELISA and recombinant PAD2 and PAD4 (Figures 23A and 23B, respectively). [Figure 23B] Potency of anti-PAD2 and anti-PAD4 antibodies against recombinant PAD The potency of clones 12, 22, and 42 was directly compared to that of prior art anti-PAD2 and anti-PAD4 antibodies using an optimized histone-H3 citrullination ELISA and recombinant PAD2 and PAD4 (Figures 23A and 23B, respectively). [Figure 24A] PAD2 / PAD4 Specificity The affinity-optimized anti-PAD4 antibodies and bispecific antibody formats were specific for PAD2 and / or PAD4 and did not bind to PAD3 (Figure 24A). The affinity-optimized clones did not bind to PAD1 (Figure 24B). [Figure 24B] PAD2 / PAD4 Specificity The affinity-optimized anti-PAD4 antibodies and bispecific antibody formats were specific for PAD2 and / or PAD4 and did not bind to PAD3 (Figure 24A). The affinity-optimized clones did not bind to PAD1 (Figure 24B). [Figure 25A]Comparative potency assays of Bis3 and DuetMab formats. PAD activity was measured with a histone H3 citrullination assay. Figure 25A: Data represent one of five experiments using different RA synovial fluid samples. Figure 25B: Whole blood. Data represent one representative sample of whole blood. Bis3 clone = clone 12. DuetMab clone = clone 06. [Figure 25B] Comparative potency assays of Bis3 and DuetMab formats. PAD activity was measured with a histone H3 citrullination assay. Figure 25A: Data represent one of five experiments using different RA synovial fluid samples. Figure 25B: Whole blood. Data represent one representative sample of whole blood. Bis3 clone = clone 12. DuetMab clone = clone 06. [Figure 26A] Efficacy of anti-PAD2 and anti-PAD4 antibodies in RA synovial fluid PAD activity was measured with a histone H3 citrullination assay using different RA synovial fluid samples: Figure 26A: Sample 1. Figure 26B: Sample 2. Figure 26C: Sample 3. Figure 26D: Sample 5. Figure 26E: Sample 4. [Figure 26B] Efficacy of anti-PAD2 and anti-PAD4 antibodies in RA synovial fluid PAD activity was measured with a histone H3 citrullination assay using different RA synovial fluid samples: Figure 26A: Sample 1. Figure 26B: Sample 2. Figure 26C: Sample 3. Figure 26D: Sample 5. Figure 26E: Sample 4. [Figure 26C] Efficacy of anti-PAD2 and anti-PAD4 antibodies in RA synovial fluid PAD activity was measured with a histone H3 citrullination assay using different RA synovial fluid samples: Figure 26A: Sample 1. Figure 26B: Sample 2. Figure 26C: Sample 3. Figure 26D: Sample 5. Figure 26E: Sample 4. [Figure 26D] Efficacy of anti-PAD2 and anti-PAD4 antibodies in RA synovial fluid PAD activity was measured with a histone H3 citrullination assay using different RA synovial fluid samples: Figure 26A: Sample 1. Figure 26B: Sample 2. Figure 26C: Sample 3. Figure 26D: Sample 5. Figure 26E: Sample 4. [Figure 26E]Efficacy of anti-PAD2 and anti-PAD4 antibodies in RA synovial fluid PAD activity was measured with a histone H3 citrullination assay using different RA synovial fluid samples: Figure 26A: Sample 1. Figure 26B: Sample 2. Figure 26C: Sample 3. Figure 26D: Sample 5. Figure 26E: Sample 4. DETAILED DESCRIPTION OF THE INVENTION

[0016] Any composition or method provided herein can be combined with one or more of any of the other compositions and methods provided herein.

[0017] 4.1 Array The antibodies or polypeptides of the invention may comprise an amino acid sequence as provided in Tables 1-58. The antibodies may comprise an amino acid sequence (VH, VL, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, LCDR3, full sequence, Fab, scFv, constant light chain, C) as provided in any of Tables 1-58. L ), heavy chain (HC), light chain (LC), CH1, CH2, CH3).

[0018] [Table 1]

[0019] [Table 2]

[0020] [Table 3]

[0021] [Table 4]

[0022] [Table 5]

[0023] Table 6

[0024] Table 7

[0025] Table 8

[0026] Table 9

[0027] Table 10

[0028] Table 11

[0029] Table 12

[0030] Table 13

[0031] Table 14

[0032] Table 15

[0033] Table 16

[0034] Table 17

[0035] Table 18

[0036] Table 19

[0037] Table 20

[0038] Table 21

[0039] Table 22

[0040] Table 23

[0041] Table 24

[0042] Table 25

[0043] Table 26

[0044] Table 27

[0045] Table 28

[0046] Table 29

[0047] Table 30

[0048] Table 31

[0049] Table 32

[0050] Table 33

[0051] Table 34

[0052] Table 35

[0053] Table 36

[0054] Table 37

[0055] Table 38

[0056] Table 39

[0057] Table 40

[0058] Table 41

[0059] Table 42

[0060] Table 43

[0061] Table 44

[0062] Table 45

[0063] Table 46

[0064] Table 47

[0065] Table 48

[0066] Table 49

[0067] Table 50

[0068] Table 51

[0069] Table 52

[0070] Table 53

[0071] Table 54

[0072] Table 55

[0073] Table 56

[0074] Table 57-1

[0075] Table 57-2

[0076] Table 57-3

[0077] Table 58-1

[0078] Table 58-2

[0079] Table 58-3

[0080] Table 58-4

[0081] Table 58-5

[0082] Table 59

[0083] Table 60

[0084] Table 61

[0085] [Table 62]

[0086] 4.2 Sequence identity

[0087] [Table 63-1]

[0088] [Table 63-2] The area defined by Kabat,

[0089] 4.3 Overview The antibody may comprise a PAD2-binding domain that specifically binds to PAD2 and / or a PAD4-binding domain that specifically binds to PAD4. The antibody may comprise a domain that specifically binds to PAD2. The antibody may comprise a domain that specifically binds to PAD4. The antibody may comprise a domain that specifically binds to PAD2 and a domain that specifically binds to PAD4. The antibody may inhibit PAD activity. The antibody may inhibit PAD-mediated citrullination of proteins. The antibody may inhibit PAD activity in synovial fluid. The antibody has an IC of ≦200 pM as measured by an H3 citrullination assay. 50 The antibody may be a human antibody.

[0090] Specific PAD2 binding can be measured by PAD2 ELISA.Specific PAD4 binding can be measured by PAD4 ELISA.

[0091] The antibody has an IC of about 700, 650, 600, 550, 540, 530, 520, 500, 480, 460, 440, 450, 430, 420, 400, 380, 360, 340, 320, 300, 280, 260, 240, 220, 200, 180, 160, 140, 120, 100, 90, 80, 70, 60, 50, 40, 30, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 4, 2, or 1 pM as measured by an H3 citrullination assay. 50 IC 50 may be related to the inhibition of PAD4 activity. 50 may be associated with the inhibition of PAD2 activity. 50 may be associated with the inhibition of the combined activity of PAD4 and PAD2.

[0092] The antibody may inhibit PAD2 activity. The antibody may inhibit PAD2-mediated protein citrullination. The antibody has an IC of ≦700 pM as measured by the H3 citrullination assay. 50 The antibody may have an IC50 of about 700, 650, 600, 550, 540, 530, 520, 500, 480, 460, 440, 450, 430, 420, 400, 380, 360, 340, 320, 300, 280, 260, 240, 220, 200, 180, 160, 140, 120, 100, 90, 80, 70, 60, 50, 40, 30, 20, 19, 18, 17, 16, 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 4, 2, or 1 pM as measured by an H3 citrullination assay. 50 IC 50 may be related to the inhibition of PAD4 activity. 50 may be associated with the inhibition of PAD2 activity. 50 may be associated with the inhibition of the combined activity of PAD4 and PAD2.

[0093] I C 50 can be measured by trypsin cleavage assay. 50can be measured by the BAEE (Nα-Benzonyl-L-arginine ethyl ester hydrochloride) citrullination assay. 50 can be measured by an H3 citrullination assay.

[0094] The antibody may inhibit PAD4 activity, and optionally, the antibody has an IC of ≦100 pM as measured by an H3 citrullination assay. 50 inhibits PAD4-mediated protein citrullination, and optionally, the PAD4 is recombinant PAD4.

[0095] The antibody may inhibit PAD2 in immune cells. The antibody may inhibit PAD4 in immune cells. The antibody may inhibit PAD2 and PAD4 in immune cells. The antibody may inhibit PAD activity in immune cells. The immune cells may be neutrophils. The immune cells may be monocytes. The immune cells may be neutrophils and monocytes.

[0096] The PAD2 inhibited by the antibody may be human, cynomolgus monkey, or mouse PAD2. The PAD4 inhibited by the antibody may be human, mouse PAD4, or cynomolgus monkey.

[0097] The antibody can include, for example, the sequence of clone 01, clone 02, clone 03, clone 04, clone 05, clone 06, clone 07, clone 08, clone 09, clone 10, clone 11, clone 12, clone 22, or clone 42, as provided in Table 1, Table 2, Table 57, and Table 58.

[0098] 4.4 Specificity The PAD2-binding domain may not specifically bind to PAD3 or PAD1. The PAD4-binding domain may not specifically bind to PAD3 or PAD1. PAD3 binding may be measured by PAD3 ELISA. PAD1 binding may be measured by PAD1 ELISA. The antibody may not specifically bind to PAD3 or PAD1. PAD3 may be human, cynomolgus monkey, or mouse PAD3. PAD1 may be human, cynomolgus monkey, or mouse PAD1. The antibody may specifically bind to PAD4 but not to PAD1, PAD2, or PAD3. The antibody may specifically bind to PAD2 but not to PAD1, PAD4, or PAD3. The antibody may specifically bind to PAD2 and PAD4 but not to PAD1 or PAD3.

[0099] The antibody may specifically bind to mouse PAD2. The antibody may specifically bind to mouse PAD4. The antibody may not specifically bind to PAD2. The antibody may not specifically bind to PAD4.

[0100] 4.5 Bispecific antibodies The antibody may be a bispecific antibody comprising a PAD2-binding domain and a PAD4-binding domain. The PAD2-binding domain may specifically bind to PAD2 but may not specifically bind to PAD1, PAD4, or PAD3. The PAD4-binding domain may specifically bind to PAD4 but may not specifically bind to PAD1, PAD3, or PAD2.

[0101] 4.6 PAD2 affinity Bispecific antibodies are antibodies that bind to human PAD2 by using the affinity (K D ) and affinity (K D ) to human PAD2. A bispecific antibody can bind to human PAD2 with the affinity (K D affinities (K D) to human PAD2. A bispecific antibody can bind to human PAD2 with the affinity (K D affinities (K D ) can bind to PAD2.

[0102] K of antibodies or bispecific antibodies against human PAD D is the K of bivalent Fab fragments of IgG containing the same PAD2-binding domain against human PAD2 D The antibody may have a K D The antibody may have an affinity (K) of about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 pM for human PAD2. D The antibody may have an affinity (K) for cynomolgus monkey PAD2 of about 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, and 160 pM. D The antibody may have an affinity (K) for mouse PAD2 of about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, or 40 pM. D The antibody may have an affinity (K) of ≦20 pM or ≦18 pM. D ) can bind to human PAD2. The affinity can be measured by surface plasmon resonance (SPR).

[0103] Bispecific antibodies are antibodies that bind to the affinity (K D ) and affinity (K D ) to cynomolgus monkey PAD2. A bispecific antibody can bind to cynomolgus monkey PAD2 with the affinity (KD affinities (K D ) to cynomolgus monkey PAD2. A bispecific antibody can bind to cynomolgus monkey PAD2 with the affinity (K D affinities (K D The KD of a bispecific antibody against cynomolgus monkey PAD2 is the KD of a bivalent Fab fragment of IgG containing the same PAD2-binding domain against cynomolgus monkey PAD2. D It may be smaller than

[0104] Bispecific antibodies are characterized by the affinity (K ) of bivalent Fab fragments of IgG containing the same PAD2-binding domain for mouse PAD2. D ) and affinity (K D ) to mouse PAD2. The bispecific antibody can bind to mouse PAD2 with the affinity (K D affinities (K D ) to mouse PAD2. The bispecific antibody can bind to mouse PAD2 with the affinity (K D affinities (K D ) can bind to PAD2. D is the K of bivalent Fab fragments of IgG containing the same PAD2-binding domain against mouse PAD2 D It may be smaller than

[0105] Affinity of antibodies to PAD2 (K D ) can be any of the affinities as provided in the Examples, particularly as provided in Tables 68 and 69. The affinity K of the antibody for PAD4 D may be within the affinity ranges as provided in the Examples, particularly as provided in Table 70 or Table 71.

[0106] Affinity (e.g., K D ) can be measured by surface plasmon resonance (SPR).

[0107] 4.7 PAD4 affinity Bispecific antibodies are characterized by the affinity (K ) of bivalent Fab fragments of IgG containing the same PAD4-binding domain for human PAD4. D ) and affinity (K D ) to human PAD4. A bispecific antibody can bind to human PAD4 with the affinity (K D affinities (K D ) to human PAD4. A bispecific antibody can bind to human PAD4 with the affinity (K D affinities (K D ) to human PAD4. A bispecific antibody can bind to PAD4 with the affinity (K D affinities (K D ) to human PAD4. A bispecific antibody can bind to PAD4 with the affinity (K D affinities (K D ) can bind to PAD4.

[0108] K of antibodies or bispecific antibodies against human PAD4 D is the K of bivalent Fab fragments of IgG containing the same PAD4-binding domain against human PAD4 D The antibody may have a K D The antibody may have an affinity of about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, or 53 pM for human PAD4.

[0109] The antibody has an affinity K of about 5-50 pM or 5-45 pM for mouse PAD4. D The antibody may have an affinity K for mouse PAD4 of about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, or 50. D may have:

[0110] Bispecific antibodies are antibodies that bind to the affinity (K D ) and affinity (K D The bispecific antibody can bind to cynomolgus monkey PAD4 with the affinity (K) of a bivalent Fab fragment of IgG containing the same PAD4-binding domain for cynomolgus monkey PAD4. D affinities (K D The bispecific antibody can bind to cynomolgus monkey PAD4 with the affinity (K) of a bivalent Fab fragment of IgG containing the same PAD4-binding domain for cynomolgus monkey PAD4. D affinities (K D ) can bind to PAD4. D is the K of bivalent IgG Fab fragments containing the same PAD4-binding domain against cynomolgus monkey PAD4 D It may be smaller than

[0111] Bispecific antibodies are characterized by the affinity (K ) of bivalent Fab fragments of IgG containing the same PAD2-binding domain for mouse PAD4. D ) and affinity (K D ) to mouse PAD4. A bispecific antibody can bind to mouse PAD4 with the affinity (K D affinities (K D) to mouse PAD4. The bispecific antibody can bind to mouse PAD2 with the affinity (K D affinities (K D ) can bind to PAD4. D is the K of bivalent Fab fragments of IgG containing the same PAD4-binding domain against mouse PAD4 D It may be smaller than

[0112] The antibodies have an affinity (K D ) can bind to human PAD4.

[0113] Affinity of antibody to PAD4, K D can be any of the affinities as provided in the Examples, in particular as provided in Table 70 or Table 71. The affinity K of the antibody for PAD4 D may be within the range of affinities provided in the Examples, particularly those provided in Table 70 or Table 71.

[0114] 4.8 Thermal stability Bispecific antibodies may have beneficial thermostability. The antibody or bispecific antibody may have a T of ≥ 40°C. onset The antibody or bispecific antibody may have a T of about 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, or 54°C. onset The antibody or bispecific antibody may have a T onset T onset can be measured by nano-differential scanning fluorimetry (DSF). onset ) is the thermal stability (T onset The thermal stability (T onset ) is the thermal stability (T onsetThe thermal stability (T onset ) is the thermal stability (T onset The thermal stability (T onset ) is the thermal stability (T onset The thermal stability (T onset ) is the thermal stability (T onset The thermal stability (T onset ) is the thermal stability (T onset ) can be within ±20°C.

[0115] Antibody thermal stability (T onset ) can be any of the values ​​provided in Table 91. Antibody thermostability (T onset ) may be within the range of values ​​provided in Table 91.

[0116] 4.9 Agglomeration The antibody may have a beneficially low risk of aggregation. The aggregation tendency of the bispecific antibody may be no more than two-fold greater than the aggregation tendency of a bivalent IgG Fab comprising the same PAD2-binding domain. The aggregation tendency of the bispecific antibody may be no more than two-fold greater than the aggregation tendency of a bivalent IgG Fab comprising the same PAD4-binding domain. The aggregation of the antibody at 40°C may be no more than two-fold greater than the aggregation of a bivalent IgG1 antibody comprising the same PAD2-binding domain or PAD4-binding domain. The aggregation of the antibody at 40°C may be no more than two-fold greater than the aggregation of a bivalent IgG1 antibody comprising the same PAD2-binding domain or PAD4-binding domain.

[0117] 4.10 Bis3 The bispecific antibody may be a bivalent bispecific antibody. The bispecific antibody may be in Bis3 format, i.e., have an scFv and an IgG binding domain (FIG. 10B). The antibody may comprise two PAD2-binding domains, such that the antibody is bivalent for PAD2, and two PAD4-binding domains, such that the antibody is also bivalent for PAD4. The PAD2 and PAD4 may be human, cynomolgus monkey, and / or mouse PAD2 or PAD4. The antibody may comprise two scFv domains, each comprising a PAD2-binding domain according to the present invention. The antibody may comprise two scFv domains, each comprising a PAD4-binding domain according to the present invention.

[0118] The antibody may comprise two Fab domains, each comprising a PAD2-binding domain according to the present invention. The antibody may comprise two Fab domains, each comprising a PAD4-binding domain according to the present invention. The antibody may be a Bis3 bispecific antibody having any of the structures shown in Figure 10B.

[0119] Bis3 bispecific antibodies have an affinity (K D ) wherein each scFv domain comprises a PAD4-binding domain according to the present invention and each Fab domain comprises a PAD2-binding domain according to the present invention. The Bis3 bispecific antibody can bind to human PAD4 with an affinity (K D ) can bind to human PAD4, wherein each scFv domain comprises a PAD2-binding domain according to the present invention, and each Fab domain comprises a PAD4-binding domain according to the present invention.

[0120] Bis3 bispecific antibodies have an affinity (K D ) wherein each scFv domain comprises a PAD4-binding domain according to the present invention and each Fab domain comprises a PAD2-binding domain according to the present invention. The Bis3 bispecific antibody can bind to human PAD2 with an affinity (K D) wherein each scFv domain comprises a PAD2-binding domain according to the invention and each Fab domain comprises a PAD4-binding domain according to the invention. The antibody has an affinity (K D The antibody can bind to human PAD2 with an affinity (K D ) can bind to human PAD4.

[0121] The Bis3 bispecific antibodies have an affinity (K) for human or cynomolgus PAD2 as provided in Table 72 or Table 73. D Bis3 may have an affinity for PAD4 (K) as provided in Table 75 or Table 74. D ).

[0122] The antibody may be a bispecific antibody comprising: (a) an IgG comprising first and second Fab domains and an Fc domain, wherein the first and second Fab domains each comprise a PAD2-binding domain that specifically binds to PAD2; and (b) a first and second scFv, wherein the first and second scFvs are each linked to the carboxy terminus of one of the heavy chains of the IgG Fc domain, and the first and second scFvs each comprise a PAD4-binding domain that specifically binds to PAD4. The Fc domain may be an IgG or IgG1 Fc domain. The PAD4-binding domain of the first and second scFvs may comprise SEQ ID NO: 39. The PAD2-binding domain of the first and second Fabs may comprise a heavy chain domain comprising SEQ ID NO: 34. The PAD2-binding domain of the first and second Fabs may comprise a light chain constant domain comprising SEQ ID NO: 36. The PAD2-binding domain of the first and second Fabs comprises a light chain domain comprising SEQ ID NO: 35.

[0123] The antibody may be a bispecific antibody comprising: (a) an IgG comprising first and second Fab domains and an Fc domain, wherein the first and second Fab domains each comprise a PAD4-binding domain that specifically binds to PAD4; and (b) a first and second scFv, wherein the first and second scFvs are each linked to the carboxy terminus of one of the heavy chains of the IgG Fc domain, and the first and second scFvs each comprise a PAD2-binding domain that specifically binds to PAD2. The first and second scFv PAD2-binding domains may comprise SEQ ID NO: 38. The first and second Fab PAD4-binding domains comprise a heavy chain domain comprising SEQ ID NO: 40. The first and second Fab PAD4-binding domains comprise a light chain domain comprising SEQ ID NO: 41. The first and second Fab domains comprise heavy chain constant domains which may comprise SEQ ID NO: 37. The first and second Fab domains comprise light chain constant domains comprising SEQ ID NO: 36.

[0124] The scFvs may be linked to the carboxy termini of the heavy chains by a peptide linker. The peptide linker may comprise SEQ ID NO: 51. The first and / or second scFv of the Bis3 bispecific antibody may comprise a VH-VL linker domain comprising SEQ ID NO: 33.

[0125] The Bis3 bispecific antibody may comprise the sequence of SEQ ID NO: 53, SEQ ID NO: 54, SEQ ID NO: 55, SEQ ID NO: 56, SEQ ID NO: 57, or SEQ ID NO: 58. The Bis3 bispecific antibody may comprise a sequence having 90% sequence identity to SEQ ID NO: 53, SEQ ID NO: 54, SEQ ID NO: 55, SEQ ID NO: 56, SEQ ID NO: 57, or SEQ ID NO: 58.

[0126] The Bis3 bispecific antibody may have the sequence of any of clones 07, 08, 09, 10, 11 or 12. The Bis3 bispecific antibody may have a sequence that is at least 90% identical to the sequence of clones 07, 08, 09, 10, 11 or 12. The Bis3 bispecific antibody may have a sequence as provided in Table 57.

[0127] 4.11 DuetMab The antibody can be a monovalent bispecific antibody. The antibody can be a DuetMab. The antigen can be a DuetMab comprising any of the structures shown in Figure 10A.

[0128] The antibody may comprise one PAD2-binding domain such that the antibody is monovalent for PAD2, or one PAD4-binding domain such that the antibody is monovalent for PAD4. The antibody may comprise an IgG comprising a first binding region comprising a first Fab, wherein the second Fab domain comprises the PAD2-binding domain, and a second binding region comprising a second Fab, wherein the second Fab comprises the PAD4-binding domain. The antibody may comprise an IgG domain with a knob / hole mutation. The IgG may comprise a kappa light chain comprising SEQ ID NO: 63. The antibody may comprise a lambda light chain comprising SEQ ID NO: 64. The antibody may comprise a kappa light chain comprising SEQ ID NO: 62. The antibody may comprise a lambda light chain comprising SEQ ID NO: 65.

[0129] A DuetMab bispecific antibody may comprise a PAD2-binding domain comprising a heavy chain comprising SEQ ID NO:73 and a light chain comprising SEQ ID NO:62, and the PAD4-binding region may comprise a heavy chain comprising SEQ ID NO:76 and a light chain comprising SEQ ID NO:65. A DuetMab bispecific antibody may comprise a PAD2-binding domain comprising a heavy chain comprising SEQ ID NO:74 and a light chain comprising SEQ ID NO:62, and a PAD4-binding region comprising a heavy chain comprising SEQ ID NO:76 and a light chain comprising SEQ ID NO:65. A DuetMab bispecific antibody may comprise a PAD4-binding region comprising a heavy chain comprising SEQ ID NO:59 or SEQ ID NO:60 and a light chain comprising SEQ ID NO:62, and a heavy chain comprising SEQ ID NO:61 and a light chain comprising SEQ ID NO:65. A DuetMab may comprise a PAD2-binding domain comprising a heavy chain comprising SEQ ID NO:70 and a light chain comprising SEQ ID NO:64, and a PAD4-binding region comprising a heavy chain comprising SEQ ID NO:67 and a light chain comprising SEQ ID NO:63. The antibody may comprise a PAD2-binding domain comprising a heavy chain comprising SEQ ID NO: 71 and a light chain comprising SEQ ID NO: 64, and a PAD4-binding region comprising a heavy chain comprising SEQ ID NO: 68 and a light chain comprising SEQ ID NO: 63. The DuetMab may comprise a PAD2-binding domain comprising a heavy chain comprising SEQ ID NO: 72 and a light chain comprising SEQ ID NO: 64, and a PAD4-binding region comprising a heavy chain comprising SEQ ID NO: 69 and a light chain comprising SEQ ID NO: 63.

[0130] DuetMab may have the sequence of any of clones 01, 02, 03, 04, 05, or 06. DuetMab bispecific antibodies may have a sequence that is at least 90% identical to the sequence of 01, 02, 03, 04, 05, or 06. Bis3 bispecific antibodies may have a sequence as provided in Table 58.

[0131] 4.12 IgG The antibody may comprise an IgG or F(ab')2 fragment. The antibody may comprise an IgG or F(ab')2 fragment that is bivalent for PAD2 or bivalent for PAD4. The antibody may be an IgG1 that is bivalent for PAD2 or PAD4. The PAD2 or PAD4 may be human, cynomolgus monkey, and / or mouse PAD2 or PAD4. The IgG or F(ab')2 fragment may comprise a PAD2-binding domain according to the present invention. The IgG or F(ab')2 fragment may comprise a PAD4-binding domain according to the present invention. The antibody may comprise two PAD2-binding domains of the present invention such that the antibody is bivalent for PAD2. The antibody may comprise two PAD4-binding domains of the present invention such that the antibody is bivalent for PAD4. The antibody may comprise two PAD4-binding domains of the present invention and not a PAD2-binding domain according to the present invention. The antibody may comprise two PAD2-binding domains of the present invention and not a PAD4-binding domain according to the present invention.

[0132] IgG may have a heavy chain with a terminal lysine. IgG may have two heavy chains with terminal lysine. IgG may have a heavy chain with a terminal lysine and a heavy chain without a terminal lysine. IgG may have two heavy chains without a terminal lysine.

[0133] 4.13 Fab The antibody can comprise a Fab fragment, wherein the Fab fragment comprises a PAD2-binding domain or a PAD4-binding domain. The Fab fragment can comprise a PAD2-binding domain, wherein the Fab has an affinity (K) of ≦20 nM, ≦10 nM, ≦6 nM, or ≦1 nM. D The antibody can comprise a Fab fragment, wherein the Fab fragment comprises a PAD4-binding domain, and the Fab binds to human PAD2 with an affinity (K) of ≦1 nM, ≦0.1 pM, ≦0.07 pM, or ≦0.05 pM. D ) binds to human PAD4. D can be measured by surface plasmon resonance (SPR).

[0134] 4.14 Variable Region The variable region of the antibody may be a human variable region. The variable region may comprise rodent or mouse complementarity determining regions (CDRs) and human framework regions (FRs). The variable region may be a primate (e.g., non-human primate) variable region. The variable region may comprise rodent or mouse CDRs and primate (e.g., non-human primate) framework regions (FRs). The variable region may comprise the CDR, VH, VL, or framework region of any of the antibodies listed in Tables 1 to 54 and 62.

[0135] 4.14.1 PAD2 lead The antibody may comprise a PAD2-binding domain, which comprises a variable heavy chain (VH) domain sequence comprising CDRs HCDR1, HCDR2, and HCDR3, and a variable light chain (VL) domain sequence comprising CDRs LCDR1, LCDR2, and LCDR3, wherein the HCDR1 amino acid sequence is SEQ ID NO: 3, the HCDR2 amino acid sequence is SEQ ID NO: 4, the HCDR3 amino acid sequence is SEQ ID NO: 5, the LCDR1 amino acid sequence is SEQ ID NO: 10, the LCDR2 amino acid sequence is SEQ ID NO: 11, and / or the LCDR3 amino acid sequence is SEQ ID NO: 12. The antibody may comprise a PAD2-binding domain, which comprises a VH domain comprising a sequence having at least 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identity to SEQ ID NO: 1. The PAD2-binding domain may comprise a VH domain comprising SEQ ID NO: 1. The PAD2-binding domain may comprise a VL domain comprising a sequence having at least 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identity to SEQ ID NO: 2. The PAD2-binding domain may comprise a VL domain sequence comprising SEQ ID NO: 2. The PAD2-binding domain may comprise a VH domain sequence comprising SEQ ID NO: 1, optionally with 1, 2, 3, 4, or 5 amino acid changes outside the CDRs. The PAD2-binding domain may comprise a VL domain sequence comprising SEQ ID NO: 2, optionally with 1, 2, 3, 4, or 5 amino acid changes outside the CDRs. The antibody may comprise the VH, VL, CDR, and framework region sequences of an antibody listed in Table 1. The antibody may be an affinity-optimized antibody of an antibody listed in Table 53 or Table 54.

[0136] 4.14.2 PAD4 lead The bispecific antibody may comprise a PAD4-binding domain comprising a variable heavy chain (VH) domain sequence comprising complementarity-determining regions (CDRs) HCDR1, HCDR2, and HCDR3, and a variable light chain (VL) domain sequence comprising CDRs LCDR1, LCDR2, and LCDR3, wherein the HCDR1 amino acid sequence is SEQ ID NO: 17, the HCDR2 amino acid sequence is SEQ ID NO: 18, the HCDR3 amino acid sequence is SEQ ID NO: 19, the LCDR1 amino acid sequence is SEQ ID NO: 24, the LCDR2 amino acid sequence is SEQ ID NO: 25, and / or the LCDR3 amino acid sequence is SEQ ID NO: 26. The PAD4-binding domain may comprise a VH domain comprising a sequence having at least 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identity to SEQ ID NO: 31. The PAD4-binding domain may comprise a VH domain sequence comprising SEQ ID NO: 31. The PAD4-binding domain may comprise a VL domain comprising a sequence having at least 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% identity to SEQ ID NO: 32. The PAD4-binding domain may comprise a VL sequence domain comprising SEQ ID NO: 32. The PAD4-binding domain may comprise a VH domain sequence comprising SEQ ID NO: 31, optionally with 1, 2, 3, 4, or 5 amino acid changes outside the CDRs. The PAD4-binding domain may comprise a VL domain sequence comprising SEQ ID NO: 32, optionally with 1, 2, 3, 4, or 5 amino acid changes outside the CDRs. The antibody may comprise the VH, VL, CDR, and framework region sequences of an antibody listed in Table 2. The antibody may be an affinity-optimized antibody of an antibody listed in Table 62.

[0137] 4.14.3 PAD4 Backup Clone The antibody or bispecific antibody may comprise any of the antibody sequences provided in Tables 3-42. The antibody may comprise the CDRs, framework regions, VH, or VL sequences of clone 42, 141LO0035 hIgG1 ngl-2, 141LO0035 hIgG1 pgl-4, 141LO0055 hIgG1 ngl-2, 141LO0030 hIgG1 ngl-2, 141LO0039 hIgG1 ngl-2, 141LO0030 hIgG1 pgl-4, 141LO0002 hIgG1 pgl-4, 141LO0002 hIgG1 pgl-3, 141LO0002 hIgG1 ngl-2, 141LO0039 hIgG1 pgl-4, PAD40175 hIgG1 ngl-2, PAD40119 hIgG1 ngl-2, PAD40141 hIgG1 ngl-2. The antibody may have a sequence that has 90% sequence identity to a VH sequence provided in any of Tables 3-42. A sequence that has 90% sequence identity to a VL sequence provided in any of Tables 3-42. The antibody may have a VH, VL, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, LCDR3 that have sequences identical to the corresponding regions of clone 42 as provided in Table 63.

[0138] 4.14.4 PAD2 Backup Clone The antibody or bispecific antibody may comprise any of the antibody sequences provided in Tables 43 to 54. The antibody may comprise the CDR, framework regions, VH, or VL sequences of clone 22: 141LO0002 hIgG1 pgl-3, 141LO0002 hIgG1 pgl-4, 141LO0002 hIgG1 ngl-2, 141LO0030 hIgG1 pgl-4, 141LO0002 hIgG1 ngl-2, 141LO0035 hIgG1 ngl-2, 141LO0039 hIgG1 pgl-4, 141LO0039 hIgG1 ngl-2, 141LO0055 hIgG1 ngl-2, 141LO0055 hIgG1 ngl-2. The antibody may have a sequence that has 90% sequence identity to a VH sequence provided in any of Tables 43-54. A sequence that has 90% sequence identity to a VL sequence provided in any of Tables 43-54. The antibody may have a VH, VL, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, LCDR3 with sequences identical to the corresponding regions of clone 42 as provided in Table 63.

[0139] 4.15 Fc modification The antibody or bispecific antibody may comprise an Fc domain, optionally an IgG1 Fc domain. The Fc domain may have null effector function. The Fc domain may comprise the mutations L234F, L235Q, and / or K322Q as numbered by the EU index as described in Kabat et al.

[14] . The Fc mutations may include at least one half-life-extending mutation. The Fc domain may comprise the mutations M252Y, S254T, and T256E as numbered by the EU index as described in Kabat et al.

[14] . The Fc domain may comprise at least one null effector function mutation and at least one half-life-extending mutation. The Fc domain may comprise the mutations L234F, L235Q, K322Q, M252Y, S254T, and T256E as numbered by the EU index as described in Kabat et al.

[14] . The Fc domain may contain the mutations L234F, L235E, P331S, M252Y, S254T, and T256E as numbered by the EU index as described in Kabat et al.

[14] . The Fc domain may contain the mutations C47, C48, or C49. H The Fc domain may have any of the mutations listed in Table 65.

[0140] The polypeptide according to the invention may comprise an Fc variant domain.

[0141] IgG Fc domains with extended half-lives are described in WO 2015 / 175874(A2)

[15] and WO 2002 / 060919(A2)

[16] . YTE increases binding to FcRn, resulting in extended serum half-life. YTE is a C H The triple mutation in 2 is: M252Y, S254T and T256E.

[0142] The modified Fc region may comprise amino acid substitutions at two or more of positions 432-437, numbered according to the EU numbering index of Kabat, relative to a human wild-type Fc region, wherein (i) positions 432 and 437 are each substituted with cysteine, (ii) position 433 is histidine or substituted with arginine, proline, threonine, lysine, serine, alanine, methionine, or asparagine, and (iii) position 434 is asparagine or substituted with arginine, tryptophan, or cysteine. (iv) position 435 is histidine or is substituted with histidine; and (v) position 436 is tyrosine or phenylalanine or is substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine, or threonine, and the modified human IgG1 has an increased half-life compared to the half-life of an IgG1 having a human wild-type Fc region.

[0143] The modified Fc region may comprise an amino acid substitution relative to a wild-type Fc region at two or more of positions 432 to 437, numbered according to the EU numbering index of Kabat; a) at least one of positions 432 and 437 is substituted with a cysteine; or b) at least one of positions 432 and 437 is substituted with an amino acid selected from the group consisting of glutamine, glutamic acid, aspartic acid, lysine, arginine, and histidine; The polypeptide has an altered half-life compared to the half-life of an IgG having a wild-type Fc region. Optionally, (i) both positions 432 and 437 are substituted with cysteine, or (ii) both positions 432 and 437 are substituted with an amino acid independently selected from the group consisting of glutamine, glutamic acid, aspartic acid, lysine, arginine, and histidine.

[0144] The polypeptide may comprise an amino acid insertion after position 437, optionally, the amino acid insertion is glutamic acid.

[0145] The binding affinity of the polypeptide to FcRn at pH 6.0 may be higher than the binding affinity of an IgG having a wild-type Fc region to FcRn at pH 6. The binding affinity of the polypeptide to FcRn at pH 7.4 may be higher than the binding affinity of an IgG having a wild-type Fc region to FcRn at pH 7.4. The KD of the polypeptide for FcRn at pH 6.0 may be less than 500 nM, and the KD at pH 7.4 is at least 1000 nM.

[0146] The polypeptide may comprise an Fc variant domain that exhibits a pH-dependent increase in binding affinity to FcRn compared to an IgG having a wild-type Fc region. The polypeptide may comprise an Fc variant domain, wherein the modified IgG Fc domain exhibits a pH-dependent decrease in binding affinity to FcRn compared to an IgG having a wild-type Fc region.

[0147] The polypeptide may comprise an Fc variant domain, wherein the modified IgG Fc domain retains a wild-type level of at least one attribute selected from the group consisting of (i) binding to at least one Fc gamma receptor, (ii) binding to Clq, or (iii) an effector function, where optionally the Fc gamma receptor is selected from the group consisting of FcγRI receptor, FcγRII receptor, and FcγRIII receptor. The polypeptide may have a reduced effector function selected from antibody-dependent cellular cytotoxicity (ADCC), complement-dependent cytotoxicity (CDC), and / or antibody-dependent cellular phagocytosis (ADCP). The polypeptide may comprise an Fc variant domain, wherein the Fc variant domain has amino acid substitutions at three or more of positions 432, 433, 434, 435, 436, or 437.

[0148] The polypeptide may comprise an Fc variant domain, the Fc variant domain having four or more amino acid substitutions at positions 432, 433, 434, 435, 436, or 437. Positions 432 and 437 may each be substituted with cysteine, position 433 may be histidine or substituted with arginine, proline, threonine, lysine, serine, alanine, methionine, or asparagine, position 434 may be asparagine or substituted with arginine, tryptophan, histidine, phenylalanine, tyrosine, serine, methionine, or threonine, position 435 may be histidine or substituted with histidine, and position 436 may be tyrosine or phenylalanine or substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine, or threonine. Position 433 can be histidine. Position 433 can be substituted with arginine, asparagine, proline, threonine, or lysine. Position 434 can be substituted with arginine, tryptophan, histidine, phenylalanine, or tyrosine. Position 434 can be substituted with arginine. Position 436 can be substituted with leucine, arginine, isoleucine, lysine, methionine, valine, or histidine. Position 433 can be histidine or can be substituted with arginine, asparagine, proline, threonine, or lysine, position 434 can be substituted with arginine, tryptophan, histidine, phenylalanine, or tyrosine, and position 436 can be substituted with leucine, arginine, isoleucine, lysine, methionine, valine, or histidine.

[0149] The polypeptide may comprise an Fc variant domain, which may comprise the amino acid sequence of positions 432-437 of CXRHXC (SEQ ID NO: 187), where position 433 is histidine or substituted with arginine, asparagine, proline, or serine, and position 436 is substituted with arginine, leucine, isoleucine, methionine, or serine. The polypeptide may comprise the amino acid sequence of positions 432-437 of CRRHXC (SEQ ID NO: 188), where position 436 is substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine, or threonine. Position 436 may also be substituted with leucine, isoleucine, serine, or threonine.

[0150] The polypeptide may comprise an Fc variant domain, which may comprise the amino acid sequence of positions 432-437 of CXRHRC (SEQ ID NO: 189), wherein position 433 is arginine, proline, threonine, lysine, serine, alanine, methionine, or asparagine. The modified IgG may comprise the amino acid sequence of positions 432-437 of ZXXHXZ (SEQ ID NO: 92), wherein the amino acid at position 432 is substituted with glutamic acid, glutamine, histidine, or aspartic acid; position 433 is histidine or substituted with arginine, alanine, lysine, threonine, leucine, proline, serine, or glutamine; position 434 is tyrosine, phenylalanine, histidine, serine, or tryptophan; position 436 is tyrosine or substituted with arginine, histidine, asparagine, lysine, leucine, methionine, threonine, or valine; and position 437 is substituted with glutamine, histidine, glutamic acid, or aspartic acid.

[0151] The Fc variant domain may include N3, YC37-YTE, YC56-YTE, YC59-YTE, Y3-YTE, Y31-YTE, Y12-YTE, Y83-YTE, Y37-YTE, and Y9-YTE, N3-YTE, N3E-YTE, SerN3-YTE, Y54-YTE, Y74-YTE, and Y8-YTE. The Fc variant domain may have a histidine at amino acid position 435. The modified IgG Fc domain may include the amino acid sequence E(R / A)(W / S / F)HRQ (SEQ ID NO: 190) at positions 432-437.

[0152] The polypeptide of claim 38, wherein the polypeptide may comprise at least one FcRn-binding portion of an Fc region of an IgG molecule, the FcRn-binding portion comprising, relative to a wild-type FcRn-binding portion, amino acid substitutions at two or more of positions 432 to 437, numbered according to the EU numbering index of Kabat, wherein at least one of positions 432 and 437 is substituted with cysteine, or (ii) at least one of positions 432 and 437 is substituted with an amino acid selected from the group consisting of glutamine, glutamic acid, aspartic acid, and histidine. Positions 432 and 437 may each be substituted with cysteine; position 433 may be histidine or may be substituted with arginine, proline, threonine, lysine, serine, alanine, methionine, or asparagine; position 434 may be asparagine or may be substituted with arginine, tryptophan, histidine, phenylalanine, tyrosine, serine, methionine, or threonine; position 435 may be histidine or may be substituted with histidine; and position 436 may be tyrosine or phenylalanine or may be substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine, or threonine.

[0153] The Fc variant domain can comprise the amino acid sequence at positions 432-437 of ZXXHXZ, where position 432 is substituted with glutamic acid, glutamine, histidine, or aspartic acid, position 433 is histidine or substituted with arginine, alanine, lysine, threonine, leucine, proline, serine, or glutamine, position 434 is substituted with tyrosine, phenylalanine, histidine, serine, or tryptophan, position 436 is tyrosine or substituted with arginine, histidine, asparagine, lysine, leucine, methionine, threonine, or valine, and position 437 is substituted with glutamine, histidine, glutamic acid, or aspartic acid. The Fc variant domain can comprise the amino acid sequence at positions 432-437 of E(R / A)(W / S / F)HRQ. Additionally, there may be an amino acid insertion after position 437, where the amino acid insertion is glutamic acid. The FcRn-binding portion of the Fc region may comprise about amino acid residues 231 to 446 of an IgG molecule, according to the EU numbering index of Kabat. The FcRn-binding portion of the Fc region may comprise about amino acid residues 216 to 446 of an IgG molecule, according to the EU numbering index of Kabat.

[0154] Variant IgG Fc domains with reduced effector function and extended half-life are described in WO 2013 / 165690(A1)

[17] . The variant IgG Fc domains include: a) a phenylalanine (F) amino acid at position 234; b) an alanine (A), asparagine (N), phenylalanine (F), glutamine (Q), or valine (V) amino acid at position 235; c) may contain an alanine (A), aspartic acid (D), glutamic acid (E), histidine (H), asparagine (N), or glutamine (Q) amino acid at position 322, or an alanine (A) or glycine (G) amino acid at position 331, wherein the numbering of amino acids is according to the EU index as in Kabat.

[0155] The Fc variant domain may comprise a phenylalanine (F) amino acid at position 234, a glutamine (Q) amino acid at position 235, and a glutamine (Q) amino acid at position 322, where amino acid numbering is according to the EU index as in Kabat.

[0156] The Fc variant domain may comprise a phenylalanine (F) amino acid at position 234, a glutamine (Q) amino acid at position 235, and a glycine (G) amino acid at position 331, where amino acid numbering is according to the EU index as in Kabat. The Fc variant domain may comprise a phenylalanine (F) amino acid at position 234, an alanine (A) amino acid at position 235, and a glutamine (Q) amino acid at position 322, where amino acid numbering is according to the EU index as in Kabat. The Fc variant domain may comprise a) a tyrosine (Y) amino acid at position 252, or a serine (S) amino acid at position 252, or a tryptophan (W) amino acid at position 252, or a threonine (T) amino acid at position 252, and / or b) a threonine (T) amino acid at position 254, and / or c) may comprise a glutamic acid (E) amino acid at position 256, or a serine (S) amino acid at position 256, or an arginine (R) amino acid at position 256, or a glutamine (Q) amino acid at position 256, or an aspartic acid (D) amino acid at position 256; Amino acid numbering follows the EU index as in Kabat.

[0157] The Fc variant domain is a) a tyrosine (Y) amino acid at position 252, and / or b) a threonine (T) amino acid at position 254, and / or c) may contain a glutamic acid (E) amino acid at position 256, where the amino acid numbering is according to the EU index as in Kabat.

[0158] The Fc variant domain is a) a tyrosine (Y) amino acid at position 252, or a serine (S) amino acid at position 252, or a tryptophan (W) amino acid at position 252, or a threonine (T) amino acid at position 252, and b) may contain a threonine (T) amino acid at position 254; Amino acid numbering follows the EU index as in Kabat.

[0159] The Fc variant domain is a) a threonine (T) amino acid at position 254, and b) may comprise a glutamic acid (E) amino acid at position 256, or a serine (S) amino acid at position 256, or an arginine (R) amino acid at position 256, or a glutamine (Q) amino acid at position 256, or an aspartic acid (D) amino acid at position 256; Amino acid numbering follows the EU index as in Kabat.

[0160] The Fc variant domain is a) a tyrosine (Y) amino acid at position 252, or a serine (S) amino acid at position 252, or a tryptophan (W) amino acid at position 252, or a threonine (T) amino acid at position 252, and b) may comprise a glutamic acid (E) amino acid at position 256, or a serine (S) amino acid at position 256, or an arginine (R) amino acid at position 256, or a glutamine (Q) amino acid at position 256, or an aspartic acid (D) amino acid at position 256; Amino acid numbering follows the EU index as in Kabat.

[0161] The Fc variant domain is a) a tyrosine (Y) amino acid at position 252 and a threonine (T) amino acid at position 254, or b) a threonine (T) amino acid at position 254 and a glutamic acid (E) amino acid at position 256, or c) may contain a tyrosine (Y) amino acid at position 252 and a glutamic acid (E) amino acid at position 256; Amino acid numbering follows the EU index as in Kabat.

[0162] The Fc variant domain may comprise a tyrosine (Y) amino acid at position 252, a threonine (T) amino acid at position 254, and a glutamic acid (E) amino acid at position 256, where amino acid numbering is according to the EU index as in Kabat.

[0163] The Fc variant domain is a) a phenylalanine (F) amino acid at position 234; b) a glutamine (Q) amino acid at position 235; c) a glutamine (Q) amino acid at position 322; d) a tyrosine (Y) amino acid at position 252; e) a threonine (T) amino acid at position 254; f) may contain a glutamic acid (E) amino acid at position 256; Amino acid numbering follows the EU index as in Kabat.

[0164] The Fc variant domain is a) a phenylalanine (F) amino acid at position 234; b) a glutamine (Q) amino acid at position 235; c) a glycine (G) amino acid at position 331; d) a tyrosine (Y) amino acid at position 252; e) a threonine (T) amino acid at position 254; f) may contain a glutamic acid (E) amino acid at position 256; Amino acid numbering follows the EU index as in Kabat.

[0165] The polypeptide may comprise a modified Fc variant domain, wherein the polypeptide has improved pharmacokinetic (PK) properties, optionally half-life, when compared to the same polypeptide comprising a wild-type Fc domain. The polypeptide may have improved FcRn binding when compared to the same polypeptide comprising a wild-type Fc domain.

[0166] The polypeptide may comprise an IgG Fc domain selected from the group consisting of a human immunoglobulin G class 1 (IgG1) Fc domain, a human immunoglobulin G class 2 (IgG2) Fc domain, a human immunoglobulin G class 3 (IgG) Fc domain, and a human immunoglobulin G class 4 (IgG4) Fc domain.

[0167] The polypeptide may comprise a modified Fc variant domain, and the polypeptide has reduced Fc-mediated effector function when compared to the same polypeptide comprising a wild-type Fc domain. The effector function may be antibody-dependent cell-mediated cytotoxicity (ADCC) and / or complement-dependent cytotoxicity (CDC). The polypeptide may have lower affinity for an Fc gamma receptor (FcγR) when compared to the same polypeptide comprising a wild-type Fc domain, and optionally the FcγR is a human FcγR. The FcγR may be FcγRI, FcγRII, FcγRIII, FcγRII, FcγRIa, FcγRIIa, FcγRIIb, FcγRIII(158V), or FcγRIII(158F).

[0168] The polypeptide may comprise a modified Fc variant domain, wherein the polypeptide binds with improved affinity to FcRn when compared to the same polypeptide comprising a wild-type Fc domain, and optionally the polypeptide has a higher affinity for FcRn at pH 6.0 than at pH 7.4.

[0169] The polypeptide may comprise a modified Fc variant domain, wherein the polypeptide binds with reduced affinity to Clq when compared to the same polypeptide comprising a wild-type Fc domain.

[0170] The polypeptide may exhibit increased thermal stability when compared to the same polypeptide comprising a FES-YTE IgG Fc domain, optionally where the thermal stability is measured by differential scanning calorimetry (DSC), and optionally the increase in thermal stability is at least 4°C.

[0171] The polypeptide may exhibit increased thermal stability when compared to the same polypeptide comprising a FES-YTE IgG Fc domain, where the thermal stability is measured by differential scanning fluorimetry (DSF), optionally the DSF fluorescent probe is Sypro Orange, and optionally the increase in thermal stability is at least 5°C.

[0172] The polypeptide may exhibit increased apparent solubility as measured using a polyethylene glycol (PEG) precipitation assay when compared to the same polypeptide comprising a FES-YTE IgG Fc domain.

[0173] The polypeptide may exhibit increased stability as measured using an accelerated stability assay when compared to the same polypeptide comprising a FES-YTE IgG Fc domain. The accelerated stability assay involves (i) incubation of the polypeptide for an extended period of time and (ii) incubation at an elevated temperature. The accelerated stability assay may be performed by incubation at a higher concentration, optionally where the extended period is at least one month, optionally where the elevated concentration is at least 25 mg / ml, and optionally where the elevated temperature is at least 40°C. The accelerated stability assay may be performed using High Performance Size Exclusion Chromatography (HPSEC) or Dynamic Light Scattering (DLS).

[0174] The Fc may comprise an RF double mutation.

[0175] The Fc may contain knob-in-hole mutations.

[0176] 4.16 Polypeptides The present invention also relates to polypeptides comprising the antibodies or bispecific antibodies of the invention. The present invention also provides polypeptides comprising one or more binding domains of an antibody defined anywhere herein. The polypeptide may comprise part or all of a PAD2-binding domain. The polypeptide may comprise part or all of a PAD4-binding domain. The polypeptide may comprise one or more CDRs as defined herein, or a binding domain such as a variable light chain or variable heavy chain domain as defined herein. The polypeptide may comprise a binding domain comprising all three CDRs (CDR1, CDR2, and CDR3) of a variable heavy chain domain sequence as defined herein. The polypeptide may comprise a binding domain comprising all three CDRs (CDR1, CDR2, and CDR3) of a variable light chain domain sequence as defined herein. The polypeptide may comprise the variable heavy chain domain of an antibody as defined herein. The polypeptide may comprise the variable light chain domain of an antibody as defined herein. The polypeptide may comprise the complete heavy chain of an antibody as defined herein. The polypeptide may comprise the complete light chain of an antibody as defined herein. The polypeptide may be an isolated polypeptide.

[0177] 4.17 Nucleic acids The present invention also relates to nucleic acids encoding one or more chains of an antibody or bispecific antibody of the invention. The present invention also relates to nucleic acids encoding a polypeptide according to the invention. The present invention also relates to vectors comprising the nucleic acids, and host cells comprising the vectors.

[0178] 4.18 Pharmaceutical Compositions The present invention also relates to pharmaceutical compositions comprising an antibody or bispecific antibody of the invention and a pharmaceutically acceptable carrier.

[0179] 4.19 Kit The present invention also relates to a kit comprising an antibody or bispecific antibody or pharmaceutical composition of the present invention. The kit may include instructions for use.

[0180] 4.20 Treatment Methods The present invention also relates to a method of treating a disease in a subject, comprising administering an antibody or pharmaceutical composition according to the present invention. The subject may have an autoimmune disease. The subject may have rheumatoid arthritis (RA). The subject may have elevated levels of PAD in synovial fluid, whole blood, or serum compared to a healthy subject. The subject may have elevated levels of PAD2 in synovial fluid, whole blood, or serum compared to a healthy subject. The subject may have elevated levels of PAD4 in synovial fluid, whole blood, or serum compared to a healthy subject. The concentration of PAD4 in the synovial fluid of the subject may be at least 200 ng / ml. The concentration of PAD2 in the synovial fluid of the subject may be at least 20 ng / ml. The concentration of PAD2 and / or PAD4 in the synovial fluid of the subject may be within the range of values ​​provided in Table 82. The concentration of PAD2 and / or PAD4 in the whole blood of the subject may be at least 1 ng / ml. The concentration of PAD2 and / or PAD4 in the whole blood of the subject may be within the range of values ​​provided in Table 83. The concentration of PAD2 or PAD4 can be determined by ELISA.

[0181] The present invention also relates to a method for treating a disease in a subject, comprising administering to the subject an anti-PAD4 antibody in combination with an anti-PAD2 antibody. The anti-PAD4 antibody and the anti-PAD2 antibody may be bivalent IgG Fab(2) fragments containing at least two binding domains for either PAD4 or PAD2, respectively. The anti-PAD2 antibody and the anti-PAD4 antibody may be administered to the subject simultaneously, separately, or sequentially.

[0182] 4.21 EPC 2000 The present invention also relates to an antibody or pharmaceutical composition of the present invention for use in a method for treating or preventing a disease in a subject. The disease may be an autoimmune disorder. The disease may be a disease characterized by increased PAD activity in a tissue compared to healthy subjects. The disease may be a disease characterized by increased PAD2 and / or PAD4 activity in a tissue compared to healthy subjects. The tissue may be synovial fluid, whole blood, or serum.

[0183] 4.22 Swiss Style The present invention also relates to an antibody of the invention or a pharmaceutical composition of the invention for the manufacture of a medicament for the treatment of an autoimmune disorder. The treatment may comprise a method of treatment according to the invention.

[0184] 5. Terminology 5.1 Antibodies The term "antibody" is used interchangeably and refers to an immunoglobulin molecule that recognizes and specifically binds to a target, such as a protein, polypeptide, peptide, carbohydrate, polynucleotide, lipid, or a combination of the foregoing, through at least one antigen recognition site within the variable region of the immunoglobulin molecule.

[0185] 5.1.1 Antibody fragments The term "antibody fragment" refers to a portion of an intact antibody. An "antigen-binding fragment," "antigen-binding domain," or "antigen-binding region" refers to a portion of an intact antibody that binds to an antigen. An antigen-binding fragment may contain an antigen-determining region of the intact antibody (e.g., a complementarity-determining region (CDR)). Examples of antigen-binding fragments of antibodies include, but are not limited to, Fab, Fab', F(ab')2, and Fv fragments, linear antibodies, and single-chain antibodies. Antigen-binding fragments of antibodies may be derived from any animal species, e.g., rodents (e.g., mice, rats, or hamsters) and humans, or may be artificially produced.

[0186] 5.1.2 Anti-PAD2 antibody The terms "anti-PAD2 antibody," "PAD2 antibody," and "antibody that binds to PAD2" are used interchangeably herein to refer to an antibody capable of binding to PAD2. The extent of binding of a PAD2 antibody to a non-PAD2 PAD may be less than about 10% of the binding of the antibody to PAD2, as measured, for example, using ForteBio or Biacore. In some embodiments provided herein, the PAD2 antibody can also bind to PAD3. In some embodiments provided herein, the PAD2 antibody does not bind to PAD3. In some embodiments provided herein, the PAD2 antibody can also bind to PAD1. In some embodiments provided herein, the PAD2 antibody does not bind to PAD1.

[0187] 5.1.3 Anti-PAD4 antibody Similarly, the terms "anti-PAD4 antibody," "PAD4 antibody," and "antibody that binds to PAD4" are used interchangeably herein to refer to an antibody capable of binding to PAD4. The extent of binding of a PAD4 antibody to a non-PAD4 PAD may be less than about 10% of the binding of the antibody to PAD4, as measured, for example, using ForteBio or Biacore. In some embodiments provided herein, the PAD4 antibody can also bind to PAD3. In some embodiments provided herein, the PAD4 antibody does not bind to PAD3. In some embodiments provided herein, the PAD4 antibody can also bind to PAD1. In some embodiments provided herein, the PAD4 antibody does not bind to PAD1.

[0188] 5.1.4 Humanized Antibodies The term "humanized" antibody or antigen-binding fragment thereof refers to forms of non-human (e.g., murine) antibodies or antigen-binding fragments thereof that are specific immunoglobulin chains, chimeric immunoglobulins, or fragments thereof that contain minimal non-human (e.g., murine) sequence. Typically, humanized antibodies or antigen-binding fragments thereof are human immunoglobulins in which residues from the complementarity-determining regions (CDRs) are replaced by residues from the CDRs of a non-human species (e.g., mouse, rat, rabbit, hamster) having the desired specificity, affinity, and capacity ("CDR-grafted") [18-20]. In some instances, Fv framework region (FR) residues of a human immunoglobulin are replaced with corresponding residues in an antibody or fragment from a non-human species having the desired specificity, affinity, and capacity. Humanized antibodies or antigen-binding fragments thereof can be further modified by substitution of additional residues either in the Fv framework regions and / or within the replaced non-human residues to improve and optimize the specificity, affinity, and / or capacity of the antibody or antigen-binding fragment. Generally, a humanized antibody or antigen-binding fragment thereof will comprise substantially all of at least one, and typically two or three, variable domains containing all or substantially all of the CDR regions corresponding to a non-human immunoglobulin, while all or substantially all of the FR regions are of human immunoglobulin consensus sequences. The humanized antibody or antigen-binding fragment thereof may also comprise at least a portion of an immunoglobulin constant region or domain (Fc), typically that of a human immunoglobulin.

[0189] 5.1.5 Human antibodies The term "human" antibody or antigen-binding fragment thereof means an antibody or antigen-binding fragment thereof having an amino acid sequence derived from the human immunoglobulin locus, and such an antibody or antigen-binding fragment is made using any technique known in the art. This definition of a human antibody or antigen-binding fragment thereof includes intact or full-length antibodies and fragments thereof.

[0190] 5.2 Binding affinity "Binding affinity" or "affinity" refers to the strength of the sum of non-covalent interactions between a single binding site of a molecule (e.g., an antibody or antigen-binding fragment thereof) and its binding partner (e.g., an antigen). Unless otherwise indicated, as used herein, "binding affinity" refers to the intrinsic binding affinity that reflects a 1:1 interaction between members of a binding pair (e.g., an antibody or antigen-binding fragment thereof and an antigen). The affinity of a molecule X for its partner Y is generally determined by the dissociation constant (K D Affinity can be expressed as the equilibrium dissociation constant (K D ) and the equilibrium association constant (K A It can be measured and / or expressed in several ways known in the art, including but not limited to, K D is k 0ff / k 0n is calculated from the quotient of A is k on / k off It is calculated from the quotient of k on refers to, for example, the association rate constant of an antibody or antigen-binding fragment thereof to an antigen, and k 0ff k refers, for example, to the dissociation of an antibody or antigen-binding fragment thereof from an antigen. on and k off can be determined by techniques known to those skilled in the art, such as Biacore® or KinExA.

[0191] 5.3 Bispecific antibodies The term "bispecific antibody" refers to an antibody containing specificity for two target molecules and includes DVD-Ig, mAb2

[21] , FIT-Ig (

[22] ), mAb-dAb, Dock-and-Lock, Fab arm exchange, SEEDbody, Triomab, LUZ-Y, Fcab, κλ-body, orthogonal Fab, scDiabody-Fc, diabody-Fc, tandem scFv-Fc, Fab-scFv-Fc, Fab-scFv, intrabody, BiTE, diabody, DART, TandAb, scDiabody, scDiabody-CH3, Diabody-CH3, Triple body, Minibody, Minibody, TriBiminibody, scFv-CH3 Formats include, but are not limited to, KIH, scFv-CH-CL-scFv, F(ab')2-scFv, scFv-KIH, Fab-scFv-Fc, tetravalent HCab, ImmTAC, knob-in-hole, knob-in-hole with a common light chain, knob-in-hole with a common light chain and charge pair, charge pair, charge pair with a common light chain, Bis3, DuetMab, DT-IgG, DutaMab, IgG(H)-scFv, scFv-(H)IgG, IgG(L)-scFv, scFv-(L)IgG, IgG(L,H)-Fv, IgG(H)-V, V(H)-IgG, IgG(L)-V, V(L)-IgG, KIH IgG-scFab, 2scFv-IgG, IgG-2scFv, scFv4-Ig, and zybody.Bispecific molecules may also be used in combination with other non-Ig formats, such as T-cell receptor binding domains, immunoglobulin superfamily domains, jawless variable lymphocyte receptors, fibronectin domains (e.g., Adnectins™), antibody constant domains where the constant domain is not a functional CH1 domain (e.g., CH3 domains, e.g., CH2 and / or CH3 of Fcabs™), scFv, (scFv)2, sc-diabodies, scFab, centrin and epitope binding domains derived from a scaffold selected from CTLA-4 (Evibody™), lipocalin domains, Z domains of Protein A (e.g., Affi The antibodies may include antibodies fused to protein A such as Protein A (e.g., Protein A domains such as ProteinA, ProteinBody™ or SpA), -A domains (e.g., Avimer™ or Maxibody™), heat shock proteins (e.g., epitope binding domains derived from GroEI and GroES), transferrin domains (e.g., TransBody™), ankyrin repeat proteins (e.g., DARPin™), peptide aptamers, C-type lectin domains (e.g., Tetranectin™), human gamma-crystallin or human ubiquitin (affilin), PDZ domains, scorpion toxins, and Kunitz-type domains of human protease inhibitors.

[0192] 5.3.1 Bis3 Bis3 format bispecific antibodies comprise an IgG molecule with two Fab domains and two scFvs, each attached to the C-terminus of a heavy chain (i.e., IgG-HC-scFv,

[23] ). The two Fab domains bind to the same target protein, and thus the molecule is symmetrical with respect to the Fab domain. The two scFvs bind different target proteins to the Fab domains, and each scFv binds the same target protein. Thus, in one embodiment, the Fab domain can bind to a first target protein (e.g., PAD2) and the scFv domain can bind to a second target protein (e.g., PAD4). Alternatively, the targets bound by the Fab domain and the scFvs can be in opposite orientations, and thus the scFv domain can bind to a first target protein (e.g., PAD2) and the Fab domain can bind to a second target protein (e.g., PAD4).

[0193] 5.3.2 DuetMab DuetMab antibodies include IgG antibodies with two heavy chains and two light chains. The two arms are asymmetric, with each arm binding to a different target protein. Thus, the antibody has a single binding domain for each of the two target proteins, resulting in a bivalent antibody overall but monovalent for each target protein. DuetMab antibodies use knob-into-hole technology for heterodimerization of two different heavy chains, increasing the efficiency of cognate heavy and light chain pairing by replacing the native disulfide bond in one of the CH1-CL interfaces with an engineered disulfide bond. Such antibodies maintain the structure and development suitability of natural IgG ([23, 24]).

[0194] 5.4 C-terminal variants Large-scale protein production involves the use of cell cultures, which are known to produce proteins that exhibit varying levels of heterogeneity. One potential source of heterogeneity involves C-terminal lysine residues, such as those typically found on the heavy chains of antibody molecules. C-terminal lysines can be lost, and as a result, individual antibodies in a production batch can vary in their C-termini with respect to whether or not a lysine residue is present ("lysine clipping")

[25] . C-terminal lysines can be present on both heavy chains of an antibody (K2), on either one of the heavy chains (K1), or absent from either of them (K0).

[0195] 5.5 Complementarity-determining regions As used herein, the term "complementarity determining region" or "CDR" refers to each of the regions of an antibody variable domain that are hypervariable in sequence and / or form structurally defined loops (hypervariable loops) and / or contain antigen-contacting residues. An antibody can contain six CDRs, e.g., three in the VH and three in the VL.

[0196] Kabat numbering is a numbering system for amino acid residues in the heavy and light chain variable regions of an antibody or antigen-binding fragment thereof. In some embodiments, CDRs can be determined according to the Kabat numbering system

[26] . Using the Kabat numbering system, the CDRs in an antibody heavy chain molecule are typically located at amino acid positions 31-35 (CDR1), 50-65 (CDR2), and 95-102 (CDR3), which can optionally include one or two additional amino acids after 35 (referred to as 35A and 35B in the Kabat numbering scheme). Using the Kabat numbering system, the CDRs in an antibody light chain molecule are typically located at amino acid positions 24-34 (CDR1), 50-56 (CDR2), and 89-97 (CDR3).

[0197] The EU index or EU numbering system is based on the sequential numbering of the first sequenced human IgG (EU antibody). The numbering scheme used herein for substitutions and insertions in the Fc region is the EU index as in Kabat

[14] . In contrast, the numbering scheme used herein for the variable regions (VH and VL) is the conventional Kabat numbering.

[0198] Chothia instead refers to the location of the structural loops.

[27] The terminus of the Chothia CDR-H1 loop, when numbered using the Kabat numbering convention, varies between H32 and H34 depending on the length of the loop (this is because the Kabat numbering scheme places the insertion at H35A and H35B; if neither 35A nor 35B is present, the loop ends at 32; if only 35A is present, the loop ends at 33; and if both 35A and 35B are present, the loop ends at 34). The AbM hypervariable regions represent a compromise between the Kabat CDRs and the Chothia structural loops and are used by Oxford Molecular's AbM antibody modeling software.

[0199] [Table 64]

[0200] 5.6 Epitope As used herein, "epitope" is a term used in the art and refers to a localized region of an antigen to which an antibody or antigen-binding fragment thereof can specifically bind. An epitope can be, for example, consecutive amino acids of a polypeptide (a linear or continuous epitope), or an epitope can arise, for example, from two or more non-contiguous regions of a polypeptide together (a conformational, non-linear, discontinuous, or discontinuous epitope). In some embodiments, the epitope to which an antibody or antigen-binding fragment thereof binds can be determined, for example, by NMR spectroscopy, X-ray diffraction crystallography studies, ELISA assays, hydrogen / deuterium exchange coupled with mass spectrometry (e.g., liquid chromatography electrospray mass spectrometry), array-based oligopeptide scanning assays, and / or mutagenesis mapping (e.g., site-directed mutagenesis mapping).

[0201] An antibody that "binds to the same epitope" as a reference antibody refers to an antibody that binds to the same amino acid residue as the reference antibody. The ability of an antibody to bind to the same epitope as a reference antibody can be determined by a hydrogen / deuterium exchange assay

[28] .

[0202] 5.7 Fc modification Combinations of mutations in the IgG Fc have been characterized to tailor immune effector function or IgG serum persistence to match the desired biological outcome of a monoclonal antibody therapeutic. Examples of IgG Fc modifications are summarized in Table 65.

[0203] [Table 65] * Amino acid numbering according to the EU index as in Kabat

[14]

[0204] The TM modification abolishes Fc effector function and is a triple mutation at the CH2 position: L234F, L235E, and P331S

[20] . The FQG, FQQ, and FAQ modifications are described in detail in WO 2013 / 65690(A1), Tsui et al.

[17] , and Borrok et al.

[30] . The FQQ modification is a more thermostable alternative to TM effector function-attenuating modifications. The YTE and N3Y modifications extend half-life. The N3Y modification is described in detail in WO 2015 / 175874(A2). The YTE modification is described in WO 2002 / 060919(A2)

[16] . The Fc mutations do not affect variable region binding affinity.

[0205] 5.8 Isolation An "isolated" polypeptide, antibody, polynucleotide, vector, cell, or composition is a polypeptide, antibody, polynucleotide, vector, cell, or composition that is in a form not found in nature. Isolated polypeptides, antibodies, polynucleotides, vectors, cells, or compositions include those that have been purified to the extent that they are no longer in the form in which they are found in nature. An isolated antibody, polynucleotide, vector, cell, or composition can be substantially pure. As used herein, "substantially pure" refers to material that is at least 50% pure (i.e., free from contaminants), at least 90% pure, at least 95% pure, at least 98% pure, or at least 99% pure.

[0206] 5.9 Knob-into-Hole Mutation The "knobs-in-hole" or "knobs-into-hole" technique refers to the mutations Y349C, T366S, L368A, and Y407V (holes) and S354C and T366W (knobs), both at the CH3-CH3 interface, to promote heteromultimerization and is described in US Patent No. 5,731,168 and US Patent No. 8,216,805 [31, 32].

[0207] Knob mutations refer to the substitutions S139C and T151W in Fc. Hole mutations refer to the substitutions Y134C, T151S, L153A, Y192V in Fc.

[0208] 5.10 Percent Identity "Percent identity" refers to the degree of identity between two sequences (e.g., amino acid sequences or nucleic acid sequences). Percent identity can be determined by aligning the two sequences and introducing gaps to maximize identity between the sequences. Alignments can be generated using programs known in the art. For purposes herein, alignment of nucleotide sequences can be performed using the blastn program set to default parameters, and alignment of amino acid sequences can be performed using the blastp program set to default parameters (see National Center for Biotechnology Information (NCBI): ncbi.nlm.nih.gov).

[0209] 5.11 RF Mutation "RF mutation" generally refers to a mutation of amino acids HY to RF in the CH3 domain of an Fc domain, for example, the mutations H435R and Y436F in the CH3 domain. RF mutations abolish binding to Protein A.

[0210] 5.12 Effect "Potency" is usually expressed as an IC50 value in nM unless otherwise specified. IC50 is the median inhibitory concentration of an antigen-binding molecule. In functional assays, IC50 is the concentration that reduces a biological response by 50% of its maximum. In ligand binding studies, IC50 is the concentration that reduces receptor binding by 50% of the maximum specific binding level. IC50 can be calculated by any number of means known in the art. Improved potency can be determined, for example, by measuring against a parent antibody (e.g., a parent antibody before germlining or a parent antibody before affinity optimization).

[0211] 5.13 Variable Region As used herein, the terms "variable region" and "variable domain" are used interchangeably and are common in the art. A variable region typically refers to a portion of an antibody, generally a portion of either the light or heavy chain, typically the amino-terminal approximately 110-120 or 110-125 amino acids in the mature heavy chain and approximately 90-115 amino acids in the mature light chain, which vary significantly in sequence among antibodies and are used to determine the binding and specificity of a particular antibody for its specific antigen. Sequence variability is concentrated in regions called complementarity-determining regions (CDRs), while the more highly conserved regions in the variable domain are called framework regions (FRs). While not wishing to be bound by any particular mechanism or theory, it is believed that the CDRs of the light and heavy chains are primarily responsible for the interaction and specificity of the antibody with its antigen.

[0212] 6 Example 1: Methods 6.1 T onset Stability assay Differential scanning fluorimetry (DSF) is a fluorescence-based protein stability assay that measures the folding state of a protein by monitoring the change in fluorescence as a function of temperature. This technique is useful for determining the midpoint temperature of thermal unfolding (T m ) and starting temperature (T onset NanoDSF provides biophysical properties such as temperature, time, or denaturant concentration. NanoDSF is a dye-free DSF method that monitors changes in intrinsic fluorescence from intrinsic tryptophan in proteins as a function of temperature, time, or denaturant concentration

[33] . Protein unfolding changes the polarity of the microenvironment around the tryptophan residues, causing a red shift in fluorescence

[18] . Based on this principle, nanoDSF can determine T by measuring the ratio of fluorescence intensities at 330 nm and 350 nm as a function of temperature. m and T onset Determine.

[0213] 6.2 Specificity assays Antibody samples were tested for binding to PAD1, PAD2, PAD3, and PAD4 using an ELISA assay. Black high-binding plates (Greiner, 781077) were prepared by adding 20 μL of a 20 μg / mL solution of streptavidin (Invitrogen, S888) in HBSS buffer (Sigma, H8264) and incubated at 4°C for 16 hours. Plates were equilibrated to room temperature, washed three times with 100 μL of PBS (Oxoid, BR0014G) containing 0.1% Tween 20 (Sigma, P2287), and 80 μL of 1% BSA in HBSS (Sigma, A7979) was added. After 1 hour, plates were washed as described above, and 20 μL of biotinylated PAD1, PAD2, PAD3, or PAD4 (prepared at 1 μg / mL in a buffer containing 50 mM HEPES pH 7.3 (VWR, J848), 10 mM CaCl2 (Sigma, 21115), 120 mM NaCl (Sigma, S5150), 5 mM DTT (Sigma, 43816), and 0.1 mM CHAPS (Sigma, 19899) was added. After 1 hour, plates were washed as described above, and 20 μL of antibody (prepared in 50 mM HEPES pH 7.3, 10 mM CaCl2, 120 mM NaCl, 5 mM DTT, and 0.1 mM CHAPS) was added. After 1.5 hours, the plate was washed as described above, and 20 μL of anti-human IgG Fc HRP (Southern Biotech, 9040-05) prepared in HBSS containing 0.5% BSA was added at a concentration of 50 ng / mL. After 1 hour, the plate was washed as described above, and 20 μL of room temperature QuantaBlu working solution was added. After 30 minutes, 20 μL of QuantaBlu stop solution was added (QuantaBlu kit ThermoFisher, 15169). Fluorescence was measured on a PHERAStar FSX plate reader (BMG Labtech) using a 330 nm excitation filter and a 420 nm emission filter.

[0214] 6.3 Interference scattering microscopy (iSCAT) Interference scattering microscopy (iSCAT) or mass spectrophotometry measures the light scattering of individual protein molecules in close proximity to a glass surface. The contrast is directly proportional to the molecular weight of the individual protein molecules / complexes. Measurement tips were prepared by cleaning polished microscope cover slides in deionized water and isopropanol, followed by drying in compressed air, and then mounting them into silicon cassette wells. Measurements were initiated by the addition of buffer (10 μL) to selected wells, followed by automatic focusing.

[0215] Protein samples were added (approximately 50–100 nM, 10 μL), and contrast from individual protein molecules was recorded within the field of view for 60 seconds (Refeyn Acquire). The resulting raw data movie was processed by image analysis (Refeyn Discover) to construct a mass histogram of the sample.

[0216] 6.4 Trypsin cleavage efficacy assay PAD activity was determined using a short peptide (Cambridge Research Biochemicals) containing arginine flanked by AlexaFluor 488, which acts as the FRET donor, and QSY7, which acts as the FRET acceptor. If the arginine is deiminated to citrulline by PAD activity, trypsin will not cleave the peptide, and fluorescence from the donor will be quenched by the acceptor. Inhibition of PAD with anti-PAD scFv prevents arginine deimination, rendering the peptide susceptible to trypsin cleavage. The resulting separation of the donor and acceptor fluorophores allows detectable emission from the donor. 2.5 ml of sample scFv was prepared in assay buffer containing 50 mM HEPES, 5 mM DTT, 10 mM CaCl2, and 0.01% CHAPS and preincubated with 2.5 μl of PAD4 at 10 nM (final concentration). Peptide substrates were prepared as stock solutions and 5 ml was added to give a final concentration of 100 nM. After the appropriate incubation time, 10 ml of 100 nM (final concentration) trypsin was added and the reaction was allowed to proceed for at least 2 minutes before being read on an EnVision plate reader (PerkinElmer, Waltham, MA).

[0217] 6.5 Histone-H3 PAD potency assay PAD2 and PAD4 activity was measured using a histone-H3 citrullination assay using different substrates as described below.

[0218] 6.5.1 Recombinant PAD Human PAD4 (RD223) was used at 0.15 ng / mL (1.8 pM), human PAD2 (RD220) was used at 0.02 ng / mL (0.24 pM), 30 min preincubation of PAD and bispecific antibody, 3 h incubation.

[0219] 6.5.2 Synovial fluid A 96-well high-binding half-area plate was coated with 1 μg / ml HIS-H3 overnight at 4°C. RA synovial fluid (DX01156) diluted in citrullination buffer was preincubated with EDTA or serial dilutions of antibody for 30 minutes, then transferred to the histone H3-coated plate and incubated at 37°C for 1.5 hours. Citrullinated histone H3 was detected by incubation with rabbit anti-human citrullinated histone H3 Ab for 1 hour, followed by incubation with goat anti-rabbit HRP-conjugated Ab for 1 hour. The color reaction was developed using UltraSensitive TMB substrate and measured at 450 nm.

[0220] 6.5.3 Whole blood Fresh whole blood was administered with bispecific Ab overnight. Plasma was collected, and PAD activity was assessed using a histone H3 PAD activity assay at multiple plasma dilutions. Whole blood from normal healthy donors was incubated with Ab overnight at 37°C. Plasma was collected and stored frozen at -80°C. A 1 / 5 plasma dilution was used for the histone-H3 PAD activity assay. PAD surface expression was assessed by FACS. Soluble PAD was assessed by ELISA.

[0221] 6.6 Histone H3cit Western blot analysis PAD enzymes citrullinate histone H3 to H3cit, which can be measured using a Western blot protein detection system. LPS exposure results in elevated H3cit levels in vivo due to PAD enzyme activity in mouse lungs. H3cit expression was analyzed using Western blot in bronchoalveolar lavage (BAL) fluid from WT and PAD4KO mice exposed to LPS and saline. Mice treated with anti-PAD2 / anti-PAD4 had lower levels of H3Cit in BAL fluid compared with WT mice. BAL fluid from PAD4KO mice lacked H3Cit.

[0222] 6.7 huFcRn affinity chromatography The affinity of the sample for huFcRN was characterized using a huFcRN-bound Sepharose column. Approximately 40 μg / 40 μL of sample was loaded onto a 1 mL column, followed by a linear gradient of 3 column volumes (CV) from buffer A (20 mM MES, 150 mM NaCl, pH 5.5) to 40% buffer B (20 mM Tris + 150 mM NaCl, pH 8.8), and then an 18 CV linear gradient from 40% to 100% buffer B. The experiment was performed at room temperature using an Agilent DAD at a flow rate of 0.5 mL / min, and the A280 and retention time of the elution profile were measured.

[0223] 6.8 Cynomolgus monkey safety study The target-association assay centers on the ability of both PAD2 and PAD4 to citrullinate histone H3. The histone H3 substrate is coated onto a plate, and active PADs in the sample deaminate the arginine residue to form citrulline. These citrullinated epitopes are then detected by standard immunoassay methods.

[0224] 6.9 Cytokine Safety Assays The potential for induction of cytokine release by the Bis3 and DuetMab formats (clone 06 and clone 12) was evaluated in eight donors (four healthy patients and four RA patients) using the following methodologies: a soluble stimulated whole blood assay and a wet-coated immobilized stimulated isolated PBMC assay. The concentrations of the cytokines IFN-γ, IL-2, IL-6, and TNF-α in collected plasma and cell culture supernatants were measured using Luminex. For each whole blood and isolated PBMC sample, negative control wells were also set up using the same lot and volume of PBS used to prepare the test items and controls.

[0225] 6.10 Generation of Padi4 knockout mouse lines A DNA targeting vector was designed and cloned to modify the mouse endogenous Padi4, peptidylarginine deaminase, type IV gene. This strategy was based on cloning LoxP sites into the introns flanking exons 7 and 10 of the Padi4 gene. Upon Cre-induced recombination, this generated a knockout KO, leaving single LoxP sites 276 bp upstream of exon 7 and 736 bp downstream of exon 10. The targeting vector was used to modify the Padi4 locus in Primogenix, PrX, mouse embryonic stem cells (C57B16 / N origin) by homologous recombination. For chimera generation, correctly targeted cells were injected into 3- and 5-day-old Balb C blastocysts. For lineage expansion, male chimeric mice were mated with C57B16 / N females. The established mouse line was crossed with the R26 Cre-deficient line to generate the final Padi4 KO allele.

[0226] 6.11 Biacore affinity analysis 6.11.1.1 Fab and Monovalent Bispecific Antibodies (DuetMabs) The affinity of anti-PAD2 antigen-binding fragments (Fabs), anti-PAD4 Fabs, or DuetMAbs for PAD species was measured at 25°C using a Biacore 8K (Cytiva). Experiments were performed using recombinant human PAD2, cynomolgus monkey PAD2, mouse PAD2, human PAD4, cynomolgus monkey PAD4, and mouse PAD4. All species were enzymatically biotinylated on the Avi tag. Anti-PAD Fabs were either expressed

[34] or obtained by papain digestion of anti-PAD IgG. After a 20-minute papain incubation, samples were injected at 0.5 ml / min onto a Superdex 200 Increase 10 / 300 GL column equilibrated in D-PBS to isolate Fabs. Streptavidin was covalently immobilized on a CM5 or C1 chip surface using standard amine coupling techniques. Recombinant biotinylated PAD2 and PAD4 species were titrated onto a streptavidin chip surface in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% Surfactant P20, 1 mM CaCl, 1 mM DTT buffer to allow Fab or DuetMab binding. Fab or DuetMab was serially diluted in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% Surfactant P20, 1 mM CaCl, 1 mM DTT and flowed over the chip at 50 μl / min, allowing 3 minutes for association and 10 minutes for dissociation. Multiple buffer-only injections were performed under the same conditions to allow double-reference subtraction of the final sensorgram set. Alternatively, Fab was serially diluted in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% Surfactant P20, 1 mM CaCl2, 1 mM DTT and injected at successively increasing concentrations over the chip at 50 μl / min (single-cycle kinetics), with 2 min of association and 10 min of dissociation at the end of a complete binding cycle. A buffer-only injection was performed under the same conditions to allow for double-reference subtraction of the final sensorgram set. All sensorgram sets were analyzed using Biacore 8K Evaluation Software. The chip surface was fully regenerated with a pulse of 3.0 M MgCl2.

[0227] 6.11.1.2 Bispecific antibodies The affinity of recombinant PAD species for anti-PAD IgG, bivalent Bis3 anti-PAD molecules, or monovalent bispecific antibodies (DuetMab) was measured at 25°C using a Biacore 8K (Cytiva). Experiments were performed using recombinant human PAD2, cynomolgus PAD2, mouse PAD2, human PAD4, cynomolgus PAD4, and mouse PAD4. Protein G' was covalently immobilized on the C1 chip surface using standard amine coupling techniques at a concentration of 20 μg / ml in 10 mM sodium acetate, pH 3.65. IgG, Bis3 molecules, or DuetMab were captured onto the surface of Protein G' at 5 μl / min in 10 mM HEPES, pH 7.4, 150 mM NaCl, 0.05% Surfactant P20, 1 mM CaCl2, and 1 mM DTT buffer to allow PAD binding. Alternatively, anti-PAD molecules were chemically biotinylated using EZ link Sulfo-NHS-LC-Biotin (Thermo) and captured on a C1-streptavidin surface (prepared as described above). PAD species were serially diluted in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% Surfactant P20, 1 mM CaCl2, and 1 mM DTT and injected at increasing concentrations over the chip at 50 μl / min (single-cycle kinetics), with 2 min of association and 10 min of dissociation at the end of a complete binding cycle. To allow double-reference subtraction of the final sensorgram set, a buffer-only injection was performed under the same conditions, and this final sensorgram set was analyzed using Biacore 8K Evaluation Software. The Protein G' chip surface was fully regenerated with a pulse of 6 M guanidine HCl in D-PBS to remove the captured molecules along with any bound PAD. On the other hand, the streptavidin surface was completely regenerated with a pulse of 3.0 M MgCl2.

[0228] 7 Example 2: Targeting PAD2 and PAD4 in the treatment of RA PAD2 and PAD4 drive citrullination in patients with RA. We confirmed that both PAD2 and PAD4 can generate citrullinated RA antigens fibroinogen β-chain

[35] and α-enolase

[36] (Figure 1).

[0229] PAD4 and PAD2 levels were also found to be increased in RA serum (Figure 3A) and synovial fluid (Figure 3B). RNA expression profiling revealed that PAD4 was highly expressed in neutrophils and monocytes (human and cynomolgus monkey) (Figures 3C and 3D). PAD2 and PAD4 were also found to be highly expressed on the cell surface of immune cells (Figures 4A and 4B).

[0230] 8 Example 3: Generation of anti-PAD2 antibodies and anti-PAD2 antibodies 8.1 Generation of PAD2 antibodies For antibody selection, we used several human antibody libraries derived from adult naive donors. Parent antibody clones were isolated from a large single-chain Fv (scFv) human antibody library derived from spleen cells from adult naive donors and cloned into a phagemid vector based on the filamentous phage M13.

[0231] PAD2-specific scFv antibodies were isolated from a phage display library through a series of iterative selection cycles against recombinant bacterially expressed human PAD2 bearing a hybrid chain, including selection cycles against recombinant bacterially expressed cynomolgus monkey or mouse PAD2

[37] . The scFvs were expressed in the bacterial periplasm and screened for their inhibitory activity in a trypsin cleavage assay. Screening hits, i.e., scFv clones that showed an inhibitory effect on PAD2 citrullination activity, were subjected to DNA sequencing. Unique scFvs were again expressed in bacteria and affinity purified. Furthermore, a subset of the selection output was subcloned for high-throughput expression and screening in an IgG format. Functional clones from either approach were reformatted into antibody heavy and light chain vectors for mammalian expression as IgG or Fab molecules. The anti-PAD2 antibodies were then affinity-optimized.

[0232] Anti-PAD2 antibodies (e.g., PAD40141, PAD40119, and PAD40175, Tables 52 to 54) have an affinity (K ) of approximately 6 to 260 nM for human PAD2. D ) (Table 69).

[0233] The anti-PAD2 antibodies were then affinity-optimized. The sequences of the obtained anti-PAD2 antibodies are shown in Tables 1 and 43 to 54. The affinities of the anti-PAD2 antibodies are shown in Tables 68 and 69.

[0234] 8.2 Generation of PAD4 antibodies For antibody selection, several human antibody libraries derived from adult naive donors were used. Parent antibody clones were isolated from a large single-chain Fv (scFv) human antibody library derived from spleen cells from adult naive donors and cloned into a phagemid vector based on the filamentous phage M13.

[0235] PAD4-specific scFv antibodies were isolated from a phage display library through a series of iterative selection cycles against recombinant mammalian-expressed human PAD4 bearing a hybrid chain, including selection cycles against recombinant mammalian-expressed cynomolgus monkey or mouse PAD4

[37] . The scFvs were expressed in bacterial periplasm and screened for their inhibitory activity in a trypsin cleavage assay. Screening hits, i.e., scFv clones that showed an inhibitory effect on PAD2 citrullination activity, were subjected to DNA sequencing. Unique scFvs were again expressed in bacteria and affinity purified. Furthermore, a subset of the selection output was subcloned for high-throughput expression and screening in an IgG format. Functional clones from either approach were reformatted into antibody heavy and light chain vectors for mammalian expression as IgG or Fab molecules.

[0236] A representative anti-PAD4 antibody (PAD40048) has an affinity (K) of approximately 87 nM for human PAD4. D The sequence of PAD40048 is provided in Table 62.

[0237] The anti-PAD4 antibodies were then affinity-optimized. The sequences of the obtained anti-PAD4 antibodies are shown in Tables 43 to 54. The affinities of the PAD4 antibodies are shown in Tables 74 and 75.

[0238] PAD4 murine surrogate antibodies were also identified and their affinities are shown in Table 80. The sequences of the surrogate affinity-optimized antibodies are shown in Tables 55 and 56.

[0239] 9 Example 4: iSCAT analysis of anti-PAD4 and anti-PAD2 binding behavior iSCAT analysis revealed that the Fab format of the anti-PAD4 affinity-optimized antibody clearly favored the heterotetrameric form, suggesting that the Fab stabilized the dimeric form of PAD4. Surprisingly, the PAD4 monomer with the Fab and the PAD4 dimer with only one Fab were completely absent, reinforcing the same conclusion (Figure 5, arrows).

[0240] Similar to Fab, the addition of IgG also favored heterotetramers, suggesting that IgG anti-PAD4 also stabilizes the dimeric form of PAD4 (Figure 6). Here, the larger size of IgG (no overlapping peaks) allowed detection of monomeric PAD4 species. Similar to Fab anti-PAD4, i) PAD4 monomers with IgG and ii) PAD4 dimers with only one IgG were completely absent. Also, there was an additional peak at approximately 600 kDa that was detectable only with IgG but not with Fab. These data suggested that a single anti-PAD4 IgG can span both binding sites of dimeric PAD4 (Figure 6).

[0241] In summary, mass spectrophotometric analysis indicated that both Fab and IgG anti-PAD4 affinity-optimized antibodies strongly favored PAD4 heterotetrameric complexes. Furthermore, IgG, but not Fab, induced additional complexes that could correspond to heterohexamers containing 2×PAD4 dimers and 2×IgG. This configuration allows for multivalency even in solution, thus making IgG substantially more potent than Fab.

[0242] Similar to PAD4, PAD2 also exhibited a monomer-dimer equilibrium (Figure 7). In contrast to PAD4, addition of the anti-PAD2 Fab affinity-optimized antibody generated both PAD2 heterodimers and heterotetramers, suggesting a different binding epitope location compared to the anti-PAD4 antibody. IgG PAD2 showed only one complex corresponding to a single IgG, suggesting that the binding epitope is positioned such that IgG binding results in a steric clash with the second binding site (arrow). Furthermore, the complete absence of the 600 kDa peak suggested that PAD2 IgG cannot form multivalent complexes like those seen with PAD4. This is consistent with the citrullination assay data, which showed lower "add-on" potency for PAD2 relative to PAD4 IgG.

[0243] In summary, PAD4 Fab and IgG anti-PAD4 binding stabilized PAD4 dimers. The epitope enabled IgG anti-PAD4 to form hexameric complexes with PAD4 dimers (Figure 8A). In contrast, PAD2 exhibited a monomer-dimer equilibrium, but anti-PAD2 binding had limited effect on dimer stabilization (Figure 8B). Furthermore, in contrast to PAD4, only a single IgG could bind PAD2, suggesting that steric hindrance of the second binding site also prevented the potential formation of hexameric complexes.

[0244] 10 Example 5: Generation of anti-PAD2 / anti-PAD4 bispecific antibodies Monovalent and bivalent bispecific antibodies were generated in DuetMab or Bis3 formats, respectively (Figures 10A and 10B).

[0245] 10.1 DuetMab Format DuetMabs with two different orientations were generated from affinity-optimized anti-PAD2 and anti-PAD4 antibodies (clone 22 and clone 42) (Figure 10A, Table 66). The sequences of exemplary DuetMabs bispecific antibodies (clones 01-06) are shown in Table 58. Clone 06 was further modified to remove the RF site from the end of the hole Fc (Table 58, SEQ ID NO:59 and SEQ ID NO:60) to abolish Protein A binding. The Fc region was modified to contain knob-in-hole mutations.

[0246] [Table 66]

[0247] 10.2 Bis3 Format Two different orientation Bis3 format bispecific antibodies were generated from affinity-optimized anti-PAD2 and anti-PAD4 antibodies (clone 22 and clone 42) (Figure 10B, Table 67). The sequences of exemplary Bis3 bispecific antibodies are shown in Table 57 (clones 07-12).

[0248] [Table 67]

[0249] 11 Example 6: Affinity 11.1 Affinities of Prior Art Anti-PAD2 and Anti-PAD4 Antibodies Aosasa et al. reported that the affinity (K ) for PAD2 ranged from 6.33 to 74 nM. D )

[13] . Therefore, the anti-PAD2 antibodies described in the prior art had low affinity for the target.

[0250] 11.2 Affinity of anti-PAD2 and anti-PAD4 The affinities of the affinity-optimized anti-PAD2 antibodies from Example 3 for human, cynomolgus monkey, and mouse PAD2 are shown in Table 68.

[0251] [Table 68-1]

[0252] [Table 68-2] Biotinylated PAD2 was captured on a CM5 / C1-streptavidin surface and anti-PAD2 Fab was flowed over. SD = Standard Deviation. nd = not determined.

[0253] Summary data for the anti-PAD2 antibodies are shown in Table 69 and Figure 13. In contrast to prior art anti-PAD2 antibodies

[13] , 10 of the anti-PAD4 clones had affinities for human PAD2 of less than 6 nM (Figure 13, Table 69). Four of the clones had affinities for human PAD2 of less than 1 nM (Figure 13, Table 69). Two of the clones had affinities for human, cynomolgus monkey, and mouse PAD2 that were less than 0.3 nM (Figure 13, Table 69).

[0254] [Table 69] Biotinylated PAD2 was captured on a CM5 / C1-streptavidin surface and anti-PAD2 Fab was applied. Data shown are the mean of n = 2-9 experiments. nd = not determined.

[0255] The affinity of the anti-PAD4 antibodies from Example 2 for human and cynomolgus monkey PAD4 is shown in Table 70.

[0256] [Table 70-1]

[0257] [Table 70-2]

[0258] [Table 70-3] Biotinylated PAD4 was captured on a CM5 / C1-streptavidin surface and anti-PAD4 Fab was flowed over. SD = standard deviation. nb = no binding.

[0259] Summary data for anti-PAD4 antibodies are shown in Table 71 and Figure 14. Affinity for human PAD4 (K D ) less than 0.1 and / or 1 nM are highlighted in FIG. 14 (dotted line).

[0260] [Table 71] Biotinylated PAD4 was captured on a CM5 / C1-streptavidin surface and anti-PAD4 Fab was flowed over. Data shown are the average of n=1-6 experiments. nb=no binding.

[0261] 11.3 Bispecific antibody affinity The affinity of different bispecific antibody formats for PAD2 was measured and is shown in Table 72 and summarized in Table 73. In this assay, the affinity of clone 22 as a Fab (141LO0035 hIgG1 pgl-4) was measured to be 10-fold higher than in the previous assay (Table 69). The recombinant PAD2 used here was different compared to when the affinity of the anti-PAD2 antibody from Example 3 was measured. Here, the tag on recombinant PAD2 (and PAD4) was on the N-terminus of the protein compared to the previous C-terminus. This potentially results in better folding of the protein or better presented epitopes, which could explain the higher affinity measured in this experiment.

[0262] The affinity of different bispecific antibody formats to PAD4 was measured and is shown in Table 74 and summarized in Table 75.

[0263] The affinity of DuetMab for PAD2 and PAD4 was comparable to that of the respective Fab and IgG in both orientations. In other words, when the optimized anti-PAD2 and anti-PAD4 antibodies were combined into the DuetMab bispecific antibody, there was surprisingly no loss of PAD2 or PAD4 affinity (Tables 72-75). The affinity (K D The affinity (K) of DuetMab for human PAD4 ranged from approximately 10 to 18 pM (Table 73). D ) ranged from approximately 40 to 58 pM (Table 75). Fc modifications did not affect the affinity of DuetMab. In summary, all of the DuetMab bispecific antibodies retained high affinity for both PAD2 and PAD4 (human, cynomolgus monkey, and mouse (PAD2 only)), regardless of orientation (i.e., whether anti-PAD2 or anti-PAD4 was "hole" or "knob").

[0264] All Bis3 bispecific antibodies retained high affinity for both PAD2 and PAD4 (human, cynomolgus monkey, and mouse (PAD2 only)), regardless of orientation (i.e., whether anti-PAD2 or anti-PAD4 was scFv or Fab) (Tables 72-75).

[0265] Bis3 format bispecific antibodies had reduced human PAD4 affinity compared to IgG (Table 75). Bis3 formats with PAD4 arms in the Fab format (e.g., clone 07) were better at maintaining affinity for both human and cynomolgus PAD4. The affinity (K D The affinity (K) for human PAD4 ranged from approximately 8 to 50 pM (Table 75). Clone 12 had an affinity (K) of approximately 30 pM for human PAD4. D The Bis3 format bispecific antibody had affinity for human PAD2 comparable to that of an IgG containing the same PAD4 binding domain (Table 73). The affinity (K D ) ranged from approximately 6 to 16 pM (Table 73). The Bis3 format with the PAD4 arm in the Fab format was better at maintaining affinity for cynomolgus PAD4 compared to DuetMab and other Bis3 formats (Table 76). Fc modifications did not significantly affect affinity for either the Bis3 or DuetMab formats (Tables 72-75).

[0266] [Table 72-1]

[0267] [Table 72-2] * = Antibody captured on the C1-Protein G surface. # = biotinylated antibody captured on C1-streptavidin surface. SD = standard deviation. nd = not determined.

[0268] [Table 73] * = Antibody captured on the C1-Protein G surface. # = biotinylated antibody captured on a C1-streptavidin surface. Data shown are the average of n = 2-8 experiments. nd = not determined.

[0269] [Table 74] * = Antibody captured on the C1-Protein G surface. # = biotinylated antibody captured on C1-streptavidin. SD = standard deviation. nd = not determined.

[0270] [Table 75] * = Antibody captured on the C1-Protein G surface. 、# = biotinylated antibody captured on a C1-streptavidin surface. Data shown are the average of n = 3-7 experiments. nd = not determined.

[0271] [Table 76] * = Antibody captured on the C1-Protein G surface. # = biotinylated antibody captured on a C1-streptavidin surface. Data shown are the average of n = 3-7 experiments. nd = not determined.

[0272] 11.4 Affinity for different PAD2 and PAD4 haplotypes The affinity of the two most dominant cynomolgus haplotypes for the Bis3 format (clone 08) was comparable and approximately two-fold lower for the least dominant cynomolgus haplotype (Tables 77 and 78). The affinity of all cynomolgus PAD4 haplotypes for clone 08 and clone 06 was within two-fold of each other. The affinity of the two most dominant haplotypes (representative clones 02 and 08) for both the Bis3 and DuetMab formats was comparable. The affinity of the bispecific antibodies (both Bis3 and DuetMab) for the three human PAD4 haplotypes was comparable (Table 79).

[0273] [Table 77] Antibodies were captured on a C1-Protein G' surface and PAD2 was flowed over. SD = standard deviation.

[0274] [Table 78] Antibodies were captured on a C1-Protein G' surface and PAD4 was flowed over. SD = standard deviation.

[0275] [Table 79] Antibodies were captured on a C1-Protein G' surface and PAD4 was flowed over. SD = standard deviation.

[0276] 11.5 Affinity of Mouse Surrogate Anti-PAD4 The affinities of the mouse surrogate anti-PAD4 are shown in Table 80.

[0277] [Table 80] Biotinylated antibodies were captured on a C1-streptavidin surface and PAD4 was flowed over. SD = standard deviation.

[0278] 12 Example 7: Efficacy 12.1 Potency of Prior Art Anti-PAD4 Antibodies The anti-PAD4 antibodies (L78-4, L119-5, L198-3 and L207-11) described in WO 2012026309(A9) [9] only showed 10-40% inhibition of PAD activity at an antibody concentration of 1000 nM.

[0279] WO 2016 / 155745(A1)

[11] suggests a mouse monoclonal antibody that is cross-reactive against PAD2, PAD4, and PAD3. It was shown that the cross-reactive antibody was only able to inhibit PAD2 activity by a maximum of approximately 25% compared to a control antibody, as measured by using a fibroblast citrullination assay.

[0280] Previously described humanized anti-PAD4 antibodies showed little or no inhibition of PAD4 activity in synovial fluid (Figure 2), as assessed by the H3 histone citrullination assay (Figure 2A) or the BAEE PAD activity assay (Cayman Chemical) (Figure 2B).

[0281] 12.2 Efficacy of anti-PAD4 and anti-PAD2 antibodies Using the PAD4 histone H3 citrullination potency assay, it initially proved difficult to distinguish the potency of the generated anti-PAD4 antibodies from that of the parent antibody described in Example 3. Using the assay, IgGs exhibited IC values ​​close to the concentration of PAD4 in the system. 50 Stacking values ​​were observed (i.e., 50 ng / ml, 0.65 nM).

[0282] We found that significantly reducing the concentration of PAD4 used in the assay (from 50 ng / ml to 15 pg / ml, a 3,333-fold reduction) improved the assay so that differences in potency could be more easily assessed. Reducing the concentration of PAD4 used in the assay required optimization of other parameters in the assay, particularly extending the PAD4 enzyme reaction incubation time and using ultra-TMB HRP detection (Figure 9A). Under these modified conditions, the affinity-optimized IgG showed improved potency compared to the parent. However, assay stacking was then observed in the low 100 fM range. We hypothesized that this was due to bivalent binding of the antibody to PAD4 homodimers. This hypothesis was confirmed by testing the potency of the antibody as a Fab fragment (Figure 9A) (see also Section 9, Example 4).

[0283] Improved potency of affinity-optimized anti-PAD4 antibodies could be observed when the antibodies were tested as Fab fragments using improved potency assays (Table 81).

[0284] [Table 81] Anti-PAD4 Fab in histone H3 assay, 3h 45min PAD4 incubation, PAD4 concentration: 0.15ng / ml

[0285] 12.3 PAD2 and PAD4 activity in RA synovial fluid The efficacy of anti-PAD2 and anti-PAD4 antibodies can be assessed, for example, using a histone-H3 activity assay for PAD activity in synovial fluid or whole blood from RA patients (Figure 11). Both PAD2 and PAD4 contribute to PAD activity (as determined by histone-H3 activity assay) in synovial fluid from RA patients (Table 82).

[0286] [Table 82] PAD4 and PAD2 levels were determined using Cayman ELISA kits.

[0287] 12.4 PAD2 and PAD4 activity in whole blood of RA patients PAD2 and PAD4 can also be detected at various concentrations in the whole blood of RA patients (Table 83).

[0288] [Table 83]

[0289] 12.5 Efficacy against recombinant PAD The potencies of DuetMab and Bis3 bispecific antibodies were compared directly using an optimized histone-H3 citrullination ELISA (see Section 12.2) and recombinant PAD2 and PAD4 (Table 84). In the DuetMab format, orientation did not affect potency (e.g., clone 01 vs. clone 02).

[0290] IC of all bispecific antibody formats and IgG clone 22 inhibits histone citrullination of recombinant PAD2 50 was lower than that reported for prior art anti-PAD2 antibodies

[13] . In particular, Aosasa et al. reported that the potencies of anti-PAD2 antibodies (S4, S10, S24, S108, S170, and S309) that inhibit recombinant PAD2 histone citrullination ranged from 7.0 to 75 nM

[27] . In contrast, the potencies (IC) of DuetMab, Bis3, and clone 22 against recombinant PAD2 in equivalent assays were 50 ) were all less than 1 nM (Table 84).

[0291] [Table 84] Histone-H3 assay using human PAD4 (RD223) at 0.15 ng / mL and human PAD2 (RD220) at 0.02 ng / mL, with a 30-minute preincubation of PAD and bispecific antibodies, followed by a 3-hour incubation.

[0292] In the DuetMab format, there was a 3-4 fold loss in PAD2 potency for the optimized anti-PAD2 lead (clone 22) in IgG format and a 7-11 fold loss in PAD4 potency for the optimized anti-PAD4 lead (clone 42) in IgG format (Table 84). There was no significant loss in potency or an increase in potency for PAD2 and PAD4 in the Bis3 format (Table 84). PAD2-Fab Bis3 (e.g., clone 08) performed better than PAD4 Fab Bis3 (e.g., clone 07) (Table 84).

[0293] 12.6 Efficacy against PAD activity in synovial fluid of RA patients Anti-PAD2 and PAD4 Abs are effector-null (TM) antibodies and have been shown to be effective in blocking extracellular PAD activity. Surprisingly, blocking PAD2 and PAD4 in the synovial fluid of RA patients demonstrated that PAD2 and PAD4 activity is non-redundant. Blocking PAD2 and PAD4 with anti-PAD2 and anti-PAD4 inhibited all PAD activity in RA synovial fluid (Figure 12).

[0294] Potency analysis of DuetMab in both orientations showed that the bispecific antibody inhibited PAD (combined PAD2 and PAD4) activity in synovial fluid (Table 85).

[0295] [Table 85] Histone-H3 activity assay, synovial fluid, 1:2000 dilution, CI95: 95% confidence interval

[0296] Potency analysis of the Bis-3 Ab format in two orientations showed that the bispecific antibody inhibited combined PAD2 and PAD4 activity in synovial fluid (Table 86). The ability of the Bis3 bispecific antibody to inhibit combined PAD activity in synovial fluid from RA patients was high and dose-dependent.

[0297] Surprisingly, the "scFv(PAD4)-Fab(PAD2)" Bis3 format (clones 08, 10 and 12) performed better than the combination of optimized anti-PAD2 (clone 22) and anti-PAD4 (clone 42) antibodies (Table 87).

[0298] [Table 86] Histone-H3 activity assay, synovial fluid, 1:2000 dilution, CI95: 95% confidence interval

[0299] [Table 87]

[0300] 12.7 Efficacy against PAD activity in whole blood The potency of the bispecific antibodies was also evaluated in whole blood-derived plasma samples. The Bis3 format was more potent than the DuetMab format (clone 12, PAD240012) in this assay (Figure 25B). Both the Duet and Bis3 format bispecific antibodies were able to completely inhibit PAD activity in both synovial fluid and whole blood (Figure 25).

[0301] [Table 88] Histone-H3 PAD activity assay, antibodies were incubated overnight in RA whole blood

[0302] 12.8 In Vivo Efficacy The in vivo efficacy of Bis3 PAD2 / PAD4 bispecific antibody was evaluated at low and high doses (Figure 22A). Bis3 bispecific antibody was able to suppress endogenous PAD activity in vivo, with rapid and nearly complete target engagement by day 57 using low doses of Bis3 bispecific antibody, and PAD activity returned to pre-treatment levels by approximately day 85. At the high dose, no detectable recovery of PAD activity was observed at the completion of the study (day 106) (Figures 22C, 22E).

[0303] The Bis3 bispecific antibody was able to suppress PAD activity in spiked plasma, with rapid and nearly complete target engagement by day 29 using low doses of the Bis3 bispecific antibody, and PAD activity returned to pre-dose levels by approximately day 85. At higher doses, no detectable recovery of PAD activity was observed at the completion of the study (day 106) (Figures 22B, 22D).

[0304] 12.9 Efficacy for different haplotypes The Bis3 format retained similar potency (within 2-fold) against human and cynomolgus PAD4 and cynomolgus PAD2 haplotypes (Table 89).

[0305] [Table 89] Bis3 = scFv(PAD4)-Fab(PAD2), clone 12

[0306] 12.10 Efficacy of anti-PAD2 and anti-PAD4 antibodies The following antibodies described in the art were cloned, expressed, and purified: anti-PAD2 antibody mAb2

[10] , anti-PD4 antibodies G8H4 and H7H4

[12] , and 4R147

[38] .

[0307] The potency of clones 12, 22, and 42 was directly compared to mAb2, G8H4, H7H4, and 4R147 using an optimized histone-H3 citrullination ELISA (Section 12.2) and recombinant PAD2 and PAD4 (Figures 23A and 23B, respectively).

[0308] Clones 12 and 22 completely inhibited hPAD2 enzyme activity, whereas anti-PAD2 antibody mAb2 only partially inhibited the activity. Clones 12 and 44 completely inhibited hPAD4 enzyme activity, whereas anti-PAD4 antibodies G8H4, H7H4, and 4R147 did not inhibit hPAD4 enzyme activity up to 50 nM.

[0309] Similarly, the potency of clone 12 was directly compared to mAbs 2, G8H4, H7H4, and 4R147 using an optimized histone-H3 citrullination ELISA in synovial fluid (Figure 26). Clone 12 completely inhibited PAD enzyme activity, whereas none of the other antibodies showed inhibitory activity.

[0310] 13 Example 8: Specificity WO 2016 / 155745(A1)

[11] suggests monoclonal antibodies that are cross-reactive against PAD2, PAD4 and PAD3.

[0311] The affinity-optimized anti-PAD4 antibodies and bispecific antibody formats were specific for PAD2 and / or PAD4 and did not bind to PAD3 (Figure 24A) or PAD1 (Figure 24B) as assessed by specificity ELISA assays (Section 6.2).

[0312] 14 Example 9: Binding behavior of bispecific antibodies The binding properties of both DuetMab orientations (PAD2λ / PAD4κ and PAD4λ / PAD2κ) were evaluated. Both DuetMab orientations had similar binding behavior. Similar to PAD4 IgG and Fab (Figures 5 and 8), Duets also stabilized PAD4 dimers (Figure 15). In contrast, but similar to PAD2 IgG and Fab (Figures 7 and 8), the binding of DuetMab to PAD2 suggests a different epitope location for PAD2 (Figure 15). Constant region modifications (YTE, TM, FQQ) did not affect the binding properties of DuetMab.

[0313] All Bis3 orientations had similar binding behavior to PAD2 and PAD4 as assessed by iSCAT (Section 6.3) (Figure 16). Similar to PAD4 IgG, the Bis3 format favored a highly stable hexameric complex, i.e., 2 × Bis3 + 2 × PAD4 dimer. Fc mutations did not affect binding behavior (Figure 16). In contrast, but similar to PAD2 IgG, Bis3 had no effect on the PAD2 dimeric state. Furthermore, steric hindrance appears to block tetramer formation (Figure 16, dotted line), as previously observed for PAD2 IgG, but in contrast to DuetMab (Figure 15).

[0314] iSCAT data showed that fibrillation was possible in the Bis3 format but not in the DuetMab (Figures 15 and 16). These data indicate that the Bis3 format with the PAD2 arm in the Fab format is less prone to form high complexes (Figure 16, "PAD2 Bis-TM-YTE 10") and therefore has a lower risk of aggregation. As expected, Fc modification did not affect binding activity (Figures 15 and 16; compare, for example, PAD2 Bis-TM-YTE 07 and PAD2 Bis-TM-YTE 11).

[0315] In the presence of PAD4 alone, the Bis3 construct, but not Duet, allowed the formation of stable hexamers as well as higher-order complexes (Figures 15 and 16). In contrast, in the presence of PAD2 alone, no hexamers or higher-order complexes were observed. Thus, dual PAD4 interactions are the primary driver of highly stable oligomers, including hexamers and potentially higher-order complexes, suggesting that "fibrillization" is possible. In the presence of both PAD4 and PAD2, DuetMab was also able to form stable hexamers, but no higher-order complexes were detected (Figures 15 and 16). The dual interactions seen in the Bis3 format are not possible in the DuetMab format (Figure 17), and therefore no higher-order complexes are observed (Table 90).

[0316] [Table 90]

[0317] 15 Example 10: Stability 15.1 Thermal stability Aggregation is closely related to the thermal stability of antibody molecules. The lower the thermal stability, the more unstable the product and the higher the degree of aggregation; on the other hand, the higher the thermal stability of the product, the lower the degree of aggregation. Promethus Nano Differential Scanning Fluorometry (Nano-DSF) was used to analyze thermal stability using standard protocols.

[0318] The YTE Fc mutations (M252Y / S254T / T256E) extend the half-life of the antibody molecule. The triple mutation (TM, L234F / L235E / P331S) attenuates the effector function of the constant region. The TM and YTE mutations were introduced into the Bis3 and DuetMab bispecific antibodies, and the sequences are provided in Tables 57 and 58. onsetStability data demonstrated that the TM-YTE Fc mutations were less stable compared to bispecific antibodies with only TM modifications, for both Bis3 and DuetMab formats (Table 91, clones 09, 10, 03, and 04). Alternative effector-null modifications (FQQ, L234F / L235Q / K322Q) did not exhibit the same instability (clones 05, 06, 11, and 12) regardless of bispecific antibody format (Figure 18).

[0319] The highest onset temperature (T onset ) was seen for the TM and FQQ-YTE DuetMab and Bis3 formats (Table 91). Clones 01, 02, 05, 06, 07, 08, 11, and 12 all had T values ​​greater than approximately 48°C. onset (FIG. 18, dotted line).

[0320] [Table 91]

[0321] 15.2 Agglomeration Accelerated stability testing using HP-SEC compared stability at 40°C and 45°C incubations to 4°C (Table 92, Figure 19). Surprisingly, the DuetMab format exhibited greater stability at both 45°C than all Bis3 formats. In the context of the TM Fc modification, the Bis3 clone 09 format exhibited lower aggregation at 45°C. In the context of the TM-YTE modification, Bis3 exhibited the highest aggregation, regardless of Bis3 orientation at 45°C. In the context of the FQQ-YTE Fc modification, the clone 11 Bis3 format exhibited the least aggregation at 45°C.

[0322] Surprisingly, at 40°C, the DuetMab format showed minimal aggregation in the context of the TM-YTE or FQQ-YTE Fc modifications, regardless of the orientation of the DuetMab. In the context of the FQQ-YTE Fc modification, the PAD2 (hole)-PAD4 (knob) orientation showed aggregation comparable to the respective IgG(TM).

[0323] For the Bis3 format, aggregation was tolerated in the context of the FQQ-YTE modification, regardless of the Bis3 orientation. The PAD2-Fab / PAD4-scFv (clones 08, 10, and 12) Bis3 format showed minimal aggregation, regardless of the Fc modification.

[0324] [Table 92] 2-week study, antibody concentration approximately 1mg / ml

[0325] 15.3 Stability Conclusions Stability data surprisingly showed that bispecific antibody formats with TM-YTE Fc modifications demonstrated reduced thermal stability that was not observed with bispecific antibodies with TM or FQQ-TM modifications (Figure 18). The stability data further demonstrated that the DuetMab format was surprisingly less prone to aggregation than the Bis3 format, regardless of the DuetMab or Bis3 orientation of the Fc modification. Furthermore, the PAD2-Fab / PAD4-scFv (clones 08, 10, and 12) Bis3 format showed minimal aggregation, regardless of Fc modification.

[0326] 16 Example 11: Safety The Bis3 format was well tolerated. In cynomolgus monkey studies, PAD2 / 4 was shown to be expressed intracellularly and on the nuclear membrane of tissue immune cells. No findings of concern were noted (Figure 20). PAD2 was detected in the CNS, but did not affect the enhancement of behavioral observations in the cynomolgus monkey study. Furthermore, no effects (e.g., adverse cytokine release) were observed in healthy blood and RA donor blood (Figure 20).

[0327] Furthermore, no safety concerns (e.g., harmful cytokine release) were observed in the blood of healthy and RA donors after exposure to Bis3 or DuetMab formats (clone 12 and clone 06) (Figure 21). None of the test samples measured cytokine levels above the upper limit of quantitation of 20,000.0 pg / ml.

[0328] 17 Example 12: Effector null mutations Constant region modifications did not affect the potency or PAD2 / PAD4 affinity of the bispecific molecules. The triple mutation (TM) (L234F / L235E / P331S) abolishes Fc effector function (Table 93). The YTE mutation (M252Y / S254T / T256E) extends the half-life of the antibody. YTE was also shown to partially rescue Fc affinity for huFcRn in the context of the TM mutations. FQQ (L234F / L235Q / K322Q) is an alternative effective null mutation for TM.

[0329] [Table 93] Affinity chromatography (Section 6.7).

[0330] References All publications mentioned herein are incorporated herein by reference. [1] ME Weinblatt et al., New England Journal of Medicine 340, 253 (1999). [2] S. Mpofu, F. Fatima, and RJ Moots, Rheumatology 44, 271 (2005). [3] EHChoy, AF Kavanaugh, and SA Jones, Nat Rev Rheumatol 9,154 (2013). [4] JL Nam, Current Opinion in Rheumatology 28, 267 (2016). [5]MHPillinger and SBABramson, Rheumatic Disease Clinics of North America 21,691(1995). [6] M. Nagashima et al., Rheumatology International 18, 113 (1998). [7] N. Thieblemont et al., Semin Immunol 28, 159 (2016). [8] KLBicker and PR Thompson, Biopolymers 99, 155 (2013). [9] M.Sato et al., International Publication No. 2012026309(A9) (May 10, 2012).

[10] CH Nielsen and D. Damgaard, International Publication No. 2014086365 (A1) (June 12, 2014).

[11] CH Nielsen and D. Damgaard, International Publication No. 2016155745 (A1) (October 6, 2016).

[12] Sato et al., International Publication No. 2016143753(A1) (September 15, 2016).

[13] M. Aosasa et al., J Immunol Res 2021, 6659960 (2021).

[14] EAKabat, Sequences of Proteins of Immunological Interest (USDepartment of Health and Human Services, Public Health Service, National Institutes of Health, 1991).

[15] P.Tsui et al., International Publication No. 2015175874(A2) (November 19, 2015).

[16] W. Dall'acqua, LS Johnson, and ES Ward, International Publication No. 2002060919 (A2) (August 8, 2002).

[17] P.Tsui, M.BORROK, and W.Dall'acqua, International Publication No. 2013165690(A1) (November 7, 2013).

[18] PT Jones et al., Nature 321,522 (1986).

[19] M. Verhoeyen, C. Milstein, and G. Winter, Science 239, 1534 (1988).

[20] L. Riechmann et al., Nature 332,323 (1988).

[21] F. RUCKER, G. Himmler, and G. Wozniak-Knopp, International Publication No. 2008003103(A2) (January 10, 2008).

[22] C. Wu, International Publication No. 2015103072(A1) (July 9, 2015).

[23] U. Brinkmann and RE Kontermann, MAbs 9, 182 (2017).

[24] Y. Mazor et al., MAbs 7, 377 (2015).

[25] V. Faid et al., Eur J Pharm Sci 159, 105730 (2021).

[26] EA Kabat and TT Wu, Ann NY Acad Sci 190,382(1971).

[27] C. Chothia and AMLesk, J Mol Biol 196, 901 (1987).

[28] SJCoales et al., Rapid Commun Mass Spectrom 23,639(2009).

[29] V. Oganesyan et al., Acta Cryst D 64,700 (2008).

[30] MJ Borrok et al., Journal of Pharmaceutical Sciences 106, 1008 (2017).

[31] PJCarter, LGPresta, and JBRidgway, U.S. Patent No. 8216805(B2) (July 10, 2012).

[32] PJCarter, LGPresta, and JBRidgway, U.S. Patent No. 5731168(A) (March 24, 1998).

[33] G. Chattopadhyay and R. Varadarajan, Protein Sci 28, 1127 (2019).

[34] SJ et al., Biotechnology and bioengineering 112, (2015).

[35] Y. Takizawa et al., Ann Rheum Dis 65, 1013 (2006).

[36] A. Kinloch et al., Arthritis Res Ther 7, R1421 (2005).

[37] V. Tj et al., Nature biotechnology 14, (1996).

[38] M.Miyamoto et al., International Publication No. 2022176970 (August 25, 2022)

Claims

1. An antibody comprising a PAD2-binding domain that specifically binds to PAD2 and / or a PAD4-binding domain that specifically binds to PAD4.

2. 2. The antibody of claim 1, wherein the antibody inhibits PAD activity, and optionally the antibody inhibits PAD-mediated citrullination of proteins.

3. Optionally, the antibody has an IC of ≦200 pM as measured by an H3 citrullination assay. ５０ The antibody of claim 2, which inhibits PAD activity in synovial fluid.

4. the antibody inhibits PAD2 activity, and optionally the antibody has an IC of ≦700 pM, optionally as measured by an H3 citrullination assay. ５０ 4. The antibody of claim 1, wherein the antibody inhibits PAD2-mediated protein citrullination, and optionally the PAD2 is recombinant PAD2.

5. the antibody inhibits PAD4 activity, and optionally the antibody has an IC of ≦100 pM, optionally as measured by an H3 citrullination assay. ５０ 5. The antibody of claim 1, wherein the antibody inhibits PAD4-mediated protein citrullination, and optionally the PAD4 is recombinant PAD4.

6. The antibody of any one of claims 1 to 5, wherein the antibody inhibits PAD2 and / or PAD4 in immune cells.

7. The antibody of claim 6 , wherein the immune cells are neutrophils and monocytes.

8. The antibody of claim 6 , wherein the immune cell is a monocyte.

9. The antibody of any one of claims 1 to 8, wherein the PAD2 is human PAD2.

10. The antibody of any one of claims 1 to 9, wherein the PAD4 is human PAD4.

11. The antibody of any one of claims 1 to 10, wherein the PAD2-binding domain does not specifically bind to PAD3, optionally as measured by PAD3 ELISA.

12. The antibody of any one of claims 1 to 11, wherein the PAD4-binding domain does not specifically bind to PAD3, optionally as measured by PAD3 ELISA.

13. The antibody of any one of claims 1 to 12, wherein the antibody does not specifically bind to PAD3, optionally as measured by PAD3 ELISA.

14. The method according to any one of claims 11 to 13, wherein the PAD3 is human PAD3.

15. The antibody of any one of claims 1 to 14, wherein the PAD2 binding domain does not specifically bind to PAD1, optionally as measured by PAD1 ELISA.

16. The antibody of any one of claims 1 to 15, wherein the PAD4 binding domain does not specifically bind to PAD1, optionally as measured by PAD1 ELISA.

17. The antibody of any one of claims 1 to 16, wherein the antibody does not specifically bind to PAD1, optionally as measured by PAD1 ELISA.

18. The antibody of any one of claims 15 to 18, wherein the PAD1 is human PAD1.

19. An antibody described in any one of claims 1 to 18, wherein the PAD2 domain specifically binds to mouse PAD2.

20. An antibody described in any one of claims 1 to 19, wherein the PAD4 domain specifically binds to mouse PAD4.

21. The antibody of any one of claims 1 to 20, wherein the PAD2 domain specifically binds to cynomolgus monkey PAD2.

22. The antibody of any one of claims 1 to 21, wherein the PAD4 domain specifically binds to cynomolgus monkey PAD4.

23. The antibody of any one of claims 1 to 18, wherein the antibody is a bispecific antibody comprising the PAD2-binding domain and the PAD4-binding domain.

24. The antibody has a K Ｄ wherein X is a bivalent Fab fragment or IgG K fragment comprising the PAD2-binding domain. Ｄ The antibody of claim 19, which is smaller.

25. The antibody has a K Ｄ wherein X is a bivalent Fab fragment or IgG K Ｄ The antibody of claim 19, which is smaller.

26. The antibody has an affinity (K Ｄ The antibody of any one of claims 19 to 21, which binds to PAD2 at the

27. The antibody has an affinity (K Ｄ The antibody of any one of claims 19 to 22, which binds to PAD4 at the nucleotide sequence (SEQ ID NO: 1).

28. The antibody has a T ｏｎｓｅｔ and optionally, T ｏｎｓｅｔ The antibody of any one of claims 19 to 23, wherein is measured by nano-differential scanning fluorimetry (DSF).

29. The antibody of any one of claims 19 to 24, wherein the antibody is a bivalent, bispecific antibody.

30. 26. The antibody of claim 25, wherein the antibody is Bis3.

31. The antibody of claim 25 or 26, wherein the antibody comprises two PAD2 binding domains such that the antibody is bivalent for PAD2, and the antibody comprises two PAD4 binding domains such that the antibody is bivalent for PAD4.

32. An antibody according to any one of claims 25 to 27, wherein each PAD2 binding domain is contained in a Fab domain and each PAD4 binding domain is contained in an scFv.

33. An antibody according to any one of claims 25 to 27, wherein each PAD4 binding domain is contained in a Fab domain and each PAD2 binding domain is contained in an scFv.

34. The antibody has an affinity (K Ｄ 30. The antibody of any one of claims 25 to 29, which binds to human PAD2 at the nucleotide sequence (SEQ ID NO: 1).

35. The antibody has an affinity (K Ｄ 31. The antibody of claim 25 or 30, which binds to human PAD4 at the amino acid sequence α, β ...

36. The antibody a) an IgG comprising first and second Fab domains and an Fc domain, wherein the first and second Fab domains each comprise a PAD2-binding domain that specifically binds to PAD2; and b) a first and a second scFv, each of which is linked to the carboxy terminus of one of the heavy chains of the Fc domain of the IgG, and each of which comprises a PAD4 binding domain that specifically binds to PAD4.

37. The antibody of claim 32, wherein the PAD4 binding domains of the first and second scFvs comprise sequence number 39.

38. The antibody of claim 32 or 33, wherein the PAD2 binding domains of the first and second Fabs comprise a heavy chain domain comprising SEQ ID NO:

34.

39. The antibody of any one of claims 32 to 34, wherein the PAD2 binding domains of the first and second Fabs comprise a light chain constant domain comprising SEQ ID NO:

36.

40. The antibody of any one of claims 32 to 35, wherein the PAD2-binding domains of the first and second Fabs comprise a light chain domain comprising SEQ ID NO:

35.

41. The antibody a) an IgG comprising first and second Fab domains and an Fc domain, wherein the first and second Fab domains each comprise a PAD4 binding domain that specifically binds to PAD4; and b) a first and a second scFv, each of which is linked to the carboxy terminus of one of the heavy chains of the Fc domain of the IgG, and each of which comprises a PAD2-binding domain that specifically binds to PAD2.

42. The antibody of claim 37, wherein the PAD2 binding domains of the first and second scFvs comprise sequence number 38.

43. The antibody of claim 37 or 38, wherein the PAD4 binding domains of the first and second Fabs comprise a heavy chain domain comprising SEQ ID NO:

40.

44. The antibody of any one of claims 37 to 39, wherein the PAD4 binding domains of the first and second Fabs comprise a light chain domain comprising SEQ ID NO:

41.

45. The antibody of any one of claims 37 to 40, wherein the first and second Fab domains comprise a heavy chain constant domain comprising SEQ ID NO:

37.

46. 42. The antibody of any one of claims 37 to 41, wherein the first and second Fab domains comprise a light chain constant domain comprising SEQ ID NO:

36.

47. The antibody of any one of claims 32 to 42, wherein the scFv is linked to the carboxy terminus of the heavy chain by a peptide linker.

48. 44. The antibody of claim 43, wherein the peptide linker comprises SEQ ID NO:

51.

49. The antibody of claim 43 or 44, wherein the first and / or second scFv comprises a VH-VL linker domain comprising SEQ ID NO:

33.

50. The antibody of any one of claims 19 to 24, wherein the antibody is a monovalent bispecific antibody.

51. 47. The antibody of claim 46, wherein the antibody is a DuetMab.

52. The antibody of claim 46 or 47, wherein the antibody comprises one PAD2 binding domain such that the antibody is monovalent for PAD2, and the antibody comprises one PAD4 binding domain such that the antibody is monovalent for PAD4.

53. The antibody comprises an IgG, and the IgG comprises a) a first binding region comprising a first Fab, wherein a second Fab domain comprises the PAD2 binding domain; and b) a second binding region comprising a second Fab, wherein the second Fab comprises the PAD4 binding domain.

54. 50. The antibody of claim 49, wherein the IgG comprises a kappa light chain comprising SEQ ID NO:

63.

55. 51. The antibody of claim 49 or 50, wherein the IgG comprises a lambda light chain comprising SEQ ID NO:

64.

56. 50. The antibody of claim 49, wherein the IgG comprises a kappa light chain comprising SEQ ID NO:

62.

57. 53. The antibody of claim 49 or 52, wherein the IgG comprises a lambda light chain comprising SEQ ID NO:

65.

58. The antibody is an IgG or F(ab') ２ An antibody according to any one of claims 1 to 18, including a fragment.

59. 55. The antibody of claim 54, wherein the antibody is an IgG1.

60. 56. The antibody of claim 54 or 55, wherein the antibody comprises two of the PAD2 binding domains such that the antibody is bivalent for PAD2.

61. 56. The antibody of claim 54 or 55, wherein the antibody comprises two of the PAD4 binding domains such that the antibody is bivalent for PAD4.

62. The antibody of any one of claims 1 to 18, wherein the antibody comprises a Fab fragment, and the Fab fragment comprises the PAD2-binding domain or the PAD4-binding domain.

63. The Fab fragment comprises the PAD2-binding domain, and the Fab has an affinity (K Ｄ 59. The antibody of claim 58, which binds to human PAD2 at the SEQ ID NO:

1.

64. the Fab fragment comprises the PAD4-binding domain, and the Fab has an affinity (K Ｄ 59. The antibody of claim 58, which binds to human PAD4 at the SEQ ID NO:

1.

65. The above K Ｄ 61. The antibody of any one of claims 22, 23, 30, 31, 59 or 60, wherein is measured by Biacore.

66. the PAD2-binding domain comprises a variable heavy (VH) domain sequence comprising complementarity-determining regions (CDRs) HCDR1, HCDR2, and HCDR3, and a variable light (VL) domain sequence comprising CDRs LCDR1, LCDR2, and LCDR3; a) the HCDR1 amino acid sequence is SEQ ID NO: 3, optionally with 1, 2 or 3 amino acid changes; b) the HCDR2 amino acid sequence is SEQ ID NO: 4, optionally with 1, 2 or 3 amino acid changes; c) the HCDR3 amino acid sequence is SEQ ID NO: 5, optionally with 1, 2 or 3 amino acid changes; d) the LCDR1 amino acid sequence is SEQ ID NO: 10, optionally with 1, 2 or 3 amino acid changes; e) the LCDR2 amino acid sequence is SEQ ID NO: 11, optionally with 1, 2 or 3 amino acid changes; and / or f) The antibody of any one of claims 1 to 65, wherein the LCDR3 amino acid sequence is SEQ ID NO: 12, optionally with 1, 2 or 3 amino acid changes.

67. The antibody of any one of claims 1 to 66, wherein the PAD2-binding domain comprises a VH domain comprising a sequence having at least 90% identity to SEQ ID NO:

1.

68. The antibody of any one of claims 1 to 67, wherein the PAD2-binding domain comprises a VH domain comprising SEQ ID NO:

1.

69. The antibody of any one of claims 1 to 68, wherein the PAD2-binding domain comprises a VL domain comprising a sequence having at least 90% identity to SEQ ID NO:

2.

70. The antibody of any one of claims 1 to 69, wherein the PAD2 binding domain comprises a VL domain sequence comprising SEQ ID NO:

2.

71. 71. The antibody of any one of claims 1 to 70, wherein the PAD2 binding domain comprises a VH domain sequence comprising SEQ ID NO: 1, and optionally has 1, 2, 3, 4 or 5 amino acid changes outside the CDRs.

72. 72. The antibody of any one of claims 1 to 71, wherein the PAD2 binding domain comprises a VL domain sequence comprising SEQ ID NO: 2, and optionally has 1, 2, 3, 4 or 5 amino acid changes outside the CDRs.

73. the PAD4-binding domain comprises a variable heavy (VH) domain sequence comprising complementarity-determining regions (CDRs) HCDR1, HCDR2, and HCDR3, and a variable light (VL) domain sequence comprising CDRs LCDR1, LCDR2, and LCDR3; a) the HCDR1 amino acid sequence is SEQ ID NO: 17; b) the HCDR2 amino acid sequence is SEQ ID NO: 18; c) the HCDR3 amino acid sequence is SEQ ID NO: 19; d) the LCDR1 amino acid sequence is SEQ ID NO: 24; e) the LCDR2 amino acid sequence is SEQ ID NO: 25, and / or f) An antibody according to any one of claims 1 to 72, wherein the LCDR3 amino acid sequence is SEQ ID NO:

26.

74. The antibody of any one of claims 1 to 73, wherein the PAD4 binding domain comprises a VH domain comprising a sequence having at least 90% identity to SEQ ID NO:

31.

75. The antibody of any one of claims 1 to 74, wherein the PAD4 binding domain comprises a VH domain sequence comprising SEQ ID NO:

31.

76. The antibody of any one of claims 1 to 75, wherein the PAD4-binding domain comprises a VL domain comprising a sequence having at least 90% identity to SEQ ID NO:

32.

77. The antibody of any one of claims 1 to 76, wherein the PAD4 binding domain comprises a VL sequence domain comprising SEQ ID NO:

32.

78. 78. The antibody of any one of claims 1 to 77, wherein the PAD4 binding domain comprises a VH domain sequence comprising SEQ ID NO: 31, and optionally has 1, 2, 3, 4 or 5 amino acid changes outside the CDRs.

79. 79. The antibody of any one of claims 1 to 78, wherein the PAD4 binding domain comprises a VL domain sequence comprising SEQ ID NO: 32, and optionally has 1, 2, 3, 4 or 5 amino acid changes outside the CDRs.

80. The antibody of any one of claims 1 to 79, wherein the antibody comprises an Fc domain, optionally an IgG1 Fc domain.

81. The antibody of claim 76, wherein the Fc domain has null effector function.

82. 78. The antibody of claim 77, wherein the Fc domain comprises the mutations L234F, L235Q, and K322Q.

83. 78. The antibody of claim 77, wherein the Fc domain comprises the mutations L234F, L235E, and P331S.

84. 80. The antibody of any one of claims 76 to 79, wherein the Fc domain comprises at least one mutation that confers half-life extension.

85. 81. The antibody of claim 80, wherein the Fc domain comprises the mutations M252Y, S254T, and T256E.

86. 82. The antibody of any one of claims 76 to 81, wherein the Fc domain has null effector function and comprises at least one mutation that confers half-life extension.

87. 83. The antibody of claim 82, wherein the Fc domain comprises the mutations L234F, L235Q, K322Q, M252Y, S254T, and T256E.

88. 83. The antibody of claim 82, wherein the Fc domain comprises the mutations L234F, L235E, P331S, M252Y, S254T, and T256E.

89. The Fc domain comprises SEQ ID NO: 47, SEQ ID NO: 48, or SEQ ID NO:

49. Ｈ 77. The antibody of claim 76, comprising two domains.

90. 32. The antibody of any one of claims 1 to 31, comprising the sequence of SEQ ID NO:53, SEQ ID NO:54, SEQ ID NO:55, SEQ ID NO:56, SEQ ID NO:57, or SEQ ID NO:

58.

91. a) the PAD2-binding domain comprises a heavy chain comprising SEQ ID NO: 73 and a light chain comprising SEQ ID NO: 62; b) The antibody of claim 47, wherein the PAD4 binding domain comprises a heavy chain comprising SEQ ID NO: 76 and a light chain comprising SEQ ID NO:

65.

92. a) the PAD2-binding domain comprises a heavy chain comprising SEQ ID NO: 74 and a light chain comprising SEQ ID NO: 62; b) The antibody of claim 47, wherein the PAD4 binding domain comprises a heavy chain comprising SEQ ID NO: 77 and a light chain comprising SEQ ID NO:

65.

93. a) the PAD2-binding domain comprises a heavy chain comprising SEQ ID NO: 59 or SEQ ID NO: 60 and a light chain comprising SEQ ID NO: 62; b) The antibody of claim 47, wherein the PAD4 binding domain comprises a heavy chain comprising SEQ ID NO: 61 and a light chain comprising SEQ ID NO:

65.

94. a) the PAD2-binding domain comprises a heavy chain comprising SEQ ID NO: 70 and a light chain comprising SEQ ID NO: 64; b) The antibody described in claim 47, wherein the PAD4 binding region comprises a heavy chain comprising SEQ ID NO: 67 and a light chain comprising SEQ ID NO:

63.

95. a) the PAD2-binding domain comprises a heavy chain comprising SEQ ID NO: 71 and a light chain comprising SEQ ID NO: 64; b) The antibody described in claim 47, wherein the PAD4 binding region comprises a heavy chain comprising SEQ ID NO: 68 and a light chain comprising SEQ ID NO:

63.

96. a) the PAD2-binding domain comprises a heavy chain comprising SEQ ID NO: 72 and a light chain comprising SEQ ID NO: 64; b) The antibody described in claim 47, wherein the PAD4 binding region comprises a heavy chain comprising SEQ ID NO: 69 and a light chain comprising SEQ ID NO:

63.

97. A polypeptide comprising an antibody according to any one of claims 1 to 96.

98. A nucleic acid encoding one or more chains of an antibody according to any one of claims 1 to 92.

99. A nucleic acid encoding the polypeptide of claim 93.

100. A vector comprising the nucleic acid of claim 94 or 95.

101. 97. A host cell comprising the vector of claim 96.

102. A pharmaceutical composition comprising the antibody of any one of claims 1 to 92 and a pharmaceutically acceptable carrier.

103. A kit comprising an antibody according to any one of claims 1 to 92 or a pharmaceutical composition according to claim 98, and optionally including instructions for use.

104. 93. A method of treating a disease in a subject, comprising administering to the subject an antibody according to any one of claims 1 to 92.

105. A method of treating a disease in a subject, comprising administering to the subject an anti-PAD4 antibody in combination with an anti-PAD2 antibody.

106. The method of claim 101, wherein the anti-PAD4 antibody is the antibody of claim 69 and the anti-PAD2 antibody is the antibody of claim 62.

107. 103. The method of claim 101 or 102, wherein the anti-PAD2 antibody and the anti-PAD4 antibody are administered to the subject simultaneously, separately, or sequentially.

108. 104. The method of any one of claims 100 to 103, wherein the disease is an autoimmune disorder.

109. 105. The method of claim 104, wherein the autoimmune disorder is rheumatoid arthritis.

110. An antibody according to any one of claims 1 to 92 or a pharmaceutical composition according to claim 98 for use in a method for treating or preventing a disease in a subject.

111. The antibody for use according to claim 1, wherein the method comprises the method according to any one of claims 100 to 105.

112. 99. Use of an antibody according to any one of claims 1 to 92 or a pharmaceutical composition according to claim 98 for the manufacture of a medicament for the treatment of an autoimmune disorder.

Images