Treatment of autoimmune diseases

Abstract

Antibodies for binding PAD and their use in therapy.

Description

Background Art

[0001] Rheumatoid arthritis (RA) is a common autoimmune disease with a chronic, progressive phenotype. It is a chronic, systemic inflammatory disease that affects both small and large joints, leading to progressive joint destruction, loss of function, chronic pain / fatigue, and increased disability.

[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-α), 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 a disease-modifying antirheumatic drug (DMARD) such as methotrexate, followed by cycles of targeted therapy such as newer biologics based on a "treat to target" approach that attempts to achieve remission, low disease activity, and / or ACR70 (70% improvement in activity). However, these approaches have had limited success, with only 20%-30% of patients achieving ACR70. There is a clear need for more effective therapies for the treatment of RA, ideally targeting the pathogenic pathways of the disease.

[0003] Histologically, RA is characterized by synovial inflammation with infiltration of T and B cells, often organized into germinal center-like structures, as well as macrophages, dendritic cells, and neutrophils [5-7], the latter of which are particularly abundant in the synovial fluid early in the disease. Although these histologic features differ little from those of other autoimmune or inflammatory conditions, it has become apparent that RA possesses unique features that cannot be simply explained by the traditional T cell-centric hypothesis of pathogenesis, in which the initial event is a loss of tolerance.

[0004] Peptidylarginine deiminases (PADs) are a family of five isoenzymes (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 called 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] WO 2012026309 A9 [9] describes anti-PAD4 antibodies.

[0006] WO 2014 / 086365 A1

[10] describes antibodies for binding to rabbit PAD2 (rPAD2).

[0007] WO 2016 / 155745 A1

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

[0008] WO 2016143753 A1

[12] describes anti-PAD4 antibodies.

[0009] Aosasa et al., 2021 described chimeric anti-PAD2 antibodies

[13] .

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

[0011] The present invention relates to anti-PAD4 antibodies with high affinity and specificity for human and cynomolgus monkey PAD4. The present invention relates to anti-PAD2 antibodies with 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 effective in inhibiting PAD4 and / or PAD2 activity and in inhibiting PAD activity in the synovial fluid of patients with rheumatoid arthritis (RA).

[0014] The present invention relates to treating autoimmune diseases by targeting both PAD2 and PAD4. The present invention particularly relates to the use of a combination of anti-PAD2 and anti-PAD4 antibodies or anti-PAD2 / 4 bispecific antibodies in treating autoimmune diseases characterized by elevated PAD activity, such as RA. The present invention is supported by data presented herein for the first time, which show that the activities of PAD2 and PAD4 in RA disease are non-redundant. Surprisingly, it was shown that targeting both PAD2 and PAD4 is necessary and sufficient to eliminate PAD activity in whole blood, serum, and synovial fluid of RA patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 : PAD2- and PAD4-dependent production of citrullinated RA antigens

[0016] ELISA measures the production of citrullinated antigens (fibrinogen β chain and α-enolase) by PAD2 and PAD4.

[0017] Figure 2: Efficacy of prior art anti-PAD4 antibodies

[0018] H3 histone citrullination assay ( Figure 2A ) or BAEE PAD activity assay (Cayman Chemical)( Figure 2B ) The potency of the prior art humanized anti-PAD4 antibodies evaluated.

[0019] Figure 3: PAD2 and PAD4 expression

[0020] 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 median. Median of PAD2 in HD < LLOD. Mann-Whitney test. ****p < 0.0001. PAD measured with Cayman ELISA kit. Figure 3B : High PAD2 and PAD4 protein levels in RA synovial fluid. RA synovial fluid (SF) n = 5. Bars indicate median. Synovial fluid diluted and PAD levels measured with Cayman ELISA kit. Figure 3C : RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkey (cyno). Figure 3D : RNA expression profiles of PAD4 in immune cells from human and cynomolgus monkey (cyno). Expression of PAD2 and PAD4 shown relative to expression in human B cells.

[0021] Figure 4: Cell surface quantification of PAD2 and PAD4

[0022] Quantification of PAD2 ( Figure 4B ) and PAD4 ( Figure 4A ) on the surface of immune cells.

[0023] Figure 5 : PAD4 and anti-PAD4 Fab

[0024] Interference scattering of individual protein molecules (iSCAT) in close proximity to the surface. Anti-PAD4 = clone 42 (48LO0063, IgG or Fab).

[0025] Figure 6 : PAD4 and anti-PAD4 IgG

[0026] Interference scattering of individual protein molecules (iSCAT) in close proximity to the surface. Anti-PAD4 = clone 42 (48LO0063, IgG or Fab).

[0027] Figure 7 : PAD2 and anti-PAD2 Fab / IgG

[0028] PAD2 and iSCAT of anti-PAD2 Fab and IgG. Anti-PAD2 = clone 22 (IgG or Fab).

[0029] Figure 8: iSCAT Overview

[0030] Figure 8A :PAD4; Figure 8B :PAD2.

[0031] Figure 9: Potency Assay Optimization

[0032] Figure 9A : Optimization of potency assay parameters. Figure 9B : PAD4 Fab Histone H3 Assay. PAD4 incubation for 3 hours and 45 minutes. [PAD4] = 15 pg.ml.

[0033] Figure 10: Bispecific Format

[0034] Figure 10A : Monovalent Duet mAb. Figure 10B : Bivalent bispecific Bis3.

[0035] Figure 11 : Histone-H3 activity assay.

[0036] The histone H3 substrate is coated on a plate, where active PAD2 in the sample deaminates arginine residues to form citrulline. These citrullinated epitopes are then detected by standard immunoassays. This assay can be used to demonstrate target engagement in the circulation and in the synovial cavity (where PAD2 activity is high) without sample dilution.

[0037] Figure 12: Inhibition of PAD activity

[0038] Figure 12A : Inhibition of PAD2 in synovial fluid (1000-fold dilution). Figure 12B : Inhibition in neutrophil supernatant (250-fold dilution). PAD activity in diluted synovial fluid samples was measured using the histone-H3 PAD activity assay. EDTA sequesters calcium and inhibits PAD activity and serves as a background control.

[0039] Figure 13 :PAD2 antibody affinity

[0040] Binding affinity K of anti-PAD2 monoclonal antibody (as Fab) D (nM). Biotinylated PAD2 was captured on a CM5 / C1-streptavidin surface. Affinity: K D (nM). Data shown are mean values, n=2-9 experiments (Table 72).

[0041] Figure 14 :PAD4 antibody affinity

[0042] Binding affinity K of anti-PAD4 monoclonal antibody (as Fab) D (nM). Biotinylated PAD4 was captured on a CM5 / C1-streptavidin surface. Affinity: K D (nM). The data shown are the average of n=1-6 experiments (Table 74).

[0043] Figure 15 :DuetMab PAD2 / PAD4 binding behavior

[0044] DuetMab's iSCAT for PAD2 / PAD4 bispecific antibody.

[0045] Figure 16 :Bis3 PAD2 / PAD4 binding behavior

[0046] iSCAT of the Bis3 PAD2 / PAD4 bispecific antibody.

[0047] Figure 17 :DuetMab combined with Bis3

[0048] Schematic diagram showing the differences between Bis3 and DuetMab formats in binding to PAD2 and PAD4 dimers.

[0049] Figure 18 :The effect of Fc modification on thermal stability

[0050] The thermal stability of the bispecific format in the context of different Fc modifications was assessed by Nano-DSF. The T values of clones 01-12 are shown. 起始 Value, (℃) (Table 92).

[0051] Figure 19: Acceleration stability

[0052] HP-SEC was used to evaluate the effects of DuetMab and Bis3 formats at 40°C ( Figure 19B ) and 45℃( Figure 19A ) tends to aggregate under the

[0053] Figure 20 :Immunohistochemistry in Cynomolgus Monkey Studies

[0054] Figure 21 : In vitro cytokine release, plate-bound antibodies

[0055] Cytokine expression following exposure to antibodies in Bis3 (clone 12) or DuetMab (clone 06) format (HD = high dose).

[0056] Figure 22: Validity of the Bis3 format

[0057] Figure 22A : Research plan. Figure 22B : PAD activity in plasma measured with a histone H3 citrullination assay from day 0 to day 36. Figure 22C : Endogenous PAD activity in plasma measured with a histone H3 citrullination assay from day 0 to day 36. Figure 22D : PAD activity in plasma measured with a histone H3 citrullination assay from day 0 to day 106. Figure 22E Endogenous PAD activity in plasma measured using a histone H3 citrullination assay from day 0 to day 106. LD: low dose. HD: high dose. Abatacept: Abatacept. Format: Bis3 (scFv PAD4; Fab: PAD2; Name: Clone 12). Numbers refer to individual experiments.

[0058] Figure 23: Efficacy of anti-PAD2 and anti-PAD4 antibodies against recombinant PAD

[0059] The potency of clones 12, 22, and 42 was directly compared to that of art anti-PAD2 and anti-PAD4 antibodies using an optimized histone-H3 citrullination ELISA and recombinant PAD2 and PAD4 ( Figure 23A and Figure 23B ).

[0060] Figure 24: PAD2 / PAD4 Specificity

[0061] Affinity-optimized anti-PAD4 antibodies and bispecific formats are specific for PAD2 and / or PAD4 and do not bind PAD3 ( Figure 24A ). Affinity-optimized clones do not bind to PAD1 ( Figure 24B ).

[0062] Figure 25: Comparative potency assays in Bis3 and DuetMab formats

[0063] PAD activity was measured using 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 whole blood sample. Bis3 clone = clone 12. DuetMab clone = clone 06.

[0064] Figure 26: Efficacy of anti-PAD2 and anti-PAD4 antibodies in RA synovial fluid

[0065] PAD activity was measured using 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

[0066] Any composition or method provided herein can be combined with one or more of any other compositions and methods provided herein.

[0067] 4.1 Sequence

[0068] The antibodies or polypeptides of the present invention may comprise an amino acid sequence as provided in Tables 1 to 58. The antibodies may have an amino acid sequence as provided in any one of Tables 1 to 58 (VH, VL, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, LCDR3, complete sequence, Fab, scFv, constant light chain (C L ), heavy chain (HC), light chain (LC), CH1, CH2, CH3).

[0069] Table 1: Anti-PAD2 (clone 22) - variable (VH / VL) (141LO0035hIgG1pgl-4)

[0070]

[0071]

[0072] Table 2: Anti-PAD4 (clone 42) - variable (VH / VL) (48LO0063hIgG1fgl-59)

[0073]

[0074] Table 3: 48LO0010hIgG1ngl-2 (anti-PAD4)

[0075]

[0076] Table 4: 48LO0032hIgG1ngl-2 (anti-PAD4)

[0077]

[0078]

[0079] Table 5: 48LO0033hIgG1ngl-3 (anti-PAD4)

[0080]

[0081]

[0082] Table 6: 48LO0036hIgG1ngl-3 (anti-PAD4)

[0083]

[0084]

[0085] Table 7: 48LO0040hIgG1ngl-3 (anti-PAD4)

[0086]

[0087]

[0088] Table 8: 48LO0048hIgG1ngl-3 (anti-PAD4)

[0089]

[0090]

[0091] Table 9: 48LO0049hIgG1ngl-3 (anti-PAD4)

[0092]

[0093]

[0094] Table 10: 48LO0049hIgG1pgl-9 (anti-PAD4)

[0095]

[0096] Table 11: 48LO0049hIgG1pgl-10 (anti-PAD4)

[0097]

[0098] Table 12: 48LO0049hIgG1pgl-12 (anti-PAD4)

[0099]

[0100]

[0101] Table 13: 48LO0049hIgG1fgl-23 (anti-PAD4)

[0102]

[0103] Table 14: 48LO0049hIgG1fgl-25 (anti-PAD4)

[0104]

[0105]

[0106] Table 15: 48LO0049hIgG1pgl-31 (anti-PAD4)

[0107]

[0108] Table 16: 48LO0051hIgG1ngl-2 (anti-PAD4)

[0109]

[0110] Table 17: 48LO0060hIgG1ngl-3 (anti-PAD4)

[0111]

[0112]

[0113] Table 18: 48LO0062hIgG1ngl-3 (anti-PAD4)

[0114]

[0115] Table 19: 48LO0063hIgG1ngl-3 (anti-PAD4)

[0116]

[0117] Table 20: 48LO0063hIgG1fgl-4 (anti-PAD4)

[0118]

[0119]

[0120] Table 21: 48LO0063hIgG1fgl-6 (anti-PAD4)

[0121]

[0122] Table 22: 48LO0063hIgG1fgl-7 (anti-PAD4)

[0123]

[0124] Table 23: 48LO0063hIgG1fgl-8 (anti-PAD4)

[0125]

[0126]

[0127] Table 24: 48LO0063hIgG1fgl-9 (anti-PAD4)

[0128]

[0129] Table 25: 48LO0063hIgG1fgl-11 (anti-PAD4)

[0130]

[0131] Table 26: 48LO0063hIgG1pgl-38 (anti-PAD4)

[0132]

[0133]

[0134] Table 27: 48LO0063hIgG1pgl-39 (anti-PAD4)

[0135]

[0136] Table 28: 48LO0063hIgG1pgl-40 (anti-PAD4)

[0137]

[0138] Table 29: 48LO0063hIgG1pgl-41 (anti-PAD4)

[0139]

[0140]

[0141] Table 30: 48LO0063hIgG1pgl-42 (anti-PAD4)

[0142]

[0143] Table 31: 48LO0063hIgG1pgl-43 (anti-PAD4)

[0144]

[0145] Table 32: 48LO0063hIgG1pgl-44 (anti-PAD4)

[0146]

[0147]

[0148] Table 33: 48LO0063hIgG1pgl-45 (anti-PAD4)

[0149]

[0150] Table 34: 48LO0063hIgG1pgl-46 (anti-PAD4)

[0151]

[0152] Table 35: 48LO0063hIgG1pgl-47 (anti-PAD4)

[0153]

[0154]

[0155] Table 36: 48LO0063hIgG1pgl-49 (anti-PAD4)

[0156]

[0157]

[0158] Table 37: 48LO0063hIgG1pgl-51 (anti-PAD4)

[0159]

[0160] Table 38: 48LO0063hIgG1pgl-52 (anti-PAD4)

[0161]

[0162] Table 39: 48LO0063hIgG1pgl-53 (anti-PAD4)

[0163]

[0164]

[0165] Table 40: 48LO0063hIgG1fgl-58 (anti-PAD4)

[0166]

[0167] Table 41: 48LO0063hIgG1fgl-60 (anti-PAD4)

[0168]

[0169] Table 42: 48LO0063hIgG1fgl-61 (anti-PAD4)

[0170]

[0171]

[0172] Table 43: 141LO0002hIgG1pgl-3 (anti-PAD2)

[0173]

[0174] Table 44: 141LO0002hIgG1pgl-4 (anti-PAD2)

[0175]

[0176]

[0177] Table 45: 141LO0002hIgG1ngl-2 (anti-PAD2)

[0178]

[0179] Table 46: 141LO0030hIgG1pgl-4 (anti-PAD2)

[0180]

[0181] Table 47: 141LO0030hIgG1ngl-2 (anti-PAD2)

[0182]

[0183]

[0184] Table 48: 141LO0035hIgG1ngl-2 (anti-PAD2)

[0185]

[0186] Table 49: 141LO0039hIgG1pgl-4 (anti-PAD2)

[0187]

[0188] Table 50: 141LO0039hIgG1ngl-2 (anti-PAD2)

[0189]

[0190]

[0191] Table 51: 141LO0055hIgG1ngl-2 (anti-PAD2)

[0192]

[0193] Table 52: PAD40119hIgG1ngl-2 (anti-PAD2)

[0194]

[0195]

[0196] Table 53: PAD40141hIgG1ngl-2 (anti-PAD2)

[0197]

[0198]

[0199] Table 54: PAD40175hIgG1ngl-2 (anti-PAD2)

[0200]

[0201] Table 55: AB1630204 (mouse anti-PAD4)

[0202]

[0203]

[0204] Table 56: AB1630205 (mouse anti-PAD4)

[0205]

[0206] Table 57: Bivalent bispecific Bis3 (clone 07-12)

[0207]

[0208]

[0209]

[0210]

[0211]

[0212]

[0213]

[0214]

[0215]

[0216] Table 58: Monovalent bispecific antibody DuetMab (clone 01-06)

[0217]

[0218]

[0219]

[0220]

[0221]

[0222]

[0223]

[0224] Table 59: Bis3 Control (R347)

[0225]

[0226]

[0227] Table 60: DuetMab Control (R347)

[0228]

[0229]

[0230] Table 61: Fc modified sequences

[0231] SEQ ID NO: Fc variant domain sequence (positions 432 to 437) 187 CXRHXC 188 CRRHXC 189 CXRHRC

[0232] Table 62: PAD40048

[0233]

[0234]

[0235] 4.2 Sequence identity

[0236] Table 63: PAD4 sequence identity %

[0237]

[0238]

[0239] The regions are defined by Kabat;

[0240] 4.3 Summary

[0241] 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 protein citrullination. The antibody may inhibit PAD activity in synovial fluid. The antibody may have an IC as measured by an H3 citrullination assay. 50 ≤200 pM. The antibody may be a human antibody.

[0242] Specific PAD2 binding can be measured by PAD2 ELISA. Specific PAD4 binding can be measured by PAD4 ELISA.

[0243] The antibody may have an H3 citrullination assay as measured by an H3 citrullination assay of about 700 pM, 650 pM, 600 pM, 550 pM, 540 pM, 530 pM, 520 pM, 500 pM, 480 pM, 460 pM, 440 pM, 450 pM, 430 pM, 420 pM, 400 pM, 380 pM, 360 pM, 340 pM, 320 pM, 300 pM, 280 pM, 260 pM, 240 pM, IC of 220pM, 200pM, 180pM, 160pM, 140pM, 120pM, 100pM, 90pM, 80pM, 70pM, 60pM, 50pM, 40pM, 30pM, 20pM, 19pM, 18pM, 17pM, 16pM, 15pM, 14pM, 13pM, 12pM, 11pM, 10pM, 9pM, 8pM, 7pM, 6pM, 4pM, 2pM, or 1pM 50 IC 50 May be related to the inhibition of PAD4 activity. 50 May be related to the inhibition of PAD2 activity. 50 May be associated with inhibition of the combined activity of PAD2 and PAD4.

[0244] The antibody can inhibit PAD2 activity. The antibody can inhibit PAD2-mediated protein citrullination. The antibody can inhibit PAD activity as measured by H3 citrullination assay, IC 50The antibody may have an H3 citrullination assay as measured by an H3 citrullination assay of about 700 pM, 650 pM, 600 pM, 550 pM, 540 pM, 530 pM, 520 pM, 500 pM, 480 pM, 460 pM, 440 pM, 450 pM, 430 pM, 420 pM, 400 pM, 380 pM, 360 pM, 340 pM, 320 pM, 300 pM, 280 pM, 260 pM, 240 pM, IC of 220pM, 200pM, 180pM, 160pM, 140pM, 120pM, 100pM, 90pM, 80pM, 70pM, 60pM, 50pM, 40pM, 30pM, 20pM, 19pM, 18pM, 17pM, 16pM, 15pM, 14pM, 13pM, 12pM, 11pM, 10pM, 9pM, 8pM, 7pM, 6pM, 4pM, 2pM, or 1pM 50 IC 50 May be related to the inhibition of PAD4 activity. 50 May be related to the inhibition of PAD2 activity. 50 May be associated with inhibition of the combined activity of PAD2 and PAD4.

[0245] IC 50 Can be measured by trypsin cleavage assay. IC 50 It can be measured by BAEE (Nα-benzoyl-L-arginine ethyl ester hydrochloride) citrullination assay. 50 Can be measured by H3 citrullination assay.

[0246] The antibody can inhibit PAD4 activity, optionally wherein the antibody inhibits PAD4-mediated protein citrullination, optionally as measured by an H3 citrullination assay, IC 50 ≤100 pM, and optionally wherein the PAD4 is recombinant PAD4.

[0247] This antibody can inhibit PAD2 in immune cells. This antibody can inhibit PAD4 in immune cells. This antibody can inhibit both PAD2 and PAD4 in immune cells. This antibody can inhibit PAD activity in immune cells. The immune cells can be neutrophils. The immune cells can be monocytes. The immune cells can be both neutrophils and monocytes.

[0248] The PAD2 inhibited by the antibody may be human, cynomolgus monkey or mouse PAD2. The PAD4 inhibited by the antibody may be human, cynomolgus monkey or mouse PAD4.

[0249] The antibody can comprise 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, e.g., as provided in Table 1, Table 2, Table 57, and Table 58.

[0250] 4.4 Specificity

[0251] 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 can be measured by a PAD3 ELISA. PAD1 binding can be measured by a PAD1 ELISA. The antibody may not specifically bind to PAD3 or PAD1. PAD3 may be human, cynomolgus macaque, or mouse PAD3. PAD1 may be human, cynomolgus macaque, or mouse PAD1. The antibody may specifically bind to PAD4, but not specifically to PAD1, PAD2, or PAD3. The antibody may specifically bind to PAD2, but not specifically to PAD1, PAD4, or PAD3. The antibody may specifically bind to PAD2 and PAD4, but not specifically to PAD1 or PAD3.

[0252] This antibody can specifically bind to mouse PAD2. This antibody can specifically bind to mouse PAD4. This antibody may not specifically bind to PAD2. This antibody may not specifically bind to PAD4.

[0253] 4.5 Bispecific Antibodies

[0254] The antibody may be a bispecific antibody comprising a PAD2 binding domain and a PAD4 binding domain. The PAD2 binding domain can specifically bind to PAD2 but does not specifically bind to PAD1, PAD4, or PAD3. The PAD4 binding domain can specifically bind to PAD4 but does not specifically bind to PAD1, PAD3, or PAD2.

[0255] 4.6PAD2 affinity

[0256] The bispecific antibody can bind to human PAD2 with an affinity (K D ) and the affinity of a bivalent Fab fragment or IgG containing the same PAD2 binding domain for human PAD2 (K D The bispecific antibody can bind to human PAD2 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD2 binding domain to human PAD2 D ) within ±5 pM. The bispecific antibody can bind to PAD2 with an affinity (K D) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD2 binding domain to human PAD2 D ) within ±10pM.

[0257] The K of the antibody or bispecific antibody against human PAD D The K value of the bivalent Fab fragment of IgG containing the same PAD2 binding domain for human PAD2 may be less than D This antibody has a K D The antibody may have an affinity for human PAD2 of about 1 pM, 2 pM, 3 pM, 4 pM, 5 pM, 6 pM, 7 pM, 8 pM, 9 pM, 10 pM, 11 pM, 12 pM, 13 pM, 14 pM, 15 pM, 16 pM, 17 pM, 18 pM, 19 pM, 20 pM, 21 pM, 22 pM, 23 pM, 24 pM, 25 pM, 26 pM, 27 pM, 28 pM, 29 pM, 30 pM (K D The antibody may have an affinity (K of about 10 pM, 20 pM, 30 pM, 40 pM, 50 pM, 60 pM, 70 pM, 80 pM, 90 pM, 100 pM, 110 pM, 120 pM, 130 pM, 140 pM, 150 pM and 160 pM for cynomolgus monkey PAD2. D The antibody may have an affinity (K) for mouse PAD2 of about 1 pM, 2 pM, 3 pM, 4 pM, 5 pM, 6 pM, 7 pM, 8 pM, 9 pM, 10 pM, 11 pM, 12 pM, 13 pM, 14 pM, 15 pM, 16 pM, 17 pM, 18 pM, 19 pM, 20 pM, 21 pM, 22 pM, 23 pM, 24 pM, 25 pM, 26 pM, 27 pM, 28 pM, 29 pM, 30 pM, 31 pM, 32 pM, 33 pM, 34 pM, 35 pM, 36 pM, 37 pM, 38 pM, 39 pM or 40 pM. D The antibody may have an affinity (K of ≤20 pM or ≤18 pM). D ) binds to human PAD2. Affinity can be measured by surface plasmon resonance (SPR).

[0258] The bispecific antibody can bind to cynomolgus monkey PAD2 with an affinity (K D ) and the affinity (K D The bispecific antibody can bind to cynomolgus monkey PAD2 with an affinity (K D ) in the affinity (K) of the bivalent Fab fragment of IgG containing the same PAD2 binding domain to cynomolgus monkey PAD2D ) within ±5 pM. The bispecific antibody can bind to PAD2 with an affinity (K D ) in the affinity (K) of the bivalent Fab fragment of IgG containing the same PAD2 binding domain to cynomolgus monkey PAD2 D ) within ±10 pM. The K of the bispecific antibody against cynomolgus monkey PAD D The K of the bivalent Fab fragment of IgG containing the same PAD2 binding domain against cynomolgus monkey PAD2 can be less than D .

[0259] The bispecific antibody can bind to mouse PAD2 with an affinity (K D ) and the affinity of the bivalent Fab fragment of IgG containing the same PAD2 binding domain to mouse PAD2 (K D The bispecific antibody can bind to mouse PAD2 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD2 binding domain to mouse PAD2 D ) within ±5 pM. The bispecific antibody can bind to PAD2 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD2 binding domain to mouse PAD2 D The K of the bispecific antibody against mouse PAD2 is within ±10 pM. D The K of the bivalent Fab fragment of IgG containing the same PAD2 binding domain against mouse PAD2 can be less than D .

[0260] The affinity of the antibody for PAD2 (K D ) can be any affinity provided in the Examples, in particular as provided in Table 68 and Table 69. The affinity of the antibody for PAD4 is K D It may be within the affinity ranges provided in the Examples, in particular as provided in Table 70 or Table 71.

[0261] Affinity (e.g. K D ) can be measured by surface plasmon resonance (SPR).

[0262] 4.7PAD4 affinity

[0263] The bispecific antibody can bind to human PAD4 with an affinity (K D ) and the affinity of the bivalent Fab fragment of IgG containing the same PAD4 binding domain for human PAD4 (K D ) are equal. The bispecific antibody can bind to human PAD4 with an affinity (K D) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD4 binding domain to human PAD4 D ) within ±5 pM. The bispecific antibody can bind to PAD4 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD4 binding domain to human PAD4 D ) within ±10 pM. The bispecific antibody can bind to PAD4 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD4 binding domain to human PAD4 D ) within ±20 pM. The bispecific antibody can bind to PAD4 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD4 binding domain to human PAD4 D ) within ±30pM.

[0264] The antibody or bispecific antibody is resistant to the K D The K of the bivalent Fab fragment of IgG containing the same PAD4 binding domain to human PAD4 may be less than D This antibody has a K D The antibody may have an affinity for human PAD4 of about 1 pM, 2 pM, 3 pM, 4 pM, 5 pM, 6 pM, 7 pM, 8 pM, 9 pM, 10 pM, 11 pM, 12 pM, 13 pM, 14 pM, 15 pM, 16 pM, 17 pM, 18 pM, 19 pM, 20 pM, 21 pM, 22 pM, 23 pM, 24 pM, 25 pM, 26 pM, 27 pM, 28 pM, 29 pM, 30 pM, 31 pM, 32 pM, 33 pM, 34 pM, 35 pM, 36 pM, 37 pM, 38 pM, 39 pM, 40 pM, 41 pM, 42 pM, 43 pM, 44 pM, 45 pM, 46 pM, 47 pM, 48 pM, 49 pM, 50 pM, 51 pM, 52 pM, 53 pM, 54 pM, 55 pM, 56 pM, 57 pM, 58 pM, 59 pM, 60 pM, 61 pM, 62 pM, 63 pM, 64 pM, 65 pM, 66 pM, 67 pM, 68 pM, 69 pM, 70 pM, or 53 pM.

[0265] The antibody may have an affinity K of about 5 pM to 50 pM or 5 pM to 45 pM for mouse PAD4. DThe antibody can have an affinity for mouse PAD4 of about 1 pM, 2 pM, 3 pM, 4 pM, 5 pM, 6 pM, 7 pM, 8 pM, 9 pM, 10 pM, 11 pM, 12 pM, 13 pM, 14 pM, 15 pM, 16 pM, 17 pM, 18 pM, 19 pM, 20 pM, 21 pM, 22 pM, 23 pM, 24 pM, 25 pM , 26pM, 27pM, 28pM, 29pM, 30pM, 31pM, 32pM, 33pM, 34pM, 35pM, 36pM, 37pM, 38pM, 39pM, 40pM, 41pM, 42pM, 43pM, 44pM, 45pM, 46pM, 47pM, 48pM, 49pM or 50pM. D .

[0266] The bispecific antibody can bind to cynomolgus monkey PAD4 with an affinity (K D ) and the affinity (K D The bispecific antibody can bind to cynomolgus monkey PAD4 with an affinity (K D ) in the affinity (K) of the bivalent Fab fragment of IgG containing the same PAD4 binding domain to cynomolgus monkey PAD4 D ) within ±1 pM. The bispecific antibody can bind to PAD4 with an affinity (K D ) in the affinity (K) of the bivalent Fab fragment of IgG containing the same PAD4 binding domain to cynomolgus monkey PAD4 D ) within ±2 pM. The K of the bispecific antibody against cynomolgus monkey PAD4 D The K of the bivalent Fab fragment of IgG containing the same PAD4 binding domain against cynomolgus monkey PAD4 can be less than D .

[0267] The bispecific antibody can bind to mouse PAD4 with an affinity (K D ) and the affinity (K) of a bivalent Fab fragment of IgG containing the same PAD2 binding domain for mouse PAD4. D ) are equal. The bispecific antibody can bind to mouse PAD4 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD4 binding domain to mouse PAD4 D ) within ±5 pM. The bispecific antibody can bind to PAD4 with an affinity (K D ) in the affinity (K) of bivalent Fab fragments of IgG containing the same PAD4 binding domain to mouse PAD2 D The K of the bispecific antibody against mouse PAD is within ±10 pM.D The K of the bivalent Fab fragment of IgG containing the same PAD4 binding domain against mouse PAD4 can be less than D .

[0268] The antibody can bind with an affinity (K of ≤40 pM or ≤35 pM) D ) binds to human PAD4.

[0269] The affinity of this antibody for PAD4 is K D It can be any affinity provided in the Examples, in particular as provided in Table 70 and Table 71. The affinity of the antibody for PAD4 is K D It may be within the affinity ranges provided in the Examples, in particular as provided in Table 70 or Table 71.

[0270] 4.8 Thermal stability

[0271] The bispecific antibody may have beneficial thermal stability. The antibody or bispecific antibody may have a T of ≥ 40°C. 起始 The antibody or bispecific antibody may have a T of about 44°C, 45°C, 46°C, 47°C, 48°C, 49°C, 50°C, 51°C, 52°C, 53°C or 54°C. 起始 The antibody or bispecific antibody may have a T of 40°C to 54°C. 起始 . T 起始 The thermal stability (T 起始 ) can be equal to or greater than the thermal stability (T 起始 The thermal stability of the bispecific antibody (T 起始 ) can be equal to or greater than the thermal stability (T 起始 The thermal stability of the bispecific antibody (T 起 The thermal stability (T 起始 ) within ±10°C. The thermal stability of the bispecific antibody (T 起始 ) can be compared with the thermal stability (T 起始 ) within ±20°C. The thermal stability of the bispecific antibody (T 起始 ) can be compared with the thermal stability (T 起始 ) within ±10°C. The thermal stability of the bispecific antibody (T 起始) can be compared with the thermal stability (T 起始 ) within ±20℃.

[0272] The thermal stability of the antibody (T 起始 ) can be any value provided in Table 91. The thermal stability (T 起始 ) can be within the range of values provided in Table 91.

[0273] 4.9 Aggregation

[0274] The antibody may have a beneficially low risk of aggregation. The propensity of the bispecific antibody to aggregate is no more than 2 times that of a bivalent Fab of an IgG comprising the same PAD2 binding domain. The propensity of the bispecific antibody to aggregate is no more than 2 times that of a bivalent Fab of an IgG comprising the same PAD4 binding domain. The aggregation of the antibody at 40°C is no more than 2 times that of a bivalent IgG1 antibody comprising the same PAD2 binding domain or PAD4 binding domain. The aggregation of the antibody at 40°C is no more than 2 times that of a bivalent IgG1 antibody comprising the same PAD2 binding domain or PAD4 binding domain.

[0275] 4.10Bis3

[0276] The bispecific antibody can be a bivalent bispecific antibody. The bispecific antibody can be in Bis3 format, i.e., having scFv and IgG binding domains ( Figure 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. PAD2 and PAD4 may be human, cynomolgus monkey, and / or mouse PAD2 or PAD4. The antibody may comprise two scFv domains, wherein each scFv domain comprises a PAD2 binding domain according to the present invention. The antibody may comprise two scFv domains, wherein each scFv domain comprises a PAD4 binding domain according to the present invention.

[0277] The antibody may comprise two Fab domains, wherein each Fab comprises a PAD2 binding domain according to the present invention. The antibody may comprise two Fab domains, wherein each Fab comprises a PAD4 binding domain according to the present invention. The antibody may comprise a Figure 10B Bis3 bispecific antibodies with any of the structures shown in .

[0278] Bis3 bispecific antibodies can bind with an affinity (K D) binds to human PAD4, 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 of about 9 pM) D ) binds 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.

[0279] Bis3 bispecific antibodies can bind with an affinity of 6 pM-7 pM (K D ) binds to human PAD2, 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 of about 16 pM). D ) binds to human PAD2, 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. The antibody can bind to human PAD2 with an affinity (K of ≤ 17 pM) D ) binds to human PAD2. The antibody can bind to human PAD2 with an affinity (K D ) binds to human PAD4.

[0280] The Bis3 bispecific antibodies may have affinities (K) for human or cynomolgus monkey PAD2 as provided in Table 72 or Table 73. D Bis3 may have an affinity (K) for PAD4 as provided in Table 75 or Table 74. D ).

[0281] The antibody may be a bispecific antibody comprising: a) an IgG comprising a first and a second Fab domain 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, wherein the first and second scFv are each linked to the carboxyl terminus of one of the heavy chains of the Fc domain of the IgG, and wherein the first and second scFv each comprise a PAD4-binding domain that specifically binds to PAD4. The Fc domain may be an IgG or IgG1 Fc domain. The first and second scFv PAD4-binding domains may comprise SEQ ID NO: 39. The first and second Fab PAD2-binding domains may comprise a heavy chain domain comprising SEQ ID NO: 34. The first and second Fab PAD2-binding domains may comprise a light chain constant domain comprising SEQ ID NO: 36. The first Fab PAD2 binding domain and the second Fab PAD2 binding domain comprise a light chain domain comprising SEQ ID NO:35.

[0282] The antibody may be a bispecific antibody comprising a) an IgG comprising a first and a second Fab domain 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, wherein the first and second scFv are each attached to the carboxyl terminus of one of the heavy chains of the Fc domain of the IgG, and wherein the first and second scFv 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 a heavy chain constant domain that may comprise SEQ ID NO: 37. The first Fab domain and the second Fab domain comprise a light chain constant domain comprising SEQ ID NO:36.

[0283] The scFv can be linked to the carboxyl terminus of the heavy chain via a peptide linker. The peptide linker can comprise SEQ ID NO: 51. The first scFv and / or the second scFv of the Bis3 bispecific antibody can comprise a VH-VL linker domain comprising SEQ ID NO: 33.

[0284] 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.

[0285] The Bis3 bispecific antibody may have the sequence of any one 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 the sequence provided in Table 57.

[0286] 4.11DuetMab

[0287] The antibody may be a monovalent bispecific antibody. The antibody may be a DuetMab. Figure 10A DuetMab with either structure shown in .

[0288] The antibody may comprise a PAD2 binding domain, such that the antibody is monovalent for PAD2, and the antibody may comprise a 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 a second Fab domain comprises a PAD2 binding domain, and a second binding region comprising a second Fab, wherein the second Fab comprises a PAD4 binding domain. The antibody may comprise an IgG domain having 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.

[0289] The 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 a PAD4 binding region comprising a heavy chain comprising SEQ ID NO: 76 and a light chain comprising SEQ ID NO: 65. The 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. The DuetMab bispecific antibody may comprise a heavy chain comprising SEQ ID NO: 59 or SEQ ID NO: 60 and a light chain comprising SEQ ID NO: 62, and a PAD4 binding region comprising a heavy chain comprising SEQ ID NO: 61 and a light chain comprising SEQ ID NO: 65. 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. 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.

[0290] DuetMab can have the sequence of any of clones 01, 02, 03, 04, 05, or 06. DuetMab bispecific antibodies can have a sequence that is at least 90% identical to the sequence of 01, 02, 03, 04, 05, or 06. Bis3 bispecific antibodies can have the sequences provided in Table 58.

[0291] 4.12IgG

[0292] The antibody may comprise an IgG or F(ab')2 fragment. The antibody may comprise an IgG or F(ab')2 fragment that is divalent for PAD2 or divalent for PAD4. The antibody may be an IgG1 that is divalent for PAD2 or PAD4. 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 of the PAD2 binding domains of the present invention, such that the antibody is divalent for PAD2. The antibody may comprise two of the PAD4 binding domains of the present invention, such that the antibody is divalent for PAD4. The antibody may comprise two of the PAD4 binding domains of the present invention, and not comprise a PAD2 binding domain according to the present invention. The antibody may comprise two PAD2 binding domains of the PAD2 binding domains of the present invention and not comprise a PAD4 binding domain according to the present invention.

[0293] IgG may comprise a heavy chain with a terminal lysine. IgG may comprise two heavy chains with a terminal lysine. IgG may comprise a heavy chain with a terminal lysine and a heavy chain without a terminal lysine. IgG may comprise two heavy chains without a terminal lysine.

[0294] 4.13Fab

[0295] The antibody may comprise a Fab fragment, wherein the Fab fragment comprises a PAD2 binding domain or a PAD4 binding domain. The Fab fragment may comprise a PAD2 binding domain, wherein the Fab binds to the PAD4 binding domain with an affinity (K) of ≤20 nM, ≤10 nM, ≤6 nM or ≤1 nM. D ) binds to human PAD2. The antibody may comprise a Fab fragment, wherein the Fab fragment comprises a PAD4 binding domain, and wherein 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 It can be measured by surface plasmon resonance (SPR).

[0296] 4.14 Variable Region

[0297] The variable region of the antibody can be a human variable region. The variable region can include rodent or mouse complementary determining regions (CDRs) and human framework regions (FRs). The variable region can be a primate (e.g., non-human primate) variable region. The variable region can include rodent or mouse CDRs and primate (e.g., non-human primate) framework regions (FRs). The variable region can include CDRs, VH, VL or framework regions of any of the antibodies described in Tables 1 to 54 and 62.

[0298] 4.14.1PAD2 leader sequence

[0299] The antibody may comprise a PAD2 binding domain, wherein the PAD2 binding domain comprises: a variable heavy (VH) domain sequence comprising CDRs HCDR1, HCDR2, and HCDR3, and a variable light (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, wherein the PAD2 binding domain comprises a VH domain comprising a sequence that is at least 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical 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 that is at least 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to SEQ ID NO: 2. The PAD2 binding domain may comprise a VL domain 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 of 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 of the CDRs. The antibody may comprise the VH, VL, CDR, and framework region sequences of the antibodies described in Table 1. The antibody may be an affinity-optimized antibody of an antibody described in Table 53 or Table 54.

[0300] 4.14.2PAD4 leader sequence

[0301] The bispecific antibody may comprise a PAD4 binding domain comprising: 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, and 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 that is at least 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to SEQ ID NO: 31. The PAD4 binding domain may comprise a VH domain comprising SEQ ID NO: 31. The PAD4 binding domain may comprise a VL domain comprising a sequence that is at least 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, or 99% identical to SEQ ID NO: 32. The PAD4 binding domain may comprise a domain comprising a VL sequence of 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 of 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 of the CDRs. The antibody may comprise the VH, VL, CDR, and framework region sequences of the antibodies described in Table 2. The antibody may be an affinity-optimized antibody of the antibodies described in Table 62.

[0302] 4.14.3PAD4 Alternate Clone

[0303] The antibody or bispecific antibody may comprise a sequence in any of the antibody sequences provided in Tables 3 to 42. The antibody may comprise the CDRs, framework regions, VH sequences, 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, or PAD40141 hIgG1 ngl-2. The antibody may have a sequence with 90% sequence identity to a VH sequence provided in any one of Tables 3 to 42. A sequence with 90% sequence identity to a VL sequence provided in any one of Tables 3 to 42. The antibody may have a VH, VL, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, LCDR3 that are identical in sequence to the corresponding regions of clone 42 provided in Table 63.

[0304] 4.14.4PAD2 Backup Clone

[0305] The antibody or bispecific antibody may comprise a sequence in any of the antibody sequences provided in Tables 43 to 54. The antibody may comprise the CDRs, framework regions, VH sequences, 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, or 141LO0055 hIgG1 ngl-2. The antibody may have a sequence with 90% sequence identity to a VH sequence provided in any one of Tables 43 to 54. A sequence with 90% sequence identity to a VL sequence provided in any one of Tables 43 to 54. The antibody may have VH, VL, HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, LCDR3 with sequences identical to the corresponding regions of clone 42 provided in Table 63.

[0306] 4.15Fc modification

[0307] The antibody or bispecific antibody may comprise an Fc domain, optionally an IgG1 Fc domain. The Fc domain may have an ineffective effector function. The Fc domain may comprise mutations L234F, L235Q and / or K322Q, as numbered by the EU index as shown in Kabat et al.

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

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

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

[14] . The Fc domain may comprise a C sequence comprising SEQ ID NO: 47, SEQ ID NO: 48 or SEQ ID NO: 49. H 2 domains. The Fc domain may have any of the mutations described in Table 65.

[0308] The polypeptides according to the present invention may comprise an Fc variant domain.

[0309] IgG Fc domains with extended half-life 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 H 2 triple mutations: M252Y, S254T and T256E.

[0310] The modified Fc region may comprise an amino acid substitution at two or more of positions 432 to 437, numbered according to the EU numbering index as of Kabat, relative to a wild-type human Fc region; wherein: (i) position 432 and position 437 are each substituted with cysteine; (ii) position 433 is histidine or is substituted with arginine, proline, threonine, lysine, serine, alanine, methionine, or asparagine; (iii) position 434 is asparagine or is substituted with arginine, tryptophan, histidine, phenylalanine, tyrosine, serine, methionine, or threonine; (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 wherein the modified human IgG1 has an increased half-life compared to the half-life of IgG1 having a wild-type human Fc region.

[0311] The modified Fc region may comprise amino acid substitutions at two or more of positions 432 to 437, numbered according to the EU numbering index as in Kabat, relative to the wild-type Fc region; wherein

[0312] a) at least one of position 432 and position 437 is substituted with cysteine; or

[0313] b) at least one of position 432 and position 437 is substituted with an amino acid selected from the group consisting of glutamine, glutamic acid, aspartic acid, lysine, arginine, and histidine;

[0314] wherein the polypeptide has an altered half-life compared to the half-life of an IgG with a wild-type Fc region. Optionally, (i) both position 432 and position 437 are substituted with cysteine; or (ii) both position 432 and position 437 are substituted with an amino acid independently selected from the group consisting of glutamine, glutamic acid, aspartic acid, lysine, arginine, and histidine.

[0315] The polypeptide may comprise an amino acid insertion after position 437, optionally wherein the amino acid insertion is glutamic acid.

[0316] The polypeptide may have a higher binding affinity for FcRn at pH 6.0 than that of an IgG with a wild-type Fc region at pH 6. The polypeptide may have a higher binding affinity for FcRn at pH 7.4 than that of an IgG with a wild-type Fc region at pH 7.4. The polypeptide may have a KD for FcRn at pH 6.0 of less than 500 nM and a KD of at least 1000 nM at pH 7.4.

[0317] The polypeptide may comprise an Fc variant domain, wherein the Fc variant domain exhibits an increased pH dependence of 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 decreased pH dependence of binding affinity to FcRn compared to an IgG having a wild-type Fc region.

[0318] The polypeptide may comprise an Fc variant domain, wherein the modified IgG Fc domain retains wild-type levels of at least one property selected from the group consisting of: (i) binding to at least one Fcγ receptor, (ii) binding to C1q, or (iii) effector function, optionally wherein the Fcγ receptor is selected from the group consisting of an FcγRI receptor, an FcγRII receptor, and an FcγRIII receptor. The polypeptide may have reduced effector function selected from the group consisting of 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.

[0319] The polypeptide may comprise an Fc variant domain, wherein the Fc variant domain has an amino acid substitution at four or more of 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; position 436 may be tyrosine or phenylalanine, or substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine, or threonine. Position 433 may be histidine. Position 433 may 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 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.

[0320] The polypeptide may comprise an Fc variant domain, wherein the Fc variant domain may comprise the amino acid sequence at positions 432 to 437 of CXRHXC (SEQ ID NO: 187), wherein 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 at positions 432 to 437 of CRRHXC (SEQ ID NO: 188), wherein position 436 is substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine, or threonine. Position 436 may be substituted with leucine, isoleucine, serine, or threonine.

[0321] The polypeptide may comprise an Fc variant domain, wherein the Fc variant domain may comprise the amino acid sequence at positions 432 to 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 at positions 432 to 437 of ZXXHXZ (SEQ ID NO: 92), wherein 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.

[0322] The Fc variant domain may comprise 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, Y8-YTE. The Fc variant domain may have a histidine at amino acid position 435. The modified IgG Fc domain may comprise the amino acid sequence of E(R / A)(W / S / F)HRQ (SEQ ID NO: 190) at positions 432 to 437.

[0323] The polypeptide may comprise at least an FcRn binding portion of an Fc region of an IgG molecule, wherein the FcRn binding portion comprises an amino acid substitution at two or more of positions 432 to 437, numbered according to the EU numbering index as in Kabat, relative to a wild-type FcRn binding portion; wherein (i) at least one of position 432 and position 437 is substituted with cysteine; or (ii) at least one of position 432 and position 437 is substituted with an amino acid selected from the group consisting of glutamine, glutamic acid, aspartic acid, and histidine. The polypeptide of claim 38, wherein (i) both position 432 and position 437 are substituted with cysteine; or (ii) both position 432 and position 437 are substituted with an amino acid independently selected from the group consisting of glutamine, glutamic acid, aspartic acid, and histidine. Position 432 and position 437 can each be substituted with cysteine; position 433 can be histidine or substituted with arginine, proline, threonine, lysine, serine, alanine, methionine or asparagine; position 434 can be asparagine or substituted with arginine, tryptophan, histidine, phenylalanine, tyrosine, serine, methionine or threonine; position 435 can be histidine or substituted with histidine; position 436 can be tyrosine or phenylalanine, or substituted with leucine, arginine, isoleucine, lysine, methionine, valine, histidine, serine or threonine.

[0324] The Fc variant domain may comprise the amino acid sequence of positions 432 to 437 of ZXXHXZ, wherein 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 may comprise the amino acid sequence of E(R / A)(W / S / F)HRQ at positions 432 to 437. An amino acid insertion may also be present after position 437, wherein the amino acid insertion is glutamic acid. The FcRn binding portion of the Fc region may comprise approximately amino acid residues 231-446 of an IgG molecule according to the EU numbering index as in Kabat. The FcRn binding portion of the Fc region may comprise approximately amino acid residues 216-446 of an IgG molecule according to the EU numbering index as in Kabat.

[0325] Variant IgG Fc domains with reduced effector function and extended half-life are described in WO2013 / 165690(A1)

[17] . The variant IgG Fc domain may comprise:

[0326] a) a phenylalanine (F) amino acid at position 234;

[0327] b) an alanine (A), asparagine (N), phenylalanine (F), glutamine (Q), or valine (V) amino acid at position 235; and,

[0328] c) 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,

[0329] The amino acid numbering is according to the EU index as in Kabat.

[0330] 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, wherein amino acid numbering is according to the EU index as in Kabat.

[0331] 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, wherein the 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, wherein the amino acid numbering is according to the EU index as in Kabat.

[0332] The Fc variant domain may comprise:

[0333] a) Tyrosine (Y) amino acid at position 252, or Serine (S) amino acid at position 252

[0334] amino acid, or a tryptophan (W) amino acid at position 252 or a threonine (T) amino acid at position 252; and / or

[0335] b) a threonine (T) amino acid at position 254; and / or

[0336] c) glutamic acid (E) amino acid at position 256, or serine (S) amino acid at position 256

[0337] amino acid, 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,

[0338] The amino acid numbering is according to the EU index as in Kabat.

[0339] The Fc variant domain may comprise:

[0340] a) a tyrosine (Y) amino acid at position 252; and / or

[0341] b) a threonine (T) amino acid at position 254; and / or

[0342] c) Glutamic acid (E) amino acid at position 256, wherein amino acid numbering is according to the EU index as in Kabat.

[0343] The Fc variant domain may comprise:

[0344] a) Tyrosine (Y) amino acid at position 252, or Serine (S) amino acid at position 252

[0345] amino acid, or a tryptophan (W) amino acid at position 252 or a threonine (T) amino acid at position 252; and

[0346] b) a threonine (T) amino acid at position 254,

[0347] The amino acid numbering is according to the EU index as in Kabat.

[0348] The Fc variant domain may comprise:

[0349] a) a threonine (T) amino acid at position 254; and

[0350] b) glutamic acid (E) amino acid at position 256, or serine (S) amino acid at position 256

[0351] amino acid, 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,

[0352] The amino acid numbering is according to the EU index as in Kabat.

[0353] The Fc variant domain may comprise:

[0354] a) Tyrosine (Y) amino acid at position 252, or Serine (S) amino acid at position 252

[0355] amino acid, or a tryptophan (W) amino acid at position 252 or a threonine (T) amino acid at position 252; and

[0356] b) glutamic acid (E) amino acid at position 256, or serine (S) amino acid at position 256

[0357] amino acid, 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,

[0358] The amino acid numbering is according to the EU index as in Kabat.

[0359] The Fc variant domain may comprise:

[0360] a) Tyrosine (Y) amino acid at position 252, and Threonine (T) at position 254

[0361] amino acids;

[0362] b) a threonine (T) amino acid at position 254 and a glutamic acid (E) amino acid at position 256;

[0363] c) a tyrosine (Y) amino acid at position 252 and a glutamic acid (E) amino acid at position 256,

[0364] The amino acid numbering is according to the EU index as in Kabat.

[0365] 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, wherein amino acid numbering is according to the EU index as in Kabat.

[0366] The Fc variant domain may comprise:

[0367] a) a phenylalanine (F) amino acid at position 234;

[0368] b) glutamine (Q) amino acid at position 235;

[0369] c) glutamine (Q) amino acid at position 322;

[0370] d) a tyrosine (Y) amino acid at position 252;

[0371] e) a threonine (T) amino acid at position 254; and,

[0372] f) a glutamic acid (E) amino acid at position 256,

[0373] The amino acid numbering is according to the EU index as in Kabat.

[0374] The Fc variant domain may comprise:

[0375] a) a phenylalanine (F) amino acid at position 234;

[0376] b) glutamine (Q) amino acid at position 235;

[0377] c) a glycine (G) amino acid at position 331;

[0378] d) a tyrosine (Y) amino acid at position 252;

[0379] e) a threonine (T) amino acid at position 254; and,

[0380] f) a glutamic acid (E) amino acid at position 256,

[0381] The amino acid numbering is according to the EU index as in Kabat.

[0382] The polypeptide may comprise a modified Fc variant domain, wherein the polypeptide has improved pharmacokinetic (PK) properties compared to the same polypeptide comprising a wild-type Fc domain, optionally wherein the PK property is half-life. The polypeptide may have improved FcRn binding when compared to the same polypeptide comprising a wild-type Fc domain.

[0383] The polypeptide may comprise an IgG Fc domain selected from the group consisting of a human immunoglobulin G1 class (IgG1) Fc domain, a human immunoglobulin G2 class (IgG2) Fc domain, a human immunoglobulin G3 class (IgG) Fc domain, and a human immunoglobulin G4 class (IgG4) Fc domain.

[0384] The polypeptide may comprise a modified Fc variant domain, wherein when compared to the same polypeptide comprising a wild-type Fc domain, the polypeptide has reduced Fc-mediated effector function. The effector function may be antibody-dependent cell-mediated cytotoxicity (ADCC) and / or complement-dependent cytotoxicity (CDC). When compared to the same polypeptide comprising a wild-type Fc domain, the polypeptide may have a lower affinity for Fc gamma receptors (FcγRs), optionally wherein the FcγRs are human FcγRs. The FcγRs may be FcγRI, FcγRII, FcγRIII, FcγRII, FcγRIa, FcγRIIa, FcγRIIb, FcγRIII (158V), FcγRIII (158F).

[0385] The polypeptide may comprise a modified Fc variant domain, wherein the polypeptide binds to FcRn with improved affinity when compared to the same polypeptide comprising a wild-type Fc domain, optionally wherein the polypeptide has a greater affinity for FcRn at pH 6.0 than at pH 7.4.

[0386] The polypeptide may comprise a modified Fc variant domain, wherein the polypeptide binds C1q with reduced affinity when compared to the same polypeptide comprising a wild-type Fc domain.

[0387] The polypeptide may exhibit increased thermal stability when compared to the same polypeptide comprising a FES-YTE IgG Fc domain, optionally wherein thermal stability is measured by differential scanning calorimetry (DSC), optionally wherein thermal stability is increased by at least 4°C.

[0388] The polypeptide may exhibit increased thermal stability when compared to the same polypeptide comprising a FES-YTE IgG Fc domain, wherein the thermal stability is measured by differential scanning fluorimetry (DSF), optionally wherein the DSF fluorescent probe is SyproOrange, optionally wherein the thermal stability is increased by at least 5°C.

[0389] The polypeptide may exhibit an increase in apparent solubility when compared to the same polypeptide comprising a FES-YTE IgG Fc domain, as measured using a polyethylene glycol (PEG) precipitation assay.

[0390] When compared to the same polypeptide comprising a FES-YTE IgG Fc domain, the polypeptide may exhibit increased stability as measured using an accelerated stability assay. The accelerated stability assay comprises: (i) incubating the polypeptide for an extended period of time, and (ii) incubating at an elevated temperature. The accelerated stability assay can be performed by incubating at a high concentration, optionally wherein the extended period of time is at least one month, optionally wherein the high concentration is at least 25 mg / ml, optionally wherein the elevated temperature is at least 40°C. The accelerated stability assay can be performed using high performance size exclusion chromatography (HPSEC) or dynamic light scattering (DLS).

[0391] The Fc may contain RF double mutations.

[0392] The Fc may comprise a knob-in-hole mutation.

[0393] 4.16 Peptides

[0394] The present invention also relates to polypeptides that constitute the antibodies or bispecific antibodies of the present invention. The present invention also provides polypeptides comprising one or more binding domains of an antibody as defined anywhere herein. The polypeptide may comprise a portion or all of a PAD2 binding domain. The polypeptide may comprise a portion or all of a PAD4 binding domain. The polypeptide may comprise a binding domain, such as one or more CDRs as defined herein, or a variable light domain or a variable heavy domain as defined herein. The polypeptide may comprise a binding domain comprising all three CDRs (CDR1, CDR2, and CDR3) of a variable heavy domain sequence as defined herein. The polypeptide may comprise a binding domain comprising all three CDRs (CDR1, CDR2, and CDR3) of a variable light domain sequence as defined herein. The polypeptide may comprise a variable heavy domain of an antibody as defined herein. The polypeptide may comprise a variable light domain of an antibody as defined herein. The polypeptide may comprise a complete heavy chain of an antibody as defined herein. The polypeptide may comprise a complete light chain of an antibody as defined herein. The polypeptide may be an isolated polypeptide.

[0395] 4.17 Nucleic Acids

[0396] The present invention also relates to nucleic acids encoding one or more chains of an antibody or bispecific antibody of the present invention. The present invention also relates to nucleic acids encoding polypeptides according to the present invention. The present invention also relates to vectors comprising the nucleic acids, and host cells comprising the vectors.

[0397] 4.18 Pharmaceutical Compositions

[0398] The present invention also relates to a pharmaceutical composition comprising the antibody or bispecific antibody of the present invention and a pharmaceutically acceptable carrier.

[0399] 4.19 kit

[0400] The present invention also relates to a kit comprising the antibody or bispecific or pharmaceutical composition of the present invention. The kit may include instructions for use.

[0401] 4.20 Treatment Methods

[0402] The present invention also relates to a method for 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 PAD levels in synovial fluid, whole blood, or serum compared to healthy subjects. The subject may have elevated PAD2 levels in synovial fluid, whole blood, or serum compared to healthy subjects. The subject may have elevated PAD4 levels in synovial fluid, whole blood, or serum compared to healthy subjects. The concentration of PAD4 in the subject's synovial fluid may be at least 200 ng / ml. The concentration of PAD2 in the subject's synovial fluid may be at least 20 ng / ml. The concentration of PAD2 and / or PAD4 in the subject's synovial fluid may be within the range of values provided in Table 82. The concentration of PAD2 and / or PAD4 in the subject's whole blood may be at least 1 ng / ml. The concentration of PAD2 and / or PAD4 in the subject's whole blood may be within the range of values provided in Table 83. The concentration of PAD2 or PAD4 can be determined by ELISA.

[0403] 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 can be bivalent IgGs comprising at least two Fab(2) fragments each of which has a binding domain specific to PAD4 or PAD2. The anti-PAD2 antibody and the anti-PAD4 antibody can be administered to the subject simultaneously, separately, or sequentially.

[0404] 4.21EPC 2000

[0405] The present invention also relates to an antibody or pharmaceutical composition of the present invention for use in a method of treating or preventing a disease in a subject. The disease may be an autoimmune disorder. The disease may be characterized by increased PAD activity in a tissue relative to healthy subjects. The disease may be characterized by increased PAD2 and / or PAD4 activity in a tissue relative to healthy subjects. The tissue may be synovial fluid, whole blood, or serum.

[0406] 4.22 Swiss

[0407] The present invention also relates to an antibody of the invention or a pharmaceutical composition of the invention for use in the manufacture of a medicament for treating an autoimmune disorder. The treatment may comprise a method of treatment according to the invention.

[0408] 5 Terminology

[0409] 5.1 Antibodies

[0410] The term "antibody" means an immunoglobulin molecule that recognizes and specifically binds 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.

[0411] 5.1.1 Antibody fragments

[0412] 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 the antigenic determining regions (e.g., complementarity-determining regions (CDRs)) of an intact antibody. 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, such as rodents (e.g., mice, rats, or hamsters) and humans, or may be artificially produced.

[0413] 5.1.2 Anti-PAD2 Antibodies

[0414] The terms "anti-PAD2 antibody," "PAD2 antibody," and "antibody that binds to PAD2" are used interchangeably herein to refer to an antibody that is 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 aspects provided herein, the PAD2 antibody is also capable of binding to PAD3. In some aspects provided herein, the PAD2 antibody does not bind to PAD3. In some aspects provided herein, the PAD2 antibody is also capable of binding to PAD1. In some aspects provided herein, the PAD2 antibody does not bind to PAD1.

[0415] 5.1.3 Anti-PAD4 Antibodies

[0416] Similarly, the terms "anti-PAD4 antibody," "PAD4 antibody," and "antibody that binds to PAD4" are used interchangeably herein to refer to an antibody that is 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 aspects provided herein, the PAD4 antibody is also capable of binding to PAD3. In some aspects provided herein, the PAD4 antibody does not bind to PAD3. In some aspects provided herein, the PAD4 antibody is also capable of binding to PAD1. In some aspects provided herein, the PAD4 antibody does not bind to PAD1.

[0417] 5.1.4 Humanized Antibodies

[0418] The term "humanized" antibody or antigen-binding fragment thereof refers to forms of non-human (e.g., murine) antibodies or antigen-binding fragments that are specific immunoglobulin chains, chimeric immunoglobulins, or fragments thereof that contain minimal non-human (e.g., murine) sequences. Typically, a humanized antibody or antigen-binding fragment thereof is a human immunoglobulin in which residues from the complementarity determining regions (CDRs) are replaced with residues from CDRs of a non-human species (e.g., mouse, rat, rabbit, hamster) with the desired specificity, affinity, and capacity ("CDR grafting") [18-20]. In some instances, the Fv framework region (FR) residues of a human immunoglobulin are replaced with corresponding residues from an antibody or fragment with the desired specificity, affinity, and capacity from a non-human species. The humanized antibody or antigen-binding fragment thereof can be further modified by replacing additional residues in the Fv framework region 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 thereof. In general, 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 non-human immunoglobulins, while all or substantially all of the FR regions are those of a human immunoglobulin consensus sequence. A 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.

[0419] 5.1.5 Human Antibodies

[0420] The term "human" antibody or antigen-binding fragment thereof means an antibody or antigen-binding fragment thereof having an amino acid sequence derived from a human immunoglobulin gene locus, wherein such antibody or antigen-binding fragment is made using any technique known in the art. This definition of human antibody or antigen-binding fragment thereof includes intact or full-length antibodies and fragments thereof.

[0421] 5.2 Binding affinity

[0422] "Binding affinity" or "affinity" generally 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 can generally be expressed by the dissociation constant (K D Affinity can be measured and / or expressed in a variety of ways known in the art, including but not limited to the equilibrium dissociation constant (K) and the equilibrium association constant (K A ). By k 0ff / k 0n Calculate K D , and by k on / k off Calculate K A .k on refers to the association rate constant of, for example, an antibody or antigen-binding fragment thereof with an antigen, and k 0ff Refers to, for example, the dissociation of an antibody or its antigen-binding fragment from an antigen. on and k off This can be done by techniques known to those skilled in the art, such as or KinExA to measure.

[0423] 5.3 Bispecific Antibodies

[0424] The term "bispecific antibody" means an antibody that has specificity for two target molecules and includes, but is not limited to, formats such as 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, triabody, minibody, minibody, TriBi minibody, scFv-CH3 KIH, scFv-CH-CL-scFv, F(ab')2-scFv, scFv-KIH, Fab-scFv-Fc, tetravalent HCab, ImmTAC, knob-in-hole, knob-in-hole with common light chain, knob-in-hole with common light chain and charge pair, charge pair, charge pair with 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 comprise an antibody fused to another non-Ig format, such as a T cell receptor binding domain; an immunoglobulin superfamily domain; an agnatian variable lymphocyte receptor; a fibronectin domain (e.g., an Adnectin TM ); antibody constant domains (e.g., CH3 domains, such as Fcab TMCH2 and / or CH3), wherein the constant domain is not a functional CH1 domain; scFv; (scFv) 2; sc-diabody; scFab; centyrin and derived from CTLA-4 (Evibody TM ) of the scaffold; lipocalin domain; protein A, such as the Z domain of protein A (e.g., Affibody TM or SpA); A domain (e.g. Avimer TM or Maxibody TM ); heat shock proteins (such as epitope binding domains from GroEI and GroES); transferrin domains (e.g., trans antibodies); ankyrin repeat proteins (e.g., DARPin TM ); peptide aptamers; C-type lectin domains (e.g., Tetranectin TM ); human γ-crystallin or human ubiquitin (affilin); a PDZ domain; a scorpion toxin; and a Kunitz-type domain of a human protease inhibitor.

[0425] 5.3.1Bis3

[0426] The Bis3 format bispecific antibody comprises an IgG molecule having two Fab domains and two scFvs, wherein each scFv is attached to the C-terminus of each heavy chain (i.e., IgG-HC-scFv,

[23] ). The two Fab domains bind to the same target protein as each other, so the molecule is symmetrical with respect to the Fab domains. The two scFvs bind different target proteins to the Fab domains, and each scFv binds to the same target protein as each other. 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 target bound by the Fab domain and the target bound by the scFv can be in opposite orientations, so that the scFv domain can bind to the first target protein (e.g., PAD2) and the Fab domain can bind to the second target protein (e.g., PAD4).

[0427] 5.3.2DuetMab

[0428] DuetMab antibodies comprise IgG antibodies with two heavy chains and two light chains. The two arms are asymmetric, with each arm binding to a different target protein. The antibody therefore has a single binding domain for each of the two target proteins, making the antibody as a whole bivalent but monovalent for each target protein. DuetMab antibodies use knob-in-hole technology for heterodimerization of two different heavy chains and increase the efficiency of homologous 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 developability properties of native IgG ([23,24]).

[0429] 5.4 C-terminal variants

[0430] Large-scale production of proteins involves the use of cell cultures that 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. The C-terminal lysine can be lost, such that individual antibodies within a production batch may vary in the presence or absence of a lysine residue at their C-termini ("lysine clipping")

[25] . The C-terminal lysine can potentially be present on both heavy chains of an antibody (K2), on either heavy chain (K1), or on no heavy chain (K0).

[0431] 5.5 Complementarity Determining Regions

[0432] As used herein, the term "complementarity determining region" or "CDR" refers to each region of an antibody variable domain that is hypervariable in sequence and / or forms structurally defined loops (hypervariable loops) and / or contains antigen contact residues. An antibody may comprise six CDRs, for example, three in VH and three in VL.

[0433] Kabat numbering is a system for numbering amino acid residues in the heavy chain variable region and light chain variable region of an antibody or antigen-binding fragment thereof. In some aspects, CDRs can be determined according to the Kabat numbering system

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

[0434] 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 for substitutions and insertions in the Fc region in this specification is the EU index as in Kabat

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

[0435] Chothia refers to the position of the structural loops

[27] . When numbered using the Kabat numbering convention, the ends of the Chothia CDR-H1 loop vary 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 both 35A and 35B are absent, the loop ends at 32; if only 35A is present, the loop ends at 33; 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.

[0436] Table 64: Kabat and Chothia

[0437]

[0438] 5.6 Epitope

[0439] As used herein, " epi-position " is the term of this area, and refers to the local area of the antigen that antibody or its Fab can specifically bind.Epi-position can be the continuous amino acids (linear or continuous epi-position) of such as polypeptide, or epi-position can be gathered together (conformation, non-linear, discontinuous or discontinuous epi-position) for example from two or more discontinuous regions of one or more polypeptide.In some respects, the epi-position that antibody or its Fab is combined can be determined by such as NMR spectroscopy, X-ray diffraction crystallography research, ELISA determination method, hydrogen / deuterium exchange combined mass spectrometry (for example, liquid chromatography electrospray mass spectrometry), oligopeptide scanning determination method based on array and / or mutagenesis mapping (for example, site-directed mutagenesis mapping).

[0440] An antibody that "binds to the same epitope" as a reference antibody is one that binds to the same amino acid residue as the reference antibody. The ability of an antibody to bind to the same epitope as the reference antibody can be determined using a hydrogen / deuterium exchange assay

[28] .

[0441] 5.7Fc modification

[0442] Multiple mutation combinations in IgG Fc have been characterized to tailor immune effector function or IgG serum persistence to suit the desired biological outcome of monoclonal antibody therapeutics. Example IgG Fc modifications are summarized in Table 65.

[0443] Table 65: IgG Fc modifications

[0444]

[0445] *Amino acid numbering is according to the EU index as in Kabat

[14]

[0446] The TM modification eliminates 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 the TM modification that reduces effector function. YTE and N3Y are modifications that increase 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.

[0447] 5.8 Separated

[0448] An "isolated" polypeptide, antibody, polynucleotide, vector, cell, or composition is one that is 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 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 of contaminants), at least 90% pure, at least 95% pure, at least 98% pure, or at least 99% pure.

[0449] 5.9 Knob-Mortar Mutation

[0450] The “knob-in-hole” or also known as “knob-into-hole” technology refers to the mutations Y349C, T366S, L368A and Y407V (hole) and S354C and T366W (knob) in the CH3-CH3 interface to promote heteromultimer formation, which has been described in patents US5731168 and US8216805 [31,32].

[0451] The knob mutation refers to the substitutions S139C and T151W in Fc. The hole mutation refers to the substitutions Y134C, T151 S, L153A, and Y192V in Fc.

[0452] 5.10 Percent Identity

[0453] "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 comparing two sequences and introducing a gap to maximize the identity between the sequences. The comparison can be produced using programs known in the art. For purposes herein, the comparison of nucleotide sequences can be performed using the blastn program set to default parameters, and the comparison of amino acid sequences can be performed using the blastp program set to default parameters (see U.S. National Center for Biotechnology Information (NCBI): ncbi.nlm.nih.gov).

[0454] 5.11RF mutation

[0455] "RF mutation" generally refers to a mutation of amino acids HY to RF in the CH3 domain of the Fc domain, such as mutations H435R and Y436F in the CH3 domain. The RF mutation abolishes binding to protein A.

[0456] 5.12 Effectiveness

[0457] Unless otherwise indicated, "potency" is generally expressed as an IC50 value in nM. 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 value. 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 many means known in the art. Improvements in potency can be determined, for example, by measuring the parent antibody (e.g., the parent antibody before germlining or the parent antibody before affinity optimization).

[0458] 5.13 Variable Region

[0459] As used herein, the terms "variable region" or "variable domain" are used interchangeably and are common in the art. Variable region generally refers to a part of an antibody, typically a part of a light chain or a heavy chain, typically about 110 to 120 amino acids or 110 to 125 amino acids at the amino terminus in a mature heavy chain, and about 90 to 115 amino acids in a mature light chain, which are widely different in sequence between antibodies and are used for the combination and specificity of a particular antibody to its specific antigen. The variability of the sequence is concentrated in those regions referred to as complementary determining regions (CDRs), while the more highly conserved regions in the variable domains are referred to as framework regions (FRs). Without wishing to be bound by any particular mechanism or theory, it is believed that the CDRs of light and heavy chains are primarily responsible for the interaction and specificity of the antibody with the antigen.

[0460] 6 Example 1: Method

[0461] 6.1T 起始 Stability assay

[0462] Differential scanning fluorimetry (DSF) is a fluorescence-based protein stability assay that measures protein folding state by monitoring changes in fluorescence with temperature. This technique provides biophysical properties such as the midpoint temperature of thermal unfolding (T m ) and the starting temperature (T 起始 Nano-DSF is a dye-free DSF method that monitors the intrinsic fluorescence of tryptophan in proteins as a function of temperature, time, or denaturant concentration

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

[18] ; using this principle, Nano-DSF determines T by measuring the ratio of fluorescence intensity at 330 nm and 350 nm as a function of temperature. m and T 起始 .

[0463] 6.2 Specificity assay

[0464] ELISA method is used to test the combination of antibody samples and PAD1, PAD2, PAD3 and PAD4. Prepare black high binding plate (Greiner, 781077) by adding 20 μ g / mL streptavidin (Invitrogen, S888) solution in HBSS buffer (Sigma, H8264) of 20 μ L, and incubate 16 hours at 4 DEG C. Plate is balanced to room temperature and washed three times with 100 μ L PBS (Oxoid, BR0014G) containing 0.1% Tween 20 (Sigma, P2287), and add the HBSS solution of 80 μ L 1% BSA (Sigma, A7979). After 1 hour, the plates were washed as previously described, and 20 μL of biotinylated PAD1, PAD2, PAD3, or PAD4 at a concentration of 1 μg / mL prepared 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) were added. After 1 hour, the plates were washed as previously described, and 20 μL of antibodies prepared in 50 mM HEPES pH 7.3, 10 mM CaCl2, 120 mM NaCl, 5 mM DTT, and 0.1 mM CHAPS were added. After 1.5 hours, plate was washed as previously described, and 20 μ L of anti-human IgG Fc HRP (Southern Biotech, 9040-05) of 50 ng / mL concentration prepared in HBSS containing 0.5% BSA were added. After 1 hour, plate was washed as previously described, and 20 μ L QuantaBlu working solutions were added at room temperature. After 30 minutes, 20 μ L QuantaBlu stop solutions (QuantaBlu test kit ThermoFisher, 15169) were added. On PHERAStar FSX microplate reader (BMG Labtech), 330 nm excitation filter and 420 nm emission filter were used to measure fluorescence.

[0465] 6.3 Interferometric Scattering Microscopy (iSCAT)

[0466] Interferometric scattering microscopy (iSCAT) or mass photometry measures light scattering from individual protein molecules in close proximity to a glass surface. The contrast ratio is proportional to the molecular weight of the individual protein molecule / complex. The measurement chip is prepared by cleaning a polished microscope coverslip in deionized water and isopropyl alcohol and then drying it in compressed air before mounting the silicon cartridge wells. The measurement is initiated by adding buffer (10 μL) to the selected wells, followed by automatic focus adjustment.

[0467] Protein samples (10 μL, approximately 50-100 nM) were added, and the contrast from a single protein molecule in the field of view was recorded for 60 seconds (Refeyn Acquire). The resulting raw data images were processed by image analysis (Refeyn Discover) to create a mass histogram of the sample.

[0468] 6.4 Trypsin cleavage efficiency assay

[0469] PAD activity was determined using a short peptide (Cambridge Research Biochemicals) containing arginine, flanked by AlexaFluor 488 acting as a FRET donor and QSY7 acting as a FRET acceptor. If arginine is deiminated to citrulline by the activity of PAD, trypsin will not cleave the peptide, and the fluorescence from the donor is quenched by the acceptor. The inhibition of PAD by anti-PAD scFv prevents arginine deimination and makes the peptide sensitive to trypsin cleavage. The separation of the resulting donor and acceptor fluorophores allows for detectable emission from the donor. 2.5 μl sample scFv was prepared in a determination buffer containing 50mM HEPES, 5mM DTT, 10mM CaCl2 and 0.01% CHAPS, and pre-incubated with 2.5ul PAD4 at 10nM (final concentration). Peptide substrate was prepared as a stock solution, and 5 μl was added to obtain a final concentration of 100nM. After the appropriate incubation time, 10 μl of 100 nM (final concentration) trypsin was added, and the reaction was allowed to proceed for at least two minutes before being read on an EnVision microplate reader (PerkinElmer, Waltham, MA).

[0470] 6.5 Histone-H3 PAD Potency Assay

[0471] PAD2 and PAD4 activities were measured using a histone-H3 citrullination assay using different substrates as follows:

[0472] 6.5.1 Recombinant PAD

[0473] 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); PAD and bispecific antibody were preincubated for 30 minutes; and incubated for 3 hours.

[0474] 6.5.2 Synovial fluid

[0475] 96-well high-binding half-area plates were 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 the antibody for 30 minutes, then transferred to the histone H3-coated plates and incubated at 37°C for 1.5 hours. Rabbit anti-human citrullinated histone H3 Ab was incubated for 1 hour to detect citrullinated histone H3, followed by a 1-hour incubation with goat anti-rabbit-HRP-conjugated Ab. UltraSensitive TMB substrate was used to generate the color reaction, which was measured at 450 nm.

[0476] 6.5.3 Whole blood

[0477] The bispecific antibody was spiked into fresh whole blood overnight. Plasma was collected and PAD activity was assessed at multiple plasma dilutions using the histone H3 PAD activity assay. Whole blood from normal healthy donors was incubated with the Ab overnight at 37°C. Plasma was harvested and stored frozen at -80°C. A 1 / 5 plasma dilution was used in the histone H3 PAD activity assay. PAD surface expression was assessed by FACS. Soluble PAD was assessed by ELISA.

[0478] 6.6 Histone H3cit Western Blot Analysis

[0479] The PAD enzyme converts histone H3 to H3cit, which can be measured using a Western blot protein detection system. Due to the activity of the PAD enzyme in the mouse lungs, LPS exposure causes elevated levels of H3cit in the body. H3cit expression was analyzed using Western blot in bronchoalveolar lavage (BAL) fluid from WT and PAD4KO mice exposed to LPS and saline. Compared to WT mice, mice administered anti-PAD2 / anti-PAD4 had lower levels of H3Cit in their BAL fluid. BAL fluid from PAD4KO mice had no H3Cit.

[0480] 6.7huFcRn affinity chromatography

[0481] A huFcRN-coupled agarose column was used to characterize the affinity of the samples for huFcRN. Approximately 40 μg / 40 μL of sample was loaded onto a 1 mL column, followed by a 3 column volume (CV) linear gradient 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 an 18 CV linear gradient from 40% to 100% buffer B. The experiment was performed using an Agilent-DAD at room temperature at a flow rate of 0.5 mL / min to measure the A280 of the elution profile and the retention time.

[0482] 6.8 Safety Study in Cynomolgus Monkeys

[0483] The target engagement assay primarily revolves around the ability of both PAD2 and PAD4 to citrullinate histone H3. A histone H3 substrate is coated on a plate, where active PADs in the sample deaminate arginine residues to form citrulline. These citrullinated epitopes are then detected by standard immunoassays.

[0484] 6.9 Cytokine Safety Assay

[0485] The potential of Bis3 and DuetMab formats (clone 06 and clone 12) to induce cytokine release was evaluated in 8 donors (4 healthy patients and 4 RA patients) using the following methodologies: a soluble stimulant whole blood assay and a wet-coated fixed stimulant isolated PBMC assay. The concentrations of cytokines IFN-γ, IL-2, IL-6, and TNF-α were measured in collected plasma and cell culture supernatants using Luminex. For each whole blood and isolated PBMC sample, a negative control well was also set up using the same batch and volume of PBS used to prepare the test items and controls.

[0486] 6.10 Generation of Padi4 Knockout Mouse Line

[0487] Design and clone DNA targeting vector to modify mouse endogenous Padi4, peptidyl arginine deaminase, IV type gene. This strategy is based on cloning LoxP sites into the introns flanking exons 7 and 10 of the Padi4 gene. After Cre-induced recombination, this will produce knockout KO, leaving a single LoxP site at 276bp upstream of exon 7 and 736bp downstream of exon 10. The targeting vector is used to modify the Padi4 locus in Primogenix PrX mouse embryonic stem cells (C57Bl6 / N source) via homologous recombination. Correctly targeted cells are injected into 3.5-day-old Balb C blastocysts to produce chimeras. Male chimeric mice are bred with C57Bl6 / N female mice to carry out system amplification. The mouse line established is bred with the R26 Cre deletion system to produce final Padi4 KO allele.

[0488] 6.11 Biacore affinity analysis

[0489] 6.11.1.1 Fab and Monovalent Bispecific Antibodies (DuetMab)

[0490] The affinity of anti-PAD2 antigen-binding fragments (Fab), anti-PAD4 Fab or DuetMabs to PAD substances was measured using Biacore 8K (Cytiva) at 25°C. Experiments were performed using recombinant human PAD2, cynomolgus macaque PAD2, mouse PAD2, human PAD4, cynomolgus macaque PAD4 and mouse PAD4. All substances were enzymatically biotinylated on the Avi tag. Anti-PAD Fab was expressed

[34] or obtained by papain digestion of anti-PAD IgG. After papain incubation for 20 minutes, the sample was loaded at 0.5 ml / min onto a Superdex 200 Increase 10 / 300 GL column equilibrated in D-PBS and the Fab was separated. Streptavidin was covalently immobilized to the surface of a CM5 or C1 chip using standard amine coupling techniques. Recombinant biotinylated PAD2 and PAD4 material was titrated onto the surface of a streptavidin chip in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% surfactant P20, 1 mM CaCl2, 1 mM DTT buffer to enable Fab or DuetMab binding. Serial dilutions of Fab or DuetMab were made in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% surfactant P20, 1 mM CaCl2, 1 mM DTT and flowed over the chip at 50 μl / min, allowing association for 3 minutes and dissociation for 10 minutes. Multiple buffer-only loads were performed under the same conditions to allow for double reference subtraction of the final sensorgram collection. Alternatively, Fab was serially diluted in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% surfactant P20, 1 mM CaCl 2 , 1 mM DTT and loaded onto the chip at 50 μl / min in increasing concentrations (single cycle kinetics), with 2 minutes of association and 10 minutes of dissociation at the end of a complete binding cycle. Buffer-only loading was performed under the same conditions to allow double reference subtraction of the final sensorgram collection. All sensorgram collections were analyzed using Biacore 8K evaluation software. The chip surface was fully regenerated with a 3.0 M MgCl 2 pulse.

[0491] 6.11.1.2 Bispecific Antibodies

[0492] The affinity of recombinant PAD materials to anti-PAD IgG, bivalent Bis3 anti-PAD molecules or monovalent bispecific antibodies (DuetMab) was measured at 25°C using Biacore 8K (Cytiva). Experiments were performed using recombinant human PAD2, cynomolgus monkey PAD2, mouse PAD2, human PAD4, cynomolgus monkey PAD4 and mouse PAD4. Protein G' was covalently immobilized to the C1 chip surface at a concentration of 20 μg / ml in 10 mM sodium acetate pH 3.65 using standard amine coupling techniques. IgG, Bis3 molecules or DuetMab were captured at 5 μl / min onto the protein G' surface in 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% surfactant P20, 1 mM CaCl2, 1 mM DTT buffer to achieve 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 above). PAD species were serially diluted in 10mM HEPES pH 7.4, 150mM NaCl, 0.05% surfactant P20, 1mM CaCl , 1mM DTT and loaded onto the chip at 50 μl / min with sequentially increasing concentrations (single cycle kinetics), associating for 2 minutes and dissociating for 10 minutes at the end of the complete binding cycle. Only buffer loading was performed under the same conditions to allow for dual reference subtraction of the final sensorgram collection, which was analyzed using Biacore 8K evaluation software. The protein G 'chip surface was completely regenerated with a pulse of 6M guanidine hydrochloride (Guanidine HCl) in D-PBS to remove captured molecules and any bound PAD. The streptavidin surface was completely regenerated with a pulse of 3.0 M MgCl2.

[0493] 7 Example 2: Targeting PAD2 and PAD4 in RA Treatment

[0494] PAD2 and PAD4 drive citrullination in RA patients. The present inventors have demonstrated that both PAD2 and PAD4 can generate citrullinated RA antigenic filaments fibrinogen β chain

[35] and α-enolase

[36] ( Figure 1 ).

[0495] It was also found that 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) ( Figure 3C and Figure 3D ). It was also found that PAD2 and PAD4 are highly expressed on the cell surface of immune cells ( Figure 4A and Figure 4B ).

[0496] 8 Example 3: Production of anti-PAD2 and anti-PAD2 antibodies

[0497] 8.1 Generation of PAD2 Antibodies

[0498] Antibody selection was performed using several human antibody libraries derived from adult naive donors. Parental antibody clones were isolated from a large single-chain Fv (scFv) human antibody library derived from spleen cells of adult naive donors and cloned into a phagemid vector based on the filamentous phage M13.

[0499] PAD2-specific scFv antibodies were isolated from phage display libraries in a series of repeated selection cycles against recombinant bacterially expressed human PAD2 with hybrid chains, including selection cycles against recombinant bacterially expressed cynomolgus monkey or mouse PAD2

[37] . The scFvs were expressed in the bacterial periplasm and screened for inhibitory activity in a trypsin cleavage assay. Screening hits, i.e., scFv clones that showed inhibition of PAD2 citrullination activity, were DNA sequenced. Unique scFvs were re-expressed in bacteria and affinity purified. Additionally, a subset of the selection outputs were subcloned for high-throughput expression and screened in 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.

[0500] Anti-PAD2 antibodies (e.g., PAD40141, PAD40119, and PAD40175, Tables 52-54) have affinities (K values) of approximately 6 nM to 260 nM for human PAD2. D )(Table 69).

[0501] The anti-PAD2 antibodies were then affinity optimized. The sequences of the resulting 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.

[0502] 8.2 Production of PAD4 Antibodies

[0503] Antibody selection was performed using several human antibody libraries derived from adult naive donors. Parental antibody clones were isolated from a large single-chain Fv (scFv) human antibody library derived from spleen cells of adult naive donors and cloned into a phagemid vector based on the filamentous phage M13.

[0504] PAD4-specific scFv antibodies were isolated from phage display libraries in a series of repeated selection cycles against recombinant mammalian-expressed human PAD4 with hybrid chains, including selection cycles against recombinant mammalian-expressed cynomolgus monkey or mouse PAD4

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

[0505] 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.

[0506] The anti-PAD4 antibodies were then affinity optimized. The sequences of the resulting anti-PAD4 antibodies are shown in Tables 43 to 54. The affinities of the PAD4 antibodies are shown in Tables 74 and 75.

[0507] PAD4 mouse surrogate antibodies were also identified, and the affinities of the surrogate antibodies are shown in Table 80. The sequences of the surrogate affinity-optimized antibodies are shown in Tables 55 and 56.

[0508] 9 Example 4: iSCAT analysis of anti-PAD4 and anti-PAD2 binding behavior

[0509] iSCAT analysis revealed that the Fab format of the anti-PAD4 affinity-optimized antibody clearly favored the heterotetrameric form, indicating that the Fab stabilized the dimeric form of PAD4. Surprisingly, the complete absence of PAD4 monomers with Fab and PAD4 dimers with only one Fab reinforced the same conclusion ( Figure 5 ,arrow).

[0510] Similar to Fab, the addition of IgG also favored heterotetramerization, indicating that IgG anti-PAD4 also stabilized the dimeric form of PAD4 ( Figure 6 Here, the larger size of the IgG (no peak overlap) allows for the detection of monomeric PAD4 species. Similar to the Fab anti-PAD4, there is a complete absence of i) PAD4 monomers with IgG and ii) PAD4 dimers with only one IgG. There is also an additional peak at approximately 600 kDa that is only detectable with IgG but not with Fab. These data indicate that a single anti-PAD4 IgG is able to span both binding sites of dimeric PAD4 ( Figure 6 ).

[0511] In summary, mass spectrometry showed that both Fab and IgG anti-PAD4 affinity-optimized antibodies strongly favored the PAD4 heterotetrameric complex. Furthermore, IgG, but not Fab, induced an additional complex that could correspond to a heterohexamer containing two PAD4 dimers and two IgGs. This configuration would enable multivalency, even in solution, making IgG substantially more potent than Fab.

[0512] Similar to PAD4, PAD2 also exhibits a monomer-dimer equilibrium ( Figure 7 ). Unlike PAD4, the addition of anti-PAD2 Fab affinity optimization antibodies produces heterodimers and heterotetramers of PAD2, indicating that the binding epitope position is different compared to the anti-PAD4 antibody. IgG PAD2 only shows a complex corresponding to a single IgG, indicating that the binding epitope is placed so that the IgG binds to the second binding site (arrow) to produce a spatial conflict. In addition, there is no peak at 600kDa, indicating that PAD2 IgG cannot form a multivalent complex as seen with PAD4. This is consistent with the citrullination assay data, which show that PAD2 has less "additional" efficacy than PAD4 IgG.

[0513] In summary, the binding of PAD4 Fab and IgG anti-PAD4 stabilizes the PAD4 dimer. This epitope allows IgG anti-PAD4 to form a hexameric complex with PAD4 dimer ( Figure 8A In contrast, although PAD2 exhibits a monomer-dimer equilibrium, anti-PAD2 binding has limited effect on dimer stabilization ( Figure 8B Furthermore, in stark contrast to PAD4, only a single IgG was able to bind to PAD2, suggesting that steric hindrance of the second binding site also prevents the potential formation of a hexameric complex.

[0514] 10 Example 5: Generation of anti-PAD2 / anti-PAD4 bispecific antibodies

[0515] Monovalent and bivalent bispecific antibodies were generated in DuetMab or Bis3 formats, respectively ( Figure 10A and Figure 10B ).

[0516] 10.1 DuetMab format

[0517] DuetMab with two different orientations was generated from affinity-optimized anti-PAD2 and anti-PAD4 antibodies (clone 22 and clone 42). Figure 10A, Table 66). The sequences of exemplary DuetMab bispecific antibodies (clones 01-06) are shown in Table 58. Clone 06 was additionally 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 eliminate Protein A binding. The Fc region was modified to contain a knob-hole mutation.

[0518] Table 66: DuetMab Orientation

[0519] λ(Pestle) κ(mortar) clone DuetMab 1st Orientation Anti-PAD2 anti-PAD4 Clone 01, Clone 03, Clone 05 DuetMab 2nd Orientation anti-PAD4 Anti-PAD2 Clone 02, Clone 04, Clone 06

[0520] 10.2Bis3 format

[0521] Generation of Bis3 format bispecific antibodies with two different orientations from affinity-optimized anti-PAD2 and anti-PAD4 antibodies (clone 22 and clone 42) Figure 10B , Table 67). The sequence of an exemplary Bis3 bispecific antibody is shown in Table 57 (clone 07-12).

[0522] Table 67: Bis3 Orientation

[0523] Fab scFv clone Bis3 1st orientation Anti-PAD2 anti-PAD4 Clone 08, Clone 10, Clone 12 Bis3 2nd orientation anti-PAD4 anti-PAD2 Clone 07, Clone 09, Clone 11

[0524] 11 Example 6: Affinity

[0525] 11.1 Affinity of Prior Art Anti-PAD2 and Anti-PAD4 Antibodies

[0526] Aosasa et al. described the affinity (K D ) in the range of 6.33 nM to 74 nM for anti-PAD2 antibodies

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

[0527] 11.2 Anti-PAD2 and Anti-PAD4 Affinity

[0528] The affinities of the affinity-optimized anti-PAD2 antibodies from Example 3 for human, cynomolgus monkey, and mouse PAD2 are shown in Table 68.

[0529] Table 68: Anti-PAD2 Fab Affinity

[0530]

[0531]

[0532] Biotinylated PAD2 was captured on a CM5 / C1-streptavidin surface and anti-PAD2 Fab flowed through. SD = standard deviation. nd = not determined.

[0533] Summary data for anti-PAD2 antibodies are shown in Tables 69 and Figure 13 . This is quite different from the prior art anti-PAD2 antibodies

[13] , where ten anti-PAD4 clones had an affinity for human PAD2 below 6 nM ( Figure 13 , Table 69). Among them, four clones had an affinity for human PAD2 lower than 1 nM ( Figure 13 , Table 69). Two of the clones had affinities below 0.3 nM for human, cynomolgus monkey, and mouse PAD2 ( Figure 13 , Table 69).

[0534] Table 69: Anti-PAD2 Fab Affinity - Summary

[0535]

[0536] Biotinylated PAD2 was captured on a CM5 / C1-streptavidin surface and anti-PAD2 Fab flowed through. Data shown are the mean of n=2-9 experiments. nd=not determined.

[0537] The affinity of the anti-PAD4 antibodies from Example 2 for human and cynomolgus monkey PAD4 is shown in Table 70.

[0538] Table 70: Anti-PAD4 Fab Affinity

[0539]

[0540]

[0541]

[0542] Biotinylated PAD4 was captured onto a CM5 / C1-streptavidin surface and anti-PAD4 Fab flowed through. SD = standard deviation. nb = no binding.

[0543] Summary data for anti-PAD4 antibodies are shown in Tables 71 and Figure 14 Affinity for human PAD4 (K D ) less than 0.1 nM and / or 1 nM in the clones Figure 14 Highlighted (dashed line).

[0544] Table 71: Anti-PAD4 Affinity - Summary

[0545]

[0546]

[0547] Biotinylated PAD4 was captured on a CM5 / C1-streptavidin surface and anti-PAD4 Fab flowed through. Data shown are the average of n = 1-6 experiments. nb = no binding.

[0548] 11.3 Bispecific Antibody Affinity

[0549] The affinity of the different bispecific 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 times higher than in the previous assay (Table 69). The recombinant PAD2 used here was different from when the affinity of the anti-PAD2 antibody from Example 3 was measured. Here, the tag on the recombinant PAD2 (and PAD4) is at the N-terminus of the protein, compared to the previous C-terminus. This could potentially lead to better folding of the protein or better presentation of the epitope, explaining the higher affinity measured in this experiment.

[0550] The affinity of the different bispecific formats for PAD4 was measured and is shown in Table 74 and summarized in Table 75.

[0551] In both orientations, the affinity of DuetMab for PAD2 and PAD4 was comparable to that of the corresponding Fab and IgG. In other words, when the optimized anti-PAD2 and anti-PAD4 antibodies were combined into the DuetMab bispecific antibody, surprisingly, there was no loss of affinity for PAD2 or PAD4 (Tables 72 to 75). D ) ranged from about 10 pM to 18 pM (Table 73). The affinity (K) of DuetMab for human PAD4 D ) in the range of approximately 40 pM-58 pM (Table 75). Fc modification had no effect on the affinity of DuetMab. In summary, all DuetMab bispecific antibodies maintained high affinity for PAD2 and PAD4 (human, cynomolgus monkey, and mouse (PAD2 only)), regardless of orientation (i.e., regardless of which of the anti-PAD2 or anti-PAD4 was the "hole" or "knob").

[0552] All Bis3 bispecific antibodies maintained high affinity for PAD2 and PAD4 (human, cynomolgus monkey, and mouse (PAD2 only)), regardless of orientation (i.e., regardless of which of the anti-PAD2 or anti-PAD4 was the scFv or Fab) (Tables 72 to 75).

[0553] Compared to IgG, the affinity of Bis3 bispecific antibodies for human PAD4 was reduced (Table 75). Bis3 formats with Fab-format PAD4 arms (e.g., clone 07) were better at maintaining affinity for both human and cynomolgus monkey PAD4. D ) ranged from about 8 pM to 50 pM (Table 75). The affinity (K) of clone 12 for human PAD4D ) was about 30 pM. The Bis3 format bispecific antibody had an affinity for human PAD2 comparable to that of an IgG containing the same PAD4 binding domain (Table 73). The affinity of DuetMab for human PAD2 (K D ) ranged from approximately 6 pM to 16 pM (Table 73). Compared to DuetMab and other Bis3 formats, the Bis3 format with a PAD4 arm in a Fab format was better at maintaining affinity for cynomolgus monkey PAD4 (Table 76). Fc modifications did not significantly affect affinity for either the Bis3 or DuetMab formats (Tables 72 to 75).

[0554] Table 72: Affinity of bispecific antibodies for PAD2

[0555]

[0556]

[0557] * = Antibody captured on the C1-Protein G' surface. # = Biotinylated antibody captured on C1-streptavidin surface. SD = Standard deviation. nd = Not determined.

[0558] Table 73: Affinity of bispecific antibodies for PAD2 - Summary

[0559]

[0560] * = Antibody captured on the C1-Protein G' surface. # = Biotinylated antibody captured on C1-streptavidin surface. Data shown are the mean of n = 2-8 experiments. nd = not determined.

[0561] Table 74: Affinity of bispecific antibodies for PAD4

[0562]

[0563]

[0564] * = Antibody captured on the C1-Protein G' surface. # = Biotinylated antibody captured on C1-streptavidin. SD = Standard deviation. nd = Not determined.

[0565] Table 75: Affinity of bispecific antibodies for PAD4 - Summary

[0566]

[0567] * = Antibodies are captured on the C1-Protein G' surface. # = Biotinylated antibody captured on C1-streptavidin surface. Data shown are the mean of n = 3-7 experiments. nd = not determined.

[0568] Table 76: Fold differences in binding affinity of bispecific antibodies

[0569]

[0570]

[0571] * = Antibody captured on the C1-Protein G' surface. # = Biotinylated antibody captured on C1-streptavidin surface. Data shown are the mean of n = 3-7 experiments. nd = not determined.

[0572] 11.4 Affinity for Different PAD2 and PAD4 Haplotypes

[0573] The two most common cynomolgus monkey haplotypes had equal affinity for the Bis3 format (e.g., clone 08) and had approximately half the affinity for the least common cynomolgus monkey haplotype (Tables 77 and 78). All cynomolgus monkey PAD4 haplotypes had affinities within 2-fold of each other for clones 08 and 06. The two most common haplotypes had equal affinity for the Bis3 and DuetMab formats (representative clones 02 and 08). The bispecific antibodies (both Bis3 and DuetMab) had equal affinity for the three human PAD4 haplotypes (Table 79).

[0574] Table 77: Affinity of Cynomolgus Monkey PAD2 Haplotypes

[0575]

[0576] Antibodies were captured onto the C1-Protein G' surface and PAD2 flowed through. SD = standard deviation.

[0577] Table 78: Affinity of Cynomolgus Monkey PAD4 Haplotypes

[0578]

[0579] Antibodies were captured onto the C1-Protein G' surface and PAD4 flowed through. SD = standard deviation.

[0580] Table 79: Affinity of human PAD4 haplotypes

[0581]

[0582] Antibodies were captured onto the C1-Protein G' surface and PAD4 flowed through. SD = standard deviation.

[0583] 11.5 Affinity of Mouse Alternative Anti-PAD4

[0584] The affinity of mouse surrogate anti-PAD4 is shown in Table 80.

[0585] Table 80: Affinity of mouse surrogate anti-PAD4 antibodies

[0586]

[0587] Biotinylated antibodies were captured onto a C1-streptavidin surface and PAD4 flowed over. SD = standard deviation.

[0588] 12 Example 7: Efficacy

[0589] 12.1 Efficacy of Prior Art Anti-PAD4 Antibodies

[0590] The anti-PAD4 antibodies (L78-4, L119-5, L198-3 and L207-11) described in WO2012026309A9 [9] showed only 10%-40% inhibition of PAD activity at an antibody concentration of 1000 nM.

[0591] WO 2016 / 155745 A1

[11] presents mouse monoclonal antibodies cross-reactive against PAD2, PAD4, and PAD3. The cross-reactive antibodies were shown to inhibit PAD2 activity by only approximately 25% compared to a control antibody, as measured using a fibrinogen citrullination assay.

[0592] As measured by H3 histone citrullination assay ( Figure 2A ) or BAEE PAD activity assay (Cayman Chemical) ( Figure 2B ), previously described humanized anti-PAD4 antibodies showed little to no inhibition of PAD4 activity in synovial fluid ( FIG. 2 ).

[0593] 12.2 Efficacy of Anti-PAD4 and Anti-PAD2 Antibodies

[0594] Using the PAD4 histone H3 citrullination potency assay, it initially proved challenging to distinguish the potency of the generated anti-PAD4 antibodies from that of the parent antibody described in Example 3. Note that using this assay, the IgG was expressed at an IC value close to the PAD4 concentration in the system (i.e., 50 ng / ml, 0.65 nM). 50 Values are stacked.

[0595] It was found that greatly 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, allowing differences in potency to be more easily assessed. Lowering 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 revised conditions, the affinity-optimized IgG showed improved potency relative to the parent. However, assay stacking was then observed in the low 100 fM range. The inventors hypothesized that this was due to bivalent binding of the antibody to the PAD4 homodimer. This hypothesis was confirmed by testing the potency of the antibody as a Fab fragment ( Figure 9A ) (See also Section 9, Example 4).

[0596] Using the improved potency assay and testing the antibodies as Fab fragments, improved potency of the affinity-optimized anti-PAD4 antibodies was observed (Table 81).

[0597] Table 81: Efficacy of anti-PAD4 antibodies

[0598] clone <![CDATA[IC 50 ]]> Fold potency increase relative to the parent PAD40048 38.7nM - 48LO0010Fab ngl-2 162.9pM 237.6 48LO0063Fab ngl-3IgG1 TM 8.9pM 4348.3

[0599] Anti-PAD4 Fab in histone H3 assay, PAD4 incubation for 3 hours 45 minutes, PAD4 concentration: 0.15 ng / ml

[0600] PAD2 and PAD4 activity in 12.3RA synovial fluid

[0601] The efficacy of anti-PAD2 and anti-PAD4 antibodies can be assessed using a histone-H3 activity assay, for example, in synovial fluid or whole blood from RA patients ( Figure 11 Both PAD2 and PAD4 contribute to PAD activity in synovial fluid from RA patients (as determined by histone-H3 activity assay) (Table 82).

[0602] Table 82: PAD2 and PAD4 levels in RA synovial fluid

[0603] RA synovial fluid PAD4 (ng / ml) PAD2 (ng / ml) Sample 1 2159 1943 Sample 2 298 21 Sample 3 719 1231 Sample 4 3 46 Sample 5 208 41

[0604] PAD4 and PAD2 levels were determined using Cayman ELISA kits.

[0605] PAD2 and PAD4 activities in whole blood of 12.4RA patients

[0606] Different concentrations of PAD2 and PAD4 can also be detected in the whole blood of RA patients (Table 83).

[0607] Table 83: PAD levels in whole blood of RA patients

[0608] RA samples CCP Status RF Status PAD4 (ng / ml) PAD2 (ng / ml) Sample 1 - - 4.2 0.7 Sample 2 - - 13.7 1.3 Sample 3 + + 78 60

[0609] 12.5 Efficacy against recombinant PAD

[0610] The potency of DuetMab and Bis3 bispecific antibodies was directly compared 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).

[0611] IC of all bispecific formats and IgG clone 22 for inhibition of histone citrullination of recombinant PAD2 50 Both are lower than the IC reported for prior art anti-PAD2 antibodies 50

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

[27] . In contrast, in an equivalent assay, the potency of DuetMab, Bis3, and Clone 22 against recombinant PAD2 (IC 50 ) were all lower than 1 nM (Table 84).

[0612] Table 84: Efficacy of bispecific antibodies (recombinant PAD2 / PAD4)

[0613]

[0614] Histone H3 assay. Human PAD4 (RD223) was used at 0.15 ng / mL; human PAD2 (RD220) was used at 0.02 ng / mL; PAD and bispecific antibody were preincubated for 30 minutes; incubated for 3 hours.

[0615] In the DuetMab format, PAD2 lost 3-4 times the potency compared to the optimized anti-PAD2 leader sequence (clone 22) in the IgG format, and PAD4 lost 7-11 times the potency compared to the optimized anti-PAD4 leader sequence (clone 42) in the IgG format (Table 84). For the Bis3 format, there was no significant potency loss or potency gain for PAD2 and PAD4 (Table 84). PAD2-Fab Bis3 (e.g., clone 08) performed better than PAD4 Fab Bis3 (e.g., clone 07) (Table 84).

[0616] 12.6 Efficacy on PAD Activity in Synovial Fluid of RA Patients

[0617] Anti-PAD2 and PAD4 Abs are effector-ineffective (TM) antibodies that show effective blocking of extracellular PAD activity. Surprisingly, blocking PAD2 and PAD4 in the synovial fluid of RA patients showed that PAD2 and PAD4 activities are non-redundant. This suggests that blocking PAD2 and PAD4 with anti-PAD2 and anti-PAD4 inhibits all PAD activity in RA synovial fluid (Figure 12).

[0618] Efficacy analysis of DuetMab in both orientations showed that the bispecific antibodies inhibited PAD activity (PAD2 and PAD4 combined) in synovial fluid (Table 85).

[0619] Table 85: DuetMab Efficacy (Synovial Fluid)

[0620] ID describe Potency pM (CI95) 48LO0063 / 141LO0035 anti-PAD2 / anti-PAD4 6.2(4.1-8.7) PAD240002 PAD2(mortar)-PAD4(pestle) 130(79-200) PAD240001 PAD4(mortar)-PAD2(pestle) 88(49-155)

[0621] Histone H3 activity assay, synovial fluid, 1:2000 dilution; CI95: 95% confidence interval

[0622] Efficacy analysis of the Bis-3 Ab format in both orientations showed that the bispecific antibodies inhibited the combined PAD2 and PAD4 activity in synovial fluid (Table 86). The ability of the Bis3 bispecific antibodies to inhibit the combined PAD activity in synovial fluid from RA patients was high and dose-dependent.

[0623] 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).

[0624] Table 86: Bis3 efficacy (synovial fluid)

[0625] name Format scFv Fab Potency (pM) (CI95) PAD240007 Bis3 PAD2 PAD4 21(19-24) PAD240008 Bis3 PAD4 PAD2 5(1.9-8.5) PAD240009 Bis3 PAD2 PAD4 30(24-37) PAD240010 Bis3 PAD4 PAD2 8(6-11) PAD240011 Bis3 PAD2 PAD4 26(21-31) PAD240012 Bis3 PAD4 PAD2 10(8.4-11.7) 141LO0035&48LO0063 IgG 16

[0626] Histone H3 activity assay, synovial fluid, 1:2000 dilution; CI95: 95% confidence interval

[0627] Table 87: Comparative Bis3 efficacy (synovial fluid)

[0628] Format Fab Potency (pM) Bis3 PAD4 21-30 Bis3 PAD2 5-8

[0629] 12.7 Potency on PAD Activity in Whole Blood

[0630] The efficacy of the bispecific antibodies was also evaluated in plasma samples from whole blood. The Bis3 format was more effective than the DuetMab format (clone 12, PAD240012) in this assay ( Figure 25BBoth Duet and Bis3 format bispecific antibodies were able to completely inhibit PAD activity in synovial fluid and whole blood ( FIG. 25 ).

[0631] Table 88: Efficacy of bispecific antibodies (whole blood)

[0632]

[0633]

[0634] Histone-H3 PAD activity assay: antibodies were incubated overnight in RA whole blood.

[0635] 12.8 In Vivo Efficacy

[0636] The in vivo efficacy of the Bis3 PAD2 / PAD4 bispecific antibody was evaluated at low and high doses ( Figure 22A The Bis3 bispecific antibody was able to suppress endogenous PAD activity in vivo, with rapid target engagement and almost complete target engagement by day 57 at low doses of the Bis3 bispecific antibody, and PAD activity returned to pre-dose levels by approximately day 85. At high doses, no detectable PAD activity recovery was observed at the completion of the study (day 106). Figure 22C 、 Figure 22E ).

[0637] Bis3 bispecific antibody was able to suppress PAD activity in spiked plasma, with rapid and nearly complete target engagement by day 29 at low doses of Bis3 bispecific antibody, and PAD activity returned to pre-dose levels by approximately day 85. At high doses, no detectable PAD activity recovery was observed at study completion (day 106). Figure 22B 、 Figure 22D ).

[0638] 12.9 Efficacy against Different Haplotypes

[0639] The Bis3 format retained similar potency (within 2-fold) against haplotypes of human and cynomolgus PAD4 and cynomolgus PAD2 (Table 89).

[0640] Table 89: Bis3 potency against PAD2 and PAD4 haplotypes (IC 50 )

[0641]

[0642]

[0643] Bis3 = scFv(PAD4)-Fab(PAD2), clone 12

[0644] 12.10 Efficacy of Anti-PAD2 and Anti-PAD4 Antibodies

[0645] 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] .

[0646] The optimized histone-H3 citrullination ELISA (see Section 12.2) and recombinant PAD2 and PAD4 (respectively Figure 23A and Figure 23B ), the potency of clones 12, 22, and 42 was directly compared to mAb2, G8H4, H7H4, and 4R147.

[0647] Clone 12 and clone 22 completely inhibited hPAD2 enzyme activity, while anti-PAD2 antibody mAb 2 only partially inhibited the activity. Clone 12 and clone 44 completely inhibited hPAD4 enzyme activity, while anti-PAD4 antibodies G8H4, H7H4 and 4R147 did not show any inhibitory activity on hPAD4 enzyme activity up to 50 nM.

[0648] Similarly, the potency of clone 12 was directly compared with mAb 2, G8H4, H7H4, and 4R147 in synovial fluid using an optimized histone-H3 citrullination ELISA (Figure 26). Clone 12 completely inhibited PAD enzyme activity, while none of the other antibodies showed any inhibitory activity.

[0649] 13 Example 8: Specificity

[0650] WO 2016 / 155745 A1

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

[0651] Affinity-optimized anti-PAD4 antibodies and bispecific formats are specific for PAD2 and / or PAD4 and do not bind PAD3 ( Figure 24A ) or PAD1( Figure 24B ), as assessed by specific ELISA assay (Section 6.2).

[0652] 14 Example 9: Binding behavior of bispecific antibodies

[0653] The binding properties of two DuetMab orientations (PAD2λ / PAD4κ and PAD4λ / PAD2κ) were evaluated. Both DuetMab orientations showed similar binding behavior. Figure 5 , Figure 8), Duet also stabilizes PAD4 dimers ( Figure 15 ). In contrast, but similar to PAD2IgG and Fab ( Figure 7, Figure 8 ), DuetMab binds to PAD2, indicating that the epitope position of PAD2 is different ( Figure 15 Constant region modifications (YTE, TM, FQQ) had no effect on the binding properties of DuetMab.

[0654] All Bis3 orientations have similar binding behaviors with PAD2 and PAD4 ( Figure 16 ), as assessed by iSCAT (Section 6.3). Similar to PAD4 IgG, the Bis3 format favors a highly stable hexameric complex, i.e., 2x Bis3 + 2x PAD4 dimer. Fc mutations had no effect on binding behavior ( Figure 16 In contrast, but similar to PAD2 IgG, Bis3 has no effect on the PAD2 dimer state. In addition, steric hindrance appears to block tetramer formation ( Figure 16 , dashed line), as previously observed for PAD2 IgG, but quite different from DuetMab ( Figure 15 ).

[0655] iSCAT data showed that fibrillation was possible with the Bis3 format, but not with DuetMab ( Figure 15 and Figure 16 These data show that the Bis3 format with the PAD2 arm in the Fab format has a lower propensity to form high complexes ( Figure 16 , "PAD2Bis-TM-YTE 10"), thus the Bis3 format with the PAD2 arm in Fab format has a lower risk of aggregation. As expected, the Fc modification had no effect on the binding activity ( Figure 15 ; Figure 16 , for example compare PAD2 Bis-TM-YTE 07 and PAD2 Bis-TM-YTE 11).

[0656] In the presence of PAD4 alone, the Bis3 construct, but not Duet, enabled the formation of stable hexamers, but also higher order complexes ( Figure 15 and Figure 16 In contrast, in the presence of PAD2 alone, no hexamers or higher-order complexes were observed. Thus, the dual PAD4 interaction is the primary driver of oligomers, including highly stable hexamers and potentially higher-order complexes, suggesting that fibrillation is possible. DuetMab was also able to form stable hexamers in the presence of both PAD4 and PAD2, but no higher-order complexes were detected ( Figure 15 and Figure 16 The double interactions seen with the Bis3 format are not possible with the DuetMab format ( Figure 17), therefore, no higher order complexes were observed (Table 90).

[0657] Table 90: Potential for formation of higher order complexes

[0658]

[0659]

[0660] 15 Example 10: Stability

[0661] 15.1 Thermal Stability

[0662] Aggregation is closely related to the thermal stability of antibody molecules. The lower the thermal stability, the less stable the product and the higher the degree of aggregation, while higher thermal stability of the product can reduce the degree of aggregation. Using standard protocols, thermal stability was analyzed using Promethus Nano-Differential Scanning Fluorescence (Nano-DSF).

[0663] The YTE Fc mutation (M252Y / S254T / T256E) extends the half-life of the antibody molecule. The triple mutation (TM, L234F / L235E / P331S) reduces the effector function of the constant region. The TM and YTE mutations were introduced into the Bis3 and DuetMab bispecific antibodies, and their sequences are provided in Tables 57 and 58. 起始 Stability data demonstrated that the TM-YTE Fc mutation was destabilizing compared to bispecific antibodies with only the TM modification for both Bis3 and DuetMab formats (Table 91, clone 09, clone 10, clone 03, and clone 04). Alternative effector null modifications (FQQ, L234F / L235Q / K322Q) did not exhibit the same instability regardless of the bispecific format (clone 05, clone 06, clone 11, and clone 12) ( Figure 18 ).

[0664] For TM and FQQ-YTE DuetMab and Bis3 formats, see the maximum onset temperature (T 起 Clone 01, clone 02, clone 05, clone 06, clone 07, clone 08, clone 11, and clone 12 all had a T greater than about 48°C. 起始 ( Figure 18 ,dotted line).

[0665] Table 91: Thermal stability of bispecific antibodies

[0666]

[0667]

[0668] 15.2 Aggregation

[0669] Accelerated stability studies using HP-SEC compared stability at 40°C and 45°C with stability at 4°C (Table 92, Figure 19). Surprisingly, the DuetMab format showed greater stability at 45°C than all Bis3 formats. In the context of TM Fc modification, the Bis3 clone 09 format showed lower aggregation at 45°C. In the context of TM-YTE modification, Bis3 showed the highest aggregation regardless of Bis3 orientation at 45°C. In the context of FQQ-YTE Fc modification, the clone 11 Bis3 format showed the least aggregation at 45°C.

[0670] Surprisingly, at 40°C, the DuetMab format showed minimal aggregation in the context of either TM-YTE or FQQ-YTE Fc modification, regardless of DuetMab orientation. In the context of FQQ-YTE Fc modification, the PAD2 (hole)-PAD4 (knob) orientation showed aggregation comparable to the corresponding IgG (TM).

[0671] For the Bis3 format, acceptable aggregation was present regardless of Bis3 orientation in the context of FQQ-YTE modification. The PAD2-Fab / PAD4-scFv (clones 08, 10, and 12) Bis3 format showed minimal aggregation regardless of Fc modification.

[0672] Table 92: Acceleration Stability

[0673]

[0674]

[0675] 2-week study, antibody concentration was approximately 1 mg / ml

[0676] 15.3 Stability Conclusion

[0677] Stability data surprisingly showed that the bispecific format with TM-YTE Fc modification exhibited decreased thermal stability, which was not observed for bispecific antibodies with TM or FQQ-TM modifications ( Figure 18 Stability data further showed that, regardless of the orientation of the Fc-modified DuetMab or Bis3, the DuetMab format was surprisingly less prone to aggregation than the Bis3 format. In addition, the PAD2-Fab / PAD4-scFv (clones 08, 10, and 12) Bis3 format exhibited minimal aggregation regardless of the Fc modification.

[0678] 16 Example 11: Security

[0679] The Bis3 format is 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 found ( Figure 20 PAD2 was detected in the CNS but had no effect on enhancing behavioral observations in cynomolgus monkey studies. No further effects (e.g., adverse cytokine release) were observed in blood from healthy and RA donors ( Figure 20 ).

[0680] No further safety concerns (e.g., adverse cytokine release) were observed in blood from healthy and RA donors after exposure to either Bis3 or DuetMab formats (clone 12 and clone 06) ( Figure 21 None of the tested samples had cytokine levels above the upper limit of quantification of 20,000.0 pg / ml.

[0681] 17 Example 12: Effector null mutation

[0682] Constant region modifications had no effect on the potency or PAD2 / PAD4 affinity of the bispecific molecules. The triple mutation (TM) (L234F / L235E / P331S) abolished Fc effector function (Table 93). The YTE mutation (M252Y / S254T / T256E) increased the half-life of the antibody. YTE was also shown to partially rescue Fc affinity for huFcRn in the context of TM mutations. FQQ (L234F / L235Q / K322Q) is an effective alternative null mutation for TM.

[0683] Table 93: Affinity for huFcRn

[0684]

[0685] Affinity chromatography (Section 6.7).

[0686] References

[0687] All publications mentioned in this specification are herein incorporated by reference.

[0688] [1]ME Weinblatt et al., New England Journal of Medicine 340,253(1999).

[0689] [2] S. Mpofu, F. Fatima, and RJ Moots, Rheumatology 44, 271 (2005).

[0690] [3]EHChoy,AFKavanaugh,and SAJones,Nat Rev Rheumatol 9,154(2013).

[0691] [4]JLNam,Current Opinion in Rheumatology 28,267(2016).

[0692] [5] MHPillinger and SBAbramson,Rheumatic Disease Clinics of North America 21,691(1995).

[0693] [6]M.Nagashima et al.,Rheumatology International 18,113(1998).

[0694] [7]N.Thieblemont et al.,Semin Immunol 28,159(2016).

[0695] [8]KLBicker and PRThompson,Biopolymers 99,155(2013).

[0696] [9]M.Sato et al.,WO2012026309A9(10May 2012).

[0697]

[10] CHNielsen and D. Damgaard, WO2014086365A1(12 June 2014).

[0698]

[11] CHNielsen and D. Damgaard, WO2016155745A1 (6 October 2016).

[0699]

[12] Sato et al.,WO2016143753A1(September 15, 2016).

[0700]

[13] M.Aosasa et al.,J Immunol Res 2021.6659960(2021).

[0701]

[14] E.A.Kabat,Sequences of Proteins of Immunological Interest(U.S.Department of Health and Human Services,Public Health Service,NationalInstitutes of Health,1991).

[0702]

[15] P.Tsui et al.,WO2015175874(A2)(19 November 2015).

[0703]

[16] W.Dall’acqua,L.S.Johnson,and E.S.Ward,WO2002060919A2(8August2002).

[0704]

[17] P.Tsui,M.BORROK,and W.Dall’acqua,WO2013165690A1(7November 2013).

[0705]

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

[0706]

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

[0707]

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

[0708]

[21] F. G.Himmler,and G.Wozniak-Knopp,WO2008003103A2(10January 2008).

[0709]

[22] C.Wu,WO2015103072A1(9 July 2015).

[0710]

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

[0711]

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

[0712]

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

[0713]

[26] E.A.Kabat and T.T.Wu,Ann N Y Acad Sci 190,382(1971).

[0714]

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

[0715]

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

[0716]

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

[0717]

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

[0718]

[31] P.J.Carter,L.G.Presta,and J.B.Ridgway,US8216805B2(10 July2012).

[0719]

[32] P.J.Carter,L.G.Presta,and J.B.Ridgway,US5731168A(24 March1998).

[0720]

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

[0721]

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

[0722]

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

[0723]

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

[0724]

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

[0725]

[38] M.Miyamoto et al.,WO2022176970(25 August 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. The antibody of claim 1, wherein the antibody inhibits PAD activity, optionally wherein the antibody inhibits PAD-mediated protein citrullination.

3. The antibody of claim 2, wherein the antibody inhibits PAD activity in synovial fluid, optionally as measured by an H3 citrullination assay, IC 50 ≤200pM.

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

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

6. The antibody of any preceding claim, wherein the antibody inhibits PAD2 and / or PAD4 in immune cells.

7. The antibody according to 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 preceding claim, wherein the PAD2 is human PAD2.

10. The antibody of any preceding claim, wherein the PAD4 is human PAD4.

11. The antibody of any preceding claim, wherein the PAD2 binding domain does not specifically bind PAD3, optionally as measured by PAD3 ELISA.

12. The antibody of any preceding claim, wherein the PAD4 binding domain does not specifically bind to PAD3, optionally as measured by PAD3 ELISA.

13. The antibody of any preceding claim, wherein the antibody does not specifically bind PAD3, optionally as measured by PAD3 ELISA.

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

15. The antibody of any preceding claim, wherein the PAD2 binding domain does not specifically bind to PAD1, optionally as measured by PAD1 ELISA.

16. The antibody of any preceding claim, wherein the PAD4 binding domain does not specifically bind to PAD1, optionally as measured by PAD1 ELISA.

17. The antibody of any preceding claim, wherein the antibody does not specifically bind 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. The antibody of any preceding claim, wherein the PAD2 domain specifically binds mouse PAD2.

20. The antibody of any preceding claim, wherein the PAD4 domain specifically binds mouse PAD4.

21. The antibody of any preceding claim, wherein the PAD2 domain specifically binds cynomolgus monkey PAD2.

22. The antibody of any preceding claim, wherein the PAD4 domain specifically binds 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 of claim 19, wherein the antibody has a K of X pM. D Binds to PAD2, wherein X < K of a bivalent Fab fragment or IgG comprising the PAD2 binding domain D .

25. The antibody of claim 19, wherein the antibody has a K of X pM. D Binds to PAD4, wherein X < K of a bivalent Fab fragment or IgG comprising the PAD4 binding domain D .

26. The antibody according to any one of claims 19 to 21, wherein the antibody binds with an affinity (K of ≤ 20 pM or ≤ 18 pM). D ) binds to PAD2.

27. The antibody according to claims 19 to 22, wherein the antibody binds with an affinity (K of ≤ 60 pM) D ) binds to PAD4.

28. The antibody according to any one of claims 19 to 23, wherein the T of the antibody is 起始 ≥40℃, optionally where T 起始 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. 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 wherein the antibody comprises two PAD4 binding domains such that the antibody is bivalent for PAD4.

32. The antibody of any one of claims 25 to 27, wherein each PAD2 binding domain is contained within a Fab domain and each PAD4 binding domain is contained within a scFv.

33. The antibody of any one of claims 25 to 27, wherein each PAD4 binding domain is contained within a Fab domain and each PAD2 binding domain is contained within a scFv.

34. The antibody according to claims 25 to 29, wherein the antibody binds with an affinity (K of ≤ 17 pM) D ) binds to human PAD2.

35. The antibody according to claim 25 or 30, wherein the antibody binds with an affinity (K of ≤ 35 pM) D ) binds to human PAD4.

36. The antibody of any one of claims 25 to 31 , wherein the antibody is a bivalent bispecific antibody comprising: a) an IgG comprising a first Fab domain and a second Fab domain and an Fc domain, wherein the first Fab domain and the second Fab domain each comprise a PAD2 binding domain that specifically binds to PAD2, and b) a first scFv and a second scFv, wherein the first scFv and the second scFv are each respectively linked to the carboxyl terminus of one of the heavy chains of the Fc domain of the IgG, and wherein the first scFv and the second scFv each comprise a PAD4 binding domain that specifically binds to PAD4.

37. The antibody of claim 32, wherein the first scFv PAD4 binding domain and the second scFv PAD4 binding domain comprise SEQ ID NO:

39.

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

34.

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

36.

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

35.

41. The antibody of any one of claims 25 to 31, wherein the antibody is a bivalent bispecific antibody comprising: a) an IgG comprising a first Fab domain and a second Fab domain and an Fc domain, wherein the first Fab domain and the second Fab domain each comprise a PAD4 binding domain that specifically binds to PAD4, and b) a first scFv and a second scFv, wherein the first scFv and the second scFv are each respectively linked to the carboxyl terminus of one of the heavy chains of the Fc domain of the IgG, and wherein the first scFv and the second scFv each comprise a PAD2 binding domain that specifically binds to PAD2.

42. The antibody of claim 37, wherein the first scFv PAD2 binding domain and the second scFv PAD2 binding domain comprise SEQ ID NO:

38.

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

40.

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

37.

46. The antibody of any one of claims 37 to 41, wherein the first Fab domain and the second Fab domain 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 carboxyl terminus of the heavy chain via a peptide linker.

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

51.

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

33.

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

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

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

53. The antibody of claim 48, wherein the antibody comprises an IgG comprising: a) comprising a first binding region of a first Fab, wherein said second Fab domain comprises said PAD2 binding domain, and b) a second binding region comprising a second Fab, wherein the second Fab comprises the PAD4 binding domain.

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

63.

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

64.

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

62.

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

65.

58. The antibody of any one of claims 1 to 18, wherein the antibody comprises an IgG or F(ab')2 fragment.

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

60. 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. 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, wherein the Fab fragment comprises the PAD2 binding domain or the PAD4 binding domain.

63. The antibody of claim 58, wherein the Fab fragment comprises the PAD2 binding domain, and wherein the Fab binds to the PAD2 domain with an affinity (K) of ≤ 20 nM, ≤ 10 nM, ≤ 6 nM, or ≤ 1 nM. D ) binds to human PAD2.

64. The antibody of claim 58, wherein the Fab fragment comprises the PAD4 binding domain, and wherein the Fab binds to the PAD4 domain with an affinity (K) of ≤1 nM, ≤0.1 pM, ≤0.07 pM, or ≤0.05 pM. D ) binds to human PAD4.

65. according to the antibody described in any one of claim 22, 23, 30, 31, 59 or 60, wherein said K D is measured by Biacore.

66. The antibody of any preceding claim, wherein 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, and wherein: 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 LCDR3 amino acid sequence is SEQ ID NO: 12, optionally with 1, 2 or 3 amino acid changes.

67. The antibody of any preceding claim, wherein the PAD2 binding domain comprises a VH domain comprising a sequence at least 90% identical to SEQ ID NO:

1.

68. The antibody of any preceding claim, wherein the PAD2 binding domain comprises a VH domain comprising SEQ ID NO:

1.

69. The antibody of any preceding claim, wherein the PAD2 binding domain comprises a VL domain comprising a sequence at least 90% identical to SEQ ID NO:

2.

70. The antibody of any preceding claim, wherein the PAD2 binding domain comprises a VL domain sequence comprising SEQ ID NO:

2.

71. The antibody of any preceding claim, wherein the PAD2 binding domain comprises a VH domain sequence comprising SEQ ID NO: 1, optionally with 1, 2, 3, 4 or 5 amino acid changes outside of the CDRs.

72. The antibody of any preceding claim, wherein the PAD2 binding domain comprises a VL domain sequence comprising SEQ ID NO: 2, optionally with 1, 2, 3, 4 or 5 amino acid changes outside of the CDRs.

73. The antibody of any preceding claim, wherein 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, and wherein: 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) the LCDR3 amino acid sequence is SEQ ID NO:

26.

74. The antibody of any preceding claim, wherein the PAD4 binding domain comprises a VH domain comprising a sequence at least 90% identical to SEQ ID NO:

31.

75. The antibody of any preceding claim, wherein the PAD4 binding domain comprises a VH domain sequence comprising SEQ ID NO:

31.

76. The antibody of any preceding claim, wherein the PAD4 binding domain comprises a VL domain comprising a sequence at least 90% identical to SEQ ID NO:

32.

77. The antibody of any preceding claim, wherein the PAD4 binding domain comprises a VL sequence domain comprising SEQ ID NO:

32.

78. The antibody of any preceding claim, wherein the PAD4 binding domain comprises a VH domain sequence comprising SEQ ID NO: 31, optionally with 1, 2, 3, 4 or 5 amino acid changes outside of the CDRs.

79. The antibody of any preceding claim, wherein the PAD4 binding domain comprises a VL domain sequence comprising SEQ ID NO: 32, optionally with 1, 2, 3, 4 or 5 amino acid changes outside of the CDRs.

80. The antibody of any preceding claim, 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. The antibody of claim 77, wherein the Fc domain comprises mutations L234F, L235Q, and K322Q.

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

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

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

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

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

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

89. The antibody of claim 76, wherein the Fc domain comprises a C sequence comprising SEQ ID NO: 47, SEQ ID NO: 48, or SEQ ID NO:

49. H 2 domains.

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

65.

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

65.

93. The antibody of claim 47, wherein: 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, and b) the PAD4 binding domain comprises a heavy chain comprising SEQ ID NO: 61 and a light chain comprising SEQ ID NO:

65.

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

63.

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

63.

96. The antibody of claim 47, wherein: a) the PAD2 binding domain comprises a heavy chain comprising SEQ ID NO: 72 and a light chain comprising SEQ ID NO: 64, and b) 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 the antibody according to any preceding claim.

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

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

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

101. 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 the antibody of any one of claims 1 to 92 or the pharmaceutical composition of claim 98, optionally comprising instructions for use.

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

105. A method of treating a disease in a subject, the method 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 wherein the anti-PAD2 antibody is the antibody of claim 62.

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

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

110. The antibody of any one of claims 1 to 92 or the pharmaceutical composition of claim 98, for use in a method of 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. Use of the antibody according to any one of claims 1 to 92 or the pharmaceutical composition according to claim 98 for the manufacture of a medicament for the treatment of an autoimmune disorder.

Images