WO2025229206A1

WO2025229206A1 - Monovalent anti-iga binding molecules and methods of use

Info

Publication number: WO2025229206A1
Application number: PCT/EP2025/062121
Authority: WO
Inventors: Sofie VOET; Sofie GABRIËLS; René Bigirimana; Christophe Blanchetot; Karen Silence; Jelke VERBEIREN; Lore DEWEIRDT
Original assignee: ArgenX BVBA
Current assignee: ArgenX BVBA
Priority date: 2024-05-03
Filing date: 2025-05-02
Publication date: 2025-11-06
Anticipated expiration: 2026-11-03
Also published as: WO2025229224A1

Abstract

The present invention relates to monovalent antigen-binding molecules that bind to IgA, such as one-armed antibodies, and their use in the treatment of disorders such as IgA mediated disorders. The monovalent antigen-binding molecules comprise an antigen-binding domain that binds to IgA and a variant Fc region that binds to the human neonatal Fc receptor with increased affinity relative to a wild-type Fc region. The monovalent antigen-binding molecules of the invention also bind IgA in a pH-dependent manner.

Description

MONOVALENT ANTI-IGA BINDING MOLECULES AND METHODS OF USE

FIELD OF THE INVENTION

The present invention relates to monovalent antigen-binding molecules that bind to IgA, such as one-armed antibodies, and their use in the treatment of disorders. The monovalent antigen-binding molecules comprise an antigen-binding domain that binds to IgA and a variant Fc region that binds to the human neonatal Fc receptor (hFcRn) with increased affinity relative to a wild-type Fc region. The monovalent antigen-binding molecules comprise a variant Fc region incorporating ABDEG™ technology or YPY mutations. The monovalent antigen-binding molecules of the invention bind IgA in a pH- dependent manner. The invention has utility in the treatment of IgA-mediated disorders.

BACKGROUND TO THE INVENTION

Immunoglobulin A (IgA) is the most prevalent antibody class produced in humans. The daily production rate of IgA is around 66 mg/kg, which is higher than all other antibody isotypes combined. Given this abundance, it has been postulated that IgA plays a significant role in immune defence (Breedveld and van Egmond, 2019; Aleyd, Heineke, and van Egmond, 2015).

In the mucosal membranes, the level of IgA is greater than all other types of antibody combined and IgA plays an important role in passive immunity in such areas. The role of IgA in the mucosal areas, such as the intestinal lumen, is relatively well-characterised. Dimeric IgA in the form of secretory IgA (slgA) is secreted by epithelial cells into extracellular secretions (such as mucus). slgA in extracellular secretions is a first line of immune defence acting as a barrier against pathogens and commensals by preventing colonisation and penetration of the mucosal epithelium so as to avoid infection and antigen leakage into the systemic circulation.

IgA is also present in the serum, typically in a monomeric form. Serum IgA is capable of binding to receptors such as FcoRI (also known as CD89), which mediates effector functions in order to initiate an inflammatory response. FcoRI is a member of the Fc receptor immunoglobulin superfamily and is expressed on cells from the myeloid lineage (including monocytes, macrophages, Kupffer cells, eosinophils, neutrophils, and certain subsets of dendritic cells) as well as on platelets (Monteiro, Kubagawa, and Cooper, 1990; Hostoffer, Krukovets, and Berger, 1993; Qian et al, 2008). Once FcoRI binds to IgA immune complexes (i.e. opsonized pathogens), there is cross-linking and an induction of pro-inflammatory responses.

The presence of excessive IgA immune complexes, IgA autoantibodies or serum IgA is thought to lead to uncontrolled and disproportionate immune cell activation, which in turn, can lead to severe tissue damage in autoimmune diseases.

SUMMARY OF THE INVENTION

There is a need to develop agents that target IgA since elevated levels of IgA and IgA immune complexes are implicated in disease. As of the filing date of this application, there are no approved products that modulate IgA that have been proven useful to treat IgA- mediated disease. The present invention addresses this need by providing monovalent antigen-binding molecules that bind to IgA, otherwise referred to as monovalent IgA- binding molecules.

The present invention provides monovalent antigen-binding molecules that are particularly suited to the treatment of IgA-mediated disorders. Advantageously, the monovalent antigen-binding molecules are able to reduce the levels of serum IgA, prevent IgA binding to its receptors (such as FcoRI) and also displace receptor-bound IgA e.g. FcoRI-bound IgA.

In further detail, the present application exemplifies monovalent antigen-binding molecules of the invention and their ability to reduce serum IgA levels in vivo. As reported herein, the monovalent antigen-binding molecules of the invention bind to IgA and the resulting complex can be internalised via neonatal Fc receptor (FcRn)-mediated cellular uptake. Once internalised, the serum IgA molecules are subjected to lysosomal degradation thereby reducing serum IgA levels. The monovalent antigen-binding molecules are recycled back to the extracellular space. The recycling of the monovalent antigen-binding molecules allows the process of serum IgA capture and degradation to be repeated by the same monovalent antigen-binding molecule. This process is referred to herein as “IgA sweeping”.

Surprisingly, the reduction of IgA observed with the monovalent antigen-binding molecules of the invention (e.g., one-armed modified antibodies) is greater as compared with bivalent IgA-binding molecules (e.g., conventional two-armed antibodies). The monovalent antigenbinding molecules of the invention improve both the speed of serum IgA reduction as well as the amount of serum IgA removal as compared to bivalent antigen-binding counterparts.

This finding is unexpected because whilst a monovalent antigen-binding molecule can only bind a single IgA antigen at any given time, a bivalent IgA-binding molecule can bind two IgA antigens simultaneously. It is therefore surprising that the monovalent antigen-binding molecules of the invention reduce serum IgA to a greater extent than bivalent IgA-binding molecules.

The results reported herein demonstrate that FcRn receptor occupancy is higher with monovalent antigen-binding molecules of the invention as compared with bivalent IgA- binding molecules. Without wishing to be bound by theory, it is believed that monovalent antigen-binding molecules reduce steric hinderance as compared to bivalent IgA-binding molecules and thereby improve access to FcRn receptor. The improved accessibility of a monovalent antigen-binding molecule-serum IgA complex to FcRn results in enhanced IgA sweeping.

Taking into account the above, the monovalent antigen-binding molecules of the invention exhibit improved IgA sweeping and are advantageous for use in the treatment of IgA- mediated diseases.

In a first aspect, the present invention provides a monovalent antigen-binding molecule comprising:

- an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences: HCDR3 comprising SEQ ID NO: 3; HCDR2 comprising SEQ ID NO: 2; HCDR1 comprising SEQ ID NO: 1 ; LCDR3 comprising SEQ ID NO: 6; LCDR2 comprising SEQ ID NO: 5; and LCDR1 comprising SEQ ID NO: 4; and

- a variant Fc region, wherein the variant Fc region comprises a first Fc domain and a second Fc domain; and wherein the first Fc domain and the second Fc domain each comprise the amino acids Y, T, E, K, F and Y at EU positions 252, 254, 256, 433, 434 and 436, respectively.

In a second aspect, the present invention provides a monovalent antigen-binding molecule comprising:

- an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences:

HCDR3 comprising SEQ ID NO: 3; HCDR2 comprising SEQ ID NO: 2; HCDR1 comprising SEQ ID NO: 1 ; LCDR3 comprising SEQ ID NO: 6; LCDR2 comprising SEQ ID NO: 5; and LCDR1 comprising SEQ ID NO: 4; and

- a variant Fc region, wherein the variant Fc region comprises a first Fc domain and a second Fc domain; and wherein the first Fc domain and the second Fc domain each comprise the amino acids Y, P and Y at EU positions 252, 308, and 434, respectively.

In some embodiments, the antigen-binding domain is a Fab.

In some embodiments, the antigen-binding domain is attached to the N-terminus of the first Fc domain. In other embodiments, the antigen-binding domain is attached to the N- terminus of the second Fc domain.

In some embodiments, the antigen-binding domain is a Fab and the C-terminus of the Fab heavy chain is attached to the N-terminus of the first Fc domain via an IgG hinge region. In some embodiments, the antigen-binding domain is a Fab and the C-terminus of the Fab heavy chain is attached to the N-terminus of the second Fc domain via an IgG hinge region.

In some embodiments, the antigen-binding domain comprises a variable heavy chain domain (VH) comprising or consisting of the amino acid sequence of SEQ ID NO: 7 or an amino acid sequence having at least 80%, 90%, 95%, 98%, 99% identity thereto, and a variable light chain domain (VL) comprising or consisting of the amino acid sequence of SEQ ID NO: 8 or an amino acid sequence having at least 80%, 90%, 95%, 98%, 99% identity thereto.

In some embodiments, the antigen-binding domain comprises a variable heavy chain domain (VH) consisting of SEQ ID NO: 7 and a variable light chain domain (VL) consisting of SEQ ID NO: 8.

The variant Fc region is, in preferred embodiments, a variant human Fc region. In most preferred embodiments, the variant Fc region is a variant IgG 1 Fc region.

In some embodiments, the variant Fc region further comprises at least one additional amino acid substitution as compared with the corresponding wild-type Fc region; and the at least one additional substitution reduces or eliminates Fc effector function.

In some embodiments, the first Fc domain and second Fc domain comprise the amino acids:

(i) A and A at EU positions 234, 235; and

(ii) G at EU position 329.

In some embodiments, the variant Fc region further comprises at least one additional amino acid substitution as compared with the corresponding wild-type Fc region; and the at least one substitution promotes dimerisation between the first Fc domain and the second Fc domain.

In some embodiments, the first Fc domain and the second Fc domain comprise knob-into- holes amino acid substitutions. In some embodiments, the first Fc domain comprises the amino acid W at EU position 366; and the second Fc domain comprises the amino acids S, A and V at EU positions 366, 368 and 407, respectively.

In some embodiments, the first Fc domain comprises the amino acids A, A, G, Y, T, E, W, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 366, 433, 434 and 436, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, T, E, K, F, Y, S, A and V at EU positions 234, 235, 329, 252, 254, 256, 433, 434, 436, 366, 368 and 407, respectively. In some embodiments, the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 73 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 74.

In some embodiments, the first Fc domain comprises the amino acids A, A, G, Y, P, Y and W at EU positions 234, 235, 329, 252, 308, 434 and 366, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, P, Y, S, A and V at EU positions 234, 235, 329, 252, 308, 434, 366, 368 and 407, respectively.

In some embodiments, the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 114 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 115.

In some embodiments, the antigen-binding domain is a Fab attached to the first Fc domain via an IgG hinge region.

In some embodiments, the first Fc domain and the second Fc domain do not comprise an N-linked glycan at EU position 297.

In some embodiments, the first Fc domain and second Fc domain comprise an afucosylated N-linked glycan at EU position 297.

In some embodiments, the first Fc domain and second Fc domain comprise an N-linked glycan having a bisecting GIcNac at EU position 297 of the Fc domains.

In some embodiments, the monovalent antigen-binding molecule is a modified lgG1 antibody having only one Fab arm.

In some embodiments, the monovalent antigen-binding molecule consists of:

(i) a first immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 75;

(ii) a second immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 76; and

(iii) an immunoglobulin light chain consisting of the amino acid sequence of SEQ ID NO: 77. In some embodiments, the monovalent antigen-binding molecule consists of:

(i) a first immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 116;

(ii) a second immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 117; and

(iii) an immunoglobulin light chain consisting of the amino acid sequence of SEQ ID NO: 77.

In another aspect, the invention provides a monovalent antigen-binding molecule that binds to IgA, consisting of:

(iii) an immunoglobulin light chain consisting of the amino acid sequence of SEQ ID NO: 77, wherein the first immunoglobulin heavy chain pairs with the second immunoglobulin heavy chain to form a variant Fc region, and wherein the first immunoglobulin heavy chain pairs with the immunoglobulin light chain to form a Fab that binds to IgA.

In another aspect, the invention provides a monovalent antigen-binding molecule that binds to IgA consisting of:

(iii) an immunoglobulin light chain consisting of the amino acid sequence of SEQ ID NO: 77, wherein the first immunoglobulin heavy chain pairs with the second immunoglobulin heavy chain to form a variant Fc region, and wherein the first immunoglobulin heavy chain pairs with the immunoglobulin light chain to form a Fab that binds to IgA. Further provided herein are isolated polynucleotides or polynucleotides, which encode the monovalent antigen-binding molecules of the invention. Also provided herein are expression vectors comprising said isolated polynucleotides or polynucleotides that are operably linked to regulatory sequences which permit expression of the monovalent antigen-binding molecules. Also provided are host cells or cell-free expression systems containing the expression vectors. Further provided are methods of producing recombinant monovalent antigen-binding molecules, the methods comprising culturing the host cells or cell free expression systems under conditions which permit expression of the monovalent antigen-binding molecule and recovering the expressed monovalent antigen-binding molecule.

In a further aspect, the present invention provides a pharmaceutical composition comprising a monovalent antigen-binding molecule of the invention and at least one pharmaceutically acceptable carrier or excipient.

In still further aspects, the present invention provides a monovalent antigen-binding molecule according to the invention or a pharmaceutical composition of the invention for use as a medicament.

Also provided are methods of treating a disorder in a subject, wherein the method comprises administering to a patient in need thereof a therapeutically effective amount of a monovalent antigen-binding molecule of the invention or a pharmaceutical composition of the invention. The disorder may preferably be an IgA-mediated disorder. In certain embodiments, the disorder is an IgA autoantibody-mediated disorder.

Also provided are methods of reducing serum IgA levels in a subject, wherein the method comprises administering to the subject a monovalent antigen-binding molecule of the invention or a pharmaceutical composition of the invention.

In some embodiments, the serum IgA level in the subject is reduced by at least 90% relative to the baseline serum IgA level.

In some embodiments, the monovalent antigen-binding molecule is administered at a dose of 3 mg/kg, 10 mg/kg, or 30 mg/kg. In some embodiments, multiple doses are administered at a weekly dosing interval, a fortnightly dosing interval or a monthly dosing interval.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 shows the results of a FcRn degradation assay in HEK hFcRn-GFP cells using flow cytometry. Solid bars represent anti-lgA antibody alone; hatched bars represent antiIgA antibodies in complex with human Mota-IgA (1 :1 ratio). An anti-FcRn antibody was added as a positive control for high FcRn degradation.

Figure 2 shows negative stain electron microscopy images of anti-lgA gClone A E HIS (A) and gClone C TA (B) in complex with human Mota-IgA after size exclusion purification (left panels) and after 1 minute of IgA co-incubation on a grid (middle panels). Two representative images of the immune complex shape are also depicted (right panels).

Figure 3 shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item (A) and a log-linear plot of the average human IgA percentage-time profiles in serum upon IP administration of the test item (B). The number of o1 KI mice included per group in the analysis is indicated with (n=x) and data is depicted as mean ± SD.

Figure 4 shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item (A) and a log-linear plot of the average human IgA percentage-time profiles in serum upon IP administration of the test item (B). The number of o1 KI mice included per group in the analysis is indicated with (n=x), and data is depicted as mean ± SD.

Figure 5 shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item (A) and a log-linear plot of the average human IgA percentage-time profiles in serum upon IP administration of the test item (B). The number of o1 KI mice included per group in the analysis is indicated with (n=x), and data is depicted as mean ± SD.

Figure 6 shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item (A) in which the number of Albumus™ mice included per group in the analysis is indicated with (n=x); and a log-linear plot of the average IgA concentration-time profiles in serum upon IP administration of human Mota- IgA on day 0 and day 2 (B) in which the number of Albumus™ mice included per group in the analysis is indicated with (n=x/y, x being the number of mice included after the first human Mota-IgA injection and y being the number of mice included after the second human Mota-IgA injection). Data is depicted as mean ± SEM. The group injected with the isotype control (G6) was only included until D7 of the study.

Figure 7 shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item (A), wherein the number of Albumus™ mice included per group in the analysis is indicated with (n=x) and a log-linear plot of the average IgA concentration-time profiles in serum upon IP administration of human Mota- IgA on day 0 and day 2 (B), wherein the number of Albumus™ mice included per group in the analysis is indicated with (n=x/y, x being the number of mice included after the first human Mota-IgA injection and y being the number of mice included after the second human Mota-IgA injection). Data is depicted as mean ± SEM.

Figure 8 shows the results of a flow cytometry experiment comparing the human IgA internalization properties of anti-lgA LALA ABDEG antibodies in HEK-FcRn and HEK WT cells. Solid bars represent anti-lgA antibody alone; hatched bars represent anti-lgA in complex with human Mota-IgA (1 :1 ratio).

Figure 9 shows the results of MSD competition and displacement experiments comparing the lgA:CD89 blocking properties of anti-lgA gClone A EK HIS and g Clone B HIS H. Blocking the interaction of human and cynomolgus monkey IgA to their corresponding CD89 receptor was tested for both clones.

Figure 10 shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item (A) and a log-linear plot of the average IgA concentration-time profiles in serum upon IP administration of human serum IgA or cynomolgus monkey serum IgA on day 0 and day 2 (B). The number of Albumus™ mice included per group in the analysis is indicated with (n=x) and data is depicted as mean ± SEM. Figure 11 shows the percentage reduction of the average human serum IgA and cynomolgus monkey serum IgA concentration normalized to the concentration of the control group (injected with human or cynomolgus monkey IgA alone) on the corresponding time point of gClone B HIS L(upper graph) and gClone A E HIS (lower graph). The number of mice included per group in the analysis is indicated with (n=x) and data are presented as mean.

Figure 12 shows the PK and PD characteristics of anti-lgA gClone B HIS H-LALA PG ABDEG TA in Albumus™ mice. The mice were injected with either human serum IgA or cynomolgus monkey serum IgA on day 0 and day 2. (A) shows a log-linear plot of the average anti-lgA concentration-time profiles in serum upon IP administration of the test item; wherein the number of Albumus™ mice included per group in the analysis is indicated with (n=x) and data is depicted as mean ± SEM. (B) shows a log-linear plot of the average IgA concentration-time profiles in serum upon IP administration of human serum IgA or cynomolgus monkey serum IgA. The number of mice included per group in the analysis is indicated with n=x/y, being x the number of mice included after the first IgA injection and y the number of mice included after the second IgA injection and the data are presented as mean ± SEM. (C) shows the percentage reduction of the average human serum IgA and cynomolgus monkey serum IgA concentration normalized to the concentration of the control group (injected with human or cynomolgus monkey IgA alone) on the corresponding time point. The number of mice included per group in the analysis is indicated with n=x/y, being x the number of mice included after the first IgA injection and y the number of mice included after the second IgA injection. Data are presented as mean.

Figure 13 shows the FcRn occupancy of anti-lgA gClone A HIS h IgG 1 LALA ABDEG (TA) and gClone B HIS H hlgG 1 LALAPG ABDEG (TA) alone or in complex with IgA (1 :3 ratio) on human PBMCs (A) and cynomolgus monkey PBMCs (B).

Figure 14 shows the results of a FcRn degradation assay in HEK hFcRn-GFP cells using flow cytometry. Solid bars represent anti-lgA antibody alone; hatched bars represent antiIgA antibodies in complex with human Mota-IgA (1 :3 ratio). An anti-FcRn antibody was added as a positive control for high FcRn degradation.

Figure 15 shows the possible O-linked glycan structures that can be attached to the hinge region of Ig A1 and compares a normal and galactose-deficient O-glycan chain (A); as well as the results of a lectin binding ELISA on IgA purified from IgA nephropathy patients (n=12) to evaluate binding to Gd-lgA1 of gClone B HIS H (B). Detection of terminal GalNac residues in the hinge region of lgA1 is done with Lectin helix pomatia agglutinin (Lectin HPA). Commercial rabbit anti-human IgA (left) and Mota hlgG 1 LALAPG ABDEG (right) included as positive and negative controls, respectively.

Figure 16 shows immunofluorescent staining for IgA immune complex deposits in a frozen kidney biopsy from an IgAN patient. Staining was done with gClone B HIS H-mlgG1 FcD, followed by the secondary anti-mouse AF568 antibody. Nuclear counterstaining was done with DAPL Imaging was done with Microscope TissueGnostics TissueFAXS system and slides were imaged at 20x magnification.

Figure 17 shows an SDS-page of secretory IgA (SlgA) and total IgA isolated from IgA nephropathy patients (IgAN) and healthy subjects (HS) under non-reducing conditions. Serum samples were further subdivided based on their lectin binding properties and were either low lectin binding (LLB) or high lectin binding (HLB). IgA samples were detected using rabbit anti-human IgA-HRP (left) or gClone B HIS H-mlgG1 FcD (N297A) (right). Monomeric IgA: 160 kDa, Secretory IgA: 320 kDa, Polymeric IgA/immune complexes: > 350 kDa.

Figure 18 shows a diagrammatic representation of the different Fc engineering strategies implemented on gClone B HIS H to improve IgA sweeping efficacy. Mutations are indicated with stars and listed below the molecules.

Figure 19 shows the results of MSD experiments comparing the lgA:CD89 blocking capacity (competition and displacement) of the gClone B HIS H - LALAPG ABDEG TA and OA variants. Blocking the interaction of human (A, C) and cynomolgus monkey IgA (B, D) to their corresponding CD89 receptor was tested for both variants. (A) hlgA:hCD89 competition activity; (B) clgA:cCD89 competition activity; (C) hlgA:hCD89 displacement activity; (D) clgA:cCD89 displacement activity.

Figure 20 shows the results of the TwoMP Mass Photometer (ReFeyn) analysis to evaluate lgA:anti-lgA complex formation with gClone B HIS H-LALAPG ABDEG TA or OA variants (1 :1 molar ratio). Analysis of the complexes without addition of DSSO (A). Complex analysis after addition of DSSO (serving chemical crosslinking) (B). Figure 21 shows SPR sensorgrams and human FcRn affinity (nM) using the IBIS’MX96 SPR imager system. The antibodies tested were: gClone B HIS H-LALAPG ABDEG twoarmed (TA) or one-armed (OA); and a two-armed (TA) wild-type IgG 1 control.

Figure 22 shows the FcRn occupancy of anti-lgA gClone B HIS H LALAPG ABDEG TA and gClone B HIS H LALAPG ABDEG OA alone or in complex with IgA (1 :3 ratio) on human PBMCs (A) and cynomolgus monkey PBMCs (B).

Figure 23 shows the results of a flow cytometry experiment comparing the human IgA internalization properties of gClone B HIS H-LALAPG ABDEG TA or OA in HEK-FcRn cells (A) or HEK-WT cells (B) using lgA:anti-lgA complexes (1 :1 molar ratio). Background signal is indicated by dotted line (human Mota-IgA alone condition).

Figure 24 shows the results of FcRn degradation assays in HEK-FcRn-GFP cells (A) and U937 cells (B) using flow cytometry. Solid bars depict anti-lgA antibodies alone; hatched bars depict lgA:anti-lgA immune complexes (1 :3 ratio). An anti-FcRn antibody was included as a positive control to induce high FcRn degradation.

Figure 25 shows log-linear plots of the average concentration-time profiles of serum human IgA and serum cynomolgus monkey IgA (A) and test item serum levels (C) upon IP administration of IgA (on DO and D2, 10mg/kg) and test item (on DO, 10 mg/kg for the TA antibody and 6.6 mg/kg for the OA format). Also, the percentage reduction of the average human serum IgA and cynomolgus monkey serum IgA concentration normalized to the concentration of the control group (injected with human or cynomolgus monkey serum IgA alone) on the corresponding time point is visualised (B). The number of mice included per group in the analysis is indicated with n=x (for the PK profile) and n=x/y (for the PD profile, x being the number of mice included after the first IgA injection and y being the number of mice included after the second IgA injection). Data is plotted as mean ± SD.

Figure 26 shows the binding of anti-lgA variants gClone B HIS H LALAPG ABDEG TA (labelled as “LALAPG ABDEG TA”) and gClone B HIS H LALAPG ABDEG OA (labelled as “LALAPG ABDEG OA”) to the IgA BCR of CD19+ CD27+ memory B cells, as assessed by flow cytometry - panel (A). A sample containing full formulation buffer (FFB) only was used as a negative control. Samples comprising HEL-ABDEG (hen egg lysozyme attached to Fc-ABDEG) were used as further negative controls. Panels (B) and (C) show the activation of the IgA BCR by measuring the phosphorylation of downstream signalling proteins PLCy2 and Syk. The same concentrations of LALAPG ABDEG TA and LALAPG ABDEG OA as in panel (A) were tested in panels (B) and (C).

Figure 27 shows SPR sensorgrams reporting the binding of anti-lgA antibody variants to human Fc gamma receptors (A) and cynomolgus Fc gamma receptors (B), as measured using the IBIS’MX96 SPR imager system. The antibodies tested were: gClone B HIS H- LALAPG ABDEG one-armed (OA) (labelled as LALAPG ABDEG OA); gClone B HIS H- ABDEG one-armed (OA) (labelled as ABDEG OA); and a two-armed (TA) wild-type IgG 1 control.

Figure 28 shows the binding of gClone B HIS H-LALAPG ABDEG OA and a control wildtype lgG1 two-armed (TA) antibody to Human C1q (A) and cynomolgus monkey C1q (B) from serum, as assessed by ELISA.

Figure 29 shows an overview of the cynomolgus monkey safety study design.

Figure 30 shows the individual change in percentage cynomolgus monkey serum IgA levels after intra-animal dose escalation of gClone B HIS H-LALAPG ABDEG TA and OA antibodies (A) and after a Single IV dose of 12 mg/kg gClone B HIS H- LALAPG ABDEG OA antibody (B).

Figure 31 shows the individual change in percentage cynomolgus monkey serum IgG levels after intra-animal dose escalation of gClone B HIS H-LALAPG ABDEG TA and OA antibodies (A) and after a single IV dose of 12 mg/kg gClone B HIS H-LALAPG ABDEG OA antibody (B).

Figure 32 shows the individual change in percentage cynomolgus monkey serum IgM levels after intra-animal dose escalation of gClone B HIS H-LALAPG ABDEG TA and OA antibodies (A) and after a Single IV dose of 12 mg/kg gClone B HIS H-LALAPG ABDEG OA antibody (B).

Figure 33 - One-armed format improves IgA sweeping efficacy. pH dependency and FcRn occupancy are crucial for IgA removal. This figure shows the individual change in percentage cynomolgus monkey serum IgA levels after a single IV dose of gClone B HIS H-LALAPG ABDEG OA antibody. Efficient IgA sweeping was observed in all three cynomolgus monkeys (one cynomolgus monkey with high baseline IgA levels, one cynomolgus monkey with medium baseline IgA levels and one cynomolgus monkey with low baseline IgA levels).

Figure 34 - Fab engineering for enhanced IgA sweeping efficacy. Improved IgA clearance in non-human primates, even when high IgA baseline levels. This figure shows the individual change in percentage cynomolgus monkey serum IgA levels after intra-animal dose escalation of gClone B HIS H-LALAPG ABDEG TA antibody. Efficient IgA sweeping was observed in both cynomolgus monkeys (both cynomolgus monkeys had high baseline IgA levels).

Figure 35 - One-armed anti-lgA Ab offers potential for rapid and deep IgA reductions. Predictions for human dosing show fast and deep reductions in IgA levels. This figure shows a prediction for IgA reduction in humans based on four different doses (3 mg/kg weekly, 10 mg/kg weekly, 10 mg/kg fortnightly, 30 mg/kg monthly).

Figure 36 shows the FcRn occupancy of gClone B HIS H hlgG1 LALAPG ABDEG (OA) and gClone B HIS H hlgG 1 LALAPG YPY (OA) alone or in complex with IgA (1 :3 ratio) on human PBMCs.

Figure 37 shows the PK and PD characteristics of test items gClone B HIS H LALAPG ABDEG OA (LALAPG ABDEG OA) and gClone B HIS H LALAPG YPY OA (LALAPG YPY OA) in Albumus™ mice. The mice were injected with I Vlg on day -3 as well as human Mota-IgA on day 0 and day 2.

(A) shows a log-linear plot of the average test item concentration-time profiles in serum upon IV administration of the test item; wherein the number of Albumus™ mice included per group in the analysis is indicated with (n=x) and data is depicted as mean ± SEM.

(B) shows the average IVIg concentration-time profiles in serum upon IP administration of I Vlg ; wherein the number of mice included per group in the analysis is indicated with (n=x) and data is depicted as mean ± SEM.

(C) shows a log-linear plot of the average IgA concentration-time profiles in serum upon IP administration of human Mota-IgA. The number of mice included per group in the analysis is indicated with (n=x/y, being x the number of mice included after the first IgA injection and y the number of mice included after the second IgA injection) and the data are presented as mean ± SEM.

(D) shows the log-linear percentage reduction of the average human Mota-IgA concentration from baseline. The number of mice included per group in the analysis is indicated with (n=x/y, being x the number of mice included after the first IgA injection and y the number of mice included after the second IgA injection). Data are presented as mean ± SEM.

DETAILED DESCRIPTION

A. Definitions

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as is commonly understood by one of ordinary skill in the art in the technical field of the invention.

“Antibody” - As used herein, the term “antibody” is intended to encompass full-length antibodies and variants thereof, including but not limited to bivalent antibodies, humanised antibodies, germlined antibodies (see definitions below). The term “antibody” is typically used herein to refer to immunoglobulin polypeptides having a combination of two heavy and two light chains wherein the polypeptide has significant specific immunoreactive activity to an antigen of interest (herein IgA). For antibodies of the IgG class, the antibodies comprise two identical light polypeptide chains of molecular weight approximately 23,000 Daltons, and two identical heavy chains of molecular weight 53,000- 70,000. The four chains are typically joined by disulfide bonds in a "Y" configuration wherein the light chains bracket the heavy chains starting at the mouth of the "Y" and continuing through the variable region. The light chains of an antibody are classified as either kappa or lambda (K,X). Each heavy chain class may be bound with either a kappa or lambda light chain. In general, the light and heavy chains are covalently bonded to each other, and the "tail" portions of the two heavy chains are bonded to each other by covalent disulfide linkages or non-covalent linkages when the immunoglobulins are generated either by hybridomas, B cells or genetically engineered host cells. In the heavy chain, the amino acid sequences run from an N-terminus at the forked ends of the Y configuration to the C- terminus at the bottom of each chain. An antibody typically comprises two antigen-binding domains that are each capable of binding an antigen. This means that a single antibody can simultaneously bind two antigen molecules and therefore antibodies are typically referred to as being “bivalent” for an antigen. The antigen-binding domains are located at the tips of the forked ends of the Y configuration. The forked ends of a “Y” configuration immunoglobin polypeptide are also referred to in the art as the “arms” of an antibody molecule. Therefore, an antibody can be referred to as having two arms. Each of the two arms typically comprise a single antigenbinding domain (i.e. a Fab fragment).

Those skilled in the art will appreciate that heavy chains are classified as gamma, mu, alpha, delta, or epsilon, (y, p, a, 8, s) with some subclasses among them (e.g., y1-y4). It is the nature of this chain that determines the "class" of the antibody as IgG, IgM, IgA, IgD or IgE, respectively. The immunoglobulin subclasses (isotypes) e.g., lgG1 , lgG2, lgG3, lgG4, lgA1 , etc. are well characterized and are known to confer functional specialization. The term “antibody” as used herein encompasses antibodies from any class or subclass of antibody.

“Monovalent antigen-binding molecule” - As used herein the term “monovalent antigenbinding molecule” (or monovalent antigen binding molecule) refers to an antigen-binding molecule having a single valency for its respective antigen i.e. IgA. The monovalent antigen-binding molecules described herein are capable of binding only a single antigen (i.e. IgA) at any given time. By way of example, a monovalent antigen-binding molecule may be a modified antibody, particularly a modified IgG antibody, with one antigen-binding arm (such as a single Fab region). In other words, the term monovalent antigen-binding molecule encompasses one-armed modified antibodies, particularly one-armed modified IgG antibodies. Figure 18 includes diagrams that show the difference in structure between bivalent antigen-binding molecules (or conventional heterotetrameric two-armed antibodies) and exemplary monovalent antigen-binding molecules of the invention (e.g. one-armed modified antibodies).

“Antigen binding domain” - As used herein “antigen binding domain” refers to any polypeptide domain that binds to an antigen (i.e. IgA herein). The term “antigen binding domain” as used herein is intended to encompass polypeptides derived from antibodies, such as Fab fragments, F(ab')2 fragments, single-chain Fvs (scFv), VH domains (VH), VL domains (VL), VHH domains and antigen binding fragments of the above. The term also encompasses synthetic antigen-binding polypeptides or antibody mimetic polypeptides such as, for example, anticalins and DARPins.

“Variable region” or “variable domain” - The terms "variable region" and "variable domain" are used herein interchangeably and are intended to have equivalent meaning. The term "variable" refers to the fact that certain portions of the variable domains VH and VL differ extensively in sequence among antibodies and are used in the binding and specificity of each particular antibody or monovalent antigen-binding molecule for its target antigen. However, the variability is not evenly distributed throughout the variable domains of antibodies. It is concentrated in three segments called "hypervariable loops" in each of the VL domain and the VH domain which form part of the antigen binding site. The first, second and third hypervariable loops of the VLambda light chain domain are referred to herein as L1 (A), L2(A) and L3(A) and may be defined as comprising residues 24-33 (L1 (A), consisting of 9, 10 or 11 amino acid residues), 49-53 (L2(A), consisting of 3 residues) and 90-96 (L3(A), consisting of 5 residues) in the VL domain (Morea et al., Methods 20:267-279 (2000)). The first, second and third hypervariable loops of the VKappa light chain domain are referred to herein as L1 (K), L2(K) and L3(K) and may be defined as comprising residues 25-33 (L1 (K), consisting of 6, 7, 8, 11 , 12 or 13 residues), 49-53 (L2(K), consisting of 3 residues) and 90-97 (L3(K), consisting of 6 residues) in the VL domain (Morea et al., Methods 20:267-279 (2000)). The first, second and third hypervariable loops of the VH domain are referred to herein as H1 , H2 and H3 and may be defined as comprising residues 25-33 (H1 , consisting of 7, 8 or 9 residues), 52-56 (H2, consisting of 3 or 4 residues) and 91-105 (H3, highly variable in length) in the VH domain (Morea et al., Methods 20:267-279 (2000)).

Unless otherwise indicated, the terms L1 , L2 and L3 respectively refer to the first, second and third hypervariable loops of a VL domain, and encompass hypervariable loops obtained from both Vkappa and Vlambda isotypes. The terms H1 , H2 and H3 respectively refer to the first, second and third hypervariable loops of the VH domain, and encompass hypervariable loops obtained from any of the known heavy chain isotypes, including y, E, 6, a or p.

The hypervariable loops L1 , L2, L3, H1 , H2 and H3 may each comprise part of a "complementarity determining region" or "CDR", as defined below. The terms "hypervariable loop" and "complementarity determining region" are not strictly synonymous, since the hypervariable loops (HVs) are defined on the basis of structure, whereas complementarity determining regions (CDRs) are defined based on sequence variability (Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD., 1983) and the limits of the HVs and the CDRs may be different in some VH and VL domains.

The CDRs of the VL and VH domains can typically be defined as comprising the following amino acids: residues 24-34 (LCDR1), 50-56 (LCDR2) and 89-97 (LCDR3) in the light chain variable domain, and residues 31-35 or 31-35b (HCDR1), 50-65 (HCDR2) and 95- 102 (HCDR3) in the heavy chain variable domain; (Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD. (1991)). Thus, the HVs may be comprised within the corresponding CDRs and references herein to the "hypervariable loops" of VH and VL domains should be interpreted as also encompassing the corresponding CDRs, and vice versa, unless otherwise indicated.

The more highly conserved portions of variable domains are called the framework region (FR), as defined below. The variable domains of native heavy and light chains each comprise four FRs (FR1 , FR2, FR3 and FR4, respectively), largely adopting a p-sheet configuration, connected by the three hypervariable loops. The hypervariable loops in each chain are held together in close proximity by the FRs and, with the hypervariable loops from the other chain, contribute to the formation of the antigen-binding site of antibodies. Structural analysis of antibodies revealed the relationship between the sequence and the shape of the binding site formed by the complementarity determining regions (Chothia et al., J. Mol. Biol. 227: 799-817 (1992)); Tramontane et al., J. Mol. Biol, 215:175-182 (1990)). Despite their high sequence variability, five of the six loops adopt just a small repertoire of main-chain conformations, called “canonical structures”. These conformations are first of all determined by the length of the loops and secondly by the presence of key residues at certain positions in the loops and in the framework regions that determine the conformation through their packing, hydrogen bonding or the ability to assume unusual main-chain conformations.

“CDR” - As used herein, the term "CDR" or "complementarity determining region" means the non-contiguous antigen binding sites found within the variable region of both heavy and light chain polypeptides. These particular regions have been described by Kabat et al., J. Biol. Chem. 252, 6609-6616 (1977) and Kabat et al., Sequences of protein of immunological interest. (1991 ), and by Chothia et al., J. Mol. Biol. 196:901-917 (1987) and by MacCallum et al., J. Mol. Biol. 262:732-745 (1996) where the definitions include overlapping or subsets of amino acid residues when compared against each other. The amino acid residues which encompass the CDRs as defined by each of the above cited references are set forth for comparison. Preferably, the term “CDR” is a CDR as defined by Kabat based on sequence comparisons.

Table 1: CDR definitions

¹Residue numbering follows the nomenclature of Kabat et al., supra ²Residue numbering follows the nomenclature of Chothia et al., supra ³Residue numbering follows the nomenclature of MacCallum et al., supra

“Framework region” - The term “framework region” or “FR region” as used herein, includes the amino acid residues that are part of the variable region, but are not part of the CDRs (e.g., using the Kabat definition of CDRs). Therefore, a variable region framework is between about 100-120 amino acids in length but includes only those amino acids outside of the CDRs. For the specific example of a heavy chain variable domain and for the CDRs as defined by Kabat et al., framework region 1 corresponds to the domain of the variable region encompassing amino acids 1-30; framework region 2 corresponds to the domain of the variable region encompassing amino acids 36-49; framework region 3 corresponds to the domain of the variable region encompassing amino acids 66-94, and framework region 4 corresponds to the domain of the variable region from amino acids 103 to the end of the variable region. The framework regions for the light chain are similarly separated by each of the light chain variable region CDRs. Similarly, using the definition of CDRs by Chothia et al. or McCallum et al. the framework region boundaries are separated by the respective CDR termini as described above. In preferred embodiments the CDRs are as defined by Kabat.

In naturally occurring antibodies, the six CDRs present on each monomeric antibody are short, non-contiguous sequences of amino acids that are specifically positioned to form the antigen binding site (i.e antigen binding domain) as the antibody assumes its three- dimensional configuration in an aqueous environment. The remainder of the heavy and light variable domains show less inter-molecular variability in amino acid sequence and are termed the framework regions. The framework regions largely adopt a p-sheet conformation and the CDRs form loops which connect, and in some cases form part of, the P-sheet structure. Thus, these framework regions act to form a scaffold that provides for positioning the six CDRs in correct orientation by inter-chain, non-covalent interactions. The antigen binding site (i.e. antigen binding domain) formed by the positioned CDRs defines a surface complementary to the epitope on the immunoreactive antigen. This complementary surface promotes the non-covalent binding of the antibody to the immunoreactive antigen epitope. The position of CDRs can be readily identified by one of ordinary skill in the art.

“Constant region” - As used herein, the term “constant region” refers to the portion of an antibody molecule or monovalent antigen-binding molecule outside of the variable domains or variable regions. Immunoglobulin light chains have a single domain “constant region”, typically referred to as the “CL” or “CL1 domain”. This domain lies C terminal to the VL domain. Immunoglobulin heavy chains differ in their constant region depending on the class of immunoglobulin (y, p, a, 8, s). Heavy chains y, a and 8 have a constant region consisting of three immunoglobulin domains (referred to as CH1 , CH2 and CH3) with a flexible hinge region separating the CH1 and CH2 domains. Heavy chains p and E have a constant region consisting of four domains (CH1-CH4). The constant domains of the heavy chain are positioned C terminal to the VH domain.

The numbering of the amino acids in the heavy and light immunoglobulin chains run from the N-terminus at the forked ends of the Y configuration to the C-terminus at the bottom of each chain. Different numbering schemes are used to define the constant domains of the immunoglobulin heavy and light chains. In accordance with the EU numbering scheme, the heavy chain constant domains of an IgG molecule are identified as follows: CH1 - amino acid residues 118-215; CH2 - amino acid residues 231-340; CH3 - amino acid residues 341-446. In accordance with the Kabat numbering scheme, the heavy chain constant domains of an IgG molecule are identified as follows: CH1 - amino acid residues 114-223; CH2 - amino acid residues 244-360; CH3 - amino acid residues 361-477.

“Fc domain” - As used herein, the “Fc domain” defines the portion of the constant region of an immunoglobulin heavy chain including the CH2 and CH3 domains. It typically defines the portion of a single immunoglobulin heavy chain beginning in the hinge region just upstream of the papain cleavage site and ending at the C-terminus of the antibody. The Fc domain typically includes some residues from the hinge region. Accordingly, a complete Fc domain typically comprises at least a portion of a hinge (e.g., upper, middle, and/or lower hinge region) domain, a CH2 domain, and a CH3 domain.

The “hinge region” includes the portion of a heavy chain molecule that joins the CH1 domain to the CH2 domain. This hinge region comprises approximately 25 residues and is flexible, thus allowing the N-terminal antigen binding region(s) to move independently. Hinge regions can be subdivided into three distinct domains: upper, middle, and lower hinge domains (Roux K.H. et a/. J. Immunol. 161 :4083-90 1998). Antigen-binding molecules of the invention comprising a “fully human” hinge region may contain one of the hinge region sequences shown in Table 2 below.

Table 2: Human hinge sequences

“Variant Fc domain” - As used herein, the term "variant Fc domain" refers to an Fc domain with one or more alterations relative to a wild-type Fc domain, for example an Fc domain with one or more alterations relative to the Fc domain of a naturally-occurring or “wild-type” human IgG. Alterations can include amino acid substitutions, additions and/or deletions, linkage of additional moieties, and/or alteration of the native glycans.

“Fc region” - As used herein, the term "Fc region" refers to the portion of a native immunoglobulin formed by the Fc domains of the two heavy chains. A native or wild-type Fc region is typically homodimeric.

“Variant Fc region” - As used herein the term “variant Fc region” refers to an Fc region comprising a first Fc domain and a second Fc domain wherein at least one of the Fc domains has one or more alterations relative to the wild-type domains of a wild-type Fc region. For example, a variant Fc region as described herein may have one or more alterations relative to the Fc region of a naturally-occurring human IgG, particularly human lgG1. The term “variant Fc region” encompasses homodimeric Fc regions wherein each of the constituent Fc domains is the same as well as heterodimeric Fc regions wherein each of the constituent Fc domains is different. For heterodimeric Fc regions, one or both of the Fc domains may be variant Fc domains.

“Specificity”- The monovalent antigen-binding molecules and modified antibodies described herein bind to a particular target antigen - IgA. It is preferred that the monovalent antigen-binding molecules and modified antibodies “specifically bind” to their target antigen, wherein the term “specifically bind” refers to the ability of any monovalent antigen-binding molecule or antibody to preferentially immunoreact with a given target i.e. IgA. The monovalent antigen-binding molecules of the present invention are monospecific and contain one binding site (i.e. antigen binding domain), which specifically binds a particular target antigen (i.e. IgA).

“Modified antibody” - As used herein, the term “modified antibody” includes synthetic forms of antibodies which are altered such that they are not naturally occurring. Examples include but are not limited to antibodies that comprise at least two heavy chain portions but not two complete heavy chains (such as, domain deleted antibodies or minibodies); heavy chain molecules joined to scFv molecules and the like. scFv molecules are known in the art and are described, e.g., in US patent 5,892,019. The term “modified antibody”, particularly as used in the context of the present invention, refers to antibodies that are monovalent, particularly one-armed modified antibodies having a single antigen-binding arm. Modified antibodies in accordance with the present invention may comprise any suitable antigen-binding domain as defined elsewhere herein linked to a variant Fc domain as defined elsewhere herein.

“Humanising substitutions” - As used herein, the term “humanising substitutions” refers to amino acid substitutions in which the amino acid residue present at a particular position in the VH or VL domain of an antibody or monovalent antigen-binding molecule is replaced with an amino acid residue which occurs at an equivalent position in a reference human VH or VL domain. The reference human VH or VL domain may be a VH or VL domain encoded by the human germline. Humanising substitutions may be made in the framework regions and/or the CDRs of the antibodies, defined herein.

“Humanised variants” - As used herein the term “humanised variant” or “humanised antibody” refers to a variant antibody or monovalent antigen-binding molecule which contains one or more “humanising substitutions” compared to a reference antibody sequence, wherein a portion of the reference antibody (e.g. the VH domain and/or the VL domain or parts thereof containing at least one CDR) has an amino acid derived from a non-human species, and the “humanising substitutions” occur within the amino acid sequence derived from a non-human species.

“Germlined variants” - The term “germlined variant” or “germlined antibody” is used herein to refer specifically to “humanised variants” in which the “humanising substitutions” result in replacement of one or more amino acid residues present at (a) particular position(s) in the VH or VL domain of an antibody with an amino acid residue which occurs at an equivalent position in a reference human VH or VL domain encoded by the human germline. It is typical that for any given “germlined variant”, the replacement amino acid residues substituted into the germlined variant are taken exclusively, or predominantly, from a single human germline-encoded VH or VL domain. The terms “humanised variant” and “germlined variant” are often used interchangeably. Introduction of one or more “humanising substitutions” into a camelid-derived (e.g. llama derived) VH or VL domain results in production of a “humanised variant” of the camelid (llama)-derived VH or VL domain. If the amino acid residues substituted in are derived predominantly or exclusively from a single human germline-encoded VH or VL domain sequence, then the result may be a “human germlined variant” of the camelid (llama)-derived VH or VL domain. “Affinity variants” - As used herein, the term “affinity variant” refers to a variant monovalent antigen-binding molecule or variant antibody which exhibits one or more changes in amino acid sequence compared to a reference monovalent antigen binding molecule/antibody, wherein the affinity variant exhibits an altered affinity for the target antigen in comparison to the reference. For example, affinity variants will exhibit a changed affinity for a target, for example IgA, as compared to a reference monovalent IgA- binding molecule or reference IgA antibody. Preferably, the affinity variant will exhibit improved affinity for the target antigen, as compared to the reference. Affinity variants typically exhibit one or more changes in amino acid sequence in the CDRs, as compared to the reference monovalent antigen binding molecule/antibody. Such substitutions may result in replacement of the original amino acid present at a given position in the CDRs with a different amino acid residue, which may be a naturally occurring amino acid residue or a non-naturally occurring amino acid residue. The amino acid substitutions may be conservative or non-conservative.

“Engineered” - As used herein the term “engineered” includes manipulation of nucleic acid or polypeptide molecules by synthetic means (e.g. by recombinant techniques, in vitro peptide synthesis, by enzymatic or chemical coupling of peptides or some combination of these techniques). Preferably, the monovalent antigen-binding molecules of the invention are engineered, including for example, humanized variants which have been engineered to improve one or more properties, such as antigen binding, stability/half-life or effector function.

“FcRn” - As used herein, the term “FcRn” refers to a neonatal Fc receptor. Exemplary FcRn molecules include human FcRn encoded by the FCAR gene as set forth in RefSeq NM_002000.

“CD89” - As used herein, the term “CD89” refers to a FcoRI Fc receptor that binds to the constant region of IgA. The receptor exhibits a relatively low affinity of ~10⁶ M“¹ for monomeric and dimeric forms of IgA. However, it is able to bind IgA immune complexes with high avidity that results in cross-linking. Cross-linking of FcoRI by IgA immune complexes (or IgA-opsonized pathogens) induces a variety of processes, including phagocytosis, antibody-dependent cellular cytotoxicity, superoxide generation, release of inflammatory mediators, and cytokines as well as antigen presentation. Exemplary CD89 molecules include human CD89 as set forth in RefSeq: NM_000569, NM_133269, N M_133271 , NM_133272 and NM_133273.

“N-linked glycan” - As used herein the term “N-linked glycan” refers to the N-linked glycan attached to the nitrogen (N) in the side chain of asparagine in the sequence (i.e. , Asn-X-Ser or Asn-X-Thr sequence, where X is any amino acid except proline) present in the CH2 domain of an Fc region. Such N-glycans are fully described in, for example, Drickamer K and Taylor ME (2006) Introduction to Glycobiology, 2^nd ed., incorporated herein by reference in its entirety.

“Afucosylated” - As used herein the term “afucosylated” refers to an N-linked glycan which lacks a core fucose molecule as described in US Pat No. 8067232, incorporated herein by reference in its entirety.

“Bisecting GIcNAc” - As used herein the term “bisecting GIcNAc” refers to an N-linked glycan having an N-acetylglucosamine (GIcNAc) molecule linked to a core mannose molecule, as described in US Pat. No. 8021856, incorporated herein by reference in its entirety.

“IgA” - As used herein, the term “IgA” refers to “immunoglobulin A” molecules or “class A immunoglobulins”. IgA is the most abundant immunoglobin class at mucosal surfaces and the second most prevalent class in human serum. At mucosal surfaces, dimeric forms of IgA predominate (such as secreted IgA (slgA)) whereas in human serum the monomeric form is the most prevalent form of IgA. There are two known isotypes of IgA - Ig A1 and lgA2. The two isotypes are distinguished from one another by the size of their hinge regions and the number of glycosylation sites. lgA1 contains a 13 amino acid hinge region with many O-linked glycolsyation sites whereas lgA2 does not contain this region and also has two additional N-linked carbohydrate chains. The distribution of these isotypes differs in the mucosal areas and serum. Serum IgA is mostly comprised of lgA1 (around 90%). In contrast, mucosal IgA consist of both isotypes and the ratio of the isotypes differs according to the specific location of the mucosal area (Cerutti 2008; Breedveld and van Egmond 2019; de Sousa-Pereira and Woof 2019). As explained in greater detail elsewhere, serum IgA, IgA autoantibodies and IgA-immune complexes have been implicated in various autoimmune disorders. “Antibody-mediated disorder” - As used herein, the term “antibody-mediated disorder” refers to any disease or disorder caused or exacerbated by the presence of an antibody in a subject. An “IgA-mediated disorder” refers to a disease or disorder caused or exacerbated by the presence of IgA antibodies, including for example IgA-immune complexes.

“Treat, treating and treatment” - As used herein, the terms "treat," "treating," and "treatment" refer to therapeutic or preventative measures described herein. The methods of "treatment" employ administration to a subject, for example, a subject having an antibody- mediated disease or disorder (e.g. autoimmune disease) or predisposed to having such a disease or disorder, an antigen-binding molecule in accordance with the present invention, in order to prevent, cure, delay, reduce the severity of, or ameliorate one or more symptoms of the disease or disorder or recurring disease or disorder, or in order to prolong the survival of a subject beyond that expected in the absence of such treatment.

“Subject” - As used herein, the term “subject” refers to any human or non-human animal. In certain embodiments, the term “subject” refers to any human or non-human mammal. In preferred embodiments, the subject is a human. In certain embodiments the subject is an adult human. As used herein, an “adult human” is a human who is at least 18 years of age.

B. Monovalent antigen-binding molecules

- a variant Fc region, wherein the variant Fc region comprises a first Fc domain and a second Fc domain; and wherein the first Fc domain and the second Fc domain each comprise the amino acids Y, T, E, K, F and Y at EU positions 252, 254, 256, 433, 434 and 436, respectively. In a second aspect, the present invention provides a monovalent antigen-binding molecule comprising:

The monovalent antigen-binding molecules of the invention are characterized in that they possess a single antigen valency. More specifically, the monovalent antigen-binding molecules have a single antigen-binding domain wherein the antigen-binding domain binds to IgA. The antigen-binding domain comprises a VH domain and a VL domain having a particular set of six CDR sequences as represented by SEQ ID NOs: 1-6. This variable region confers particular functional properties, including pH-dependent IgA binding, as described herein below. The monovalent antigen-binding molecules described herein are also referred to as monovalent IgA-binding molecules.

The monovalent antigen-binding molecules of the invention are further characterized in that they possess a variant Fc region comprising the ABDEG™ amino acid signature. More specifically, the variant Fc region comprises a first Fc domain and a second Fc domain wherein both Fc domains comprise the amino acids Y, T, E, K, F and Y at EU positions 252, 254, 256, 433, 434 and 436, respectively. As described in more detail elsewhere herein, variant Fc regions possessing the ABDEG™ amino acid signature exhibit enhanced binding to the human neonatal Fc receptor FcRn. The combination of the single antigenbinding domain characterized by the CDR sequences as represented by SEQ ID NOs: 1-6 and the variant Fc region comprising the ABDEG™ amino acid signature results in an antigen-binding molecule that exhibits effective IgA sweeping activity in vivo.

As an alternative to the variant Fc region comprising the ABDEG™ amino acid signature, the monovalent antigen-binding molecules of the invention may be further characterized in that they possess a variant Fc region comprising “YPY” mutations. More specifically, the variant Fc region comprises a first Fc domain and a second Fc domain wherein both Fc domains comprise the amino acids Y, P and Y at EU positions 252, 308 and 434, respectively. As described in more detail elsewhere herein, variant Fc regions possessing the YPY mutations exhibit enhanced binding to the human neonatal Fc receptor FcRn. The combination of the single antigen-binding domain characterized by the CDR sequences as represented by SEQ ID NOs: 1-6 and the variant Fc region comprising the YPY mutations results in an antigen-binding molecule that exhibits effective IgA sweeping activity in vivo.

Embodiments pertaining to structural and functional features of the monovalent antigenbinding molecules of the invention are set forth below.

(I) Monovalent IgA-binding molecules

The monovalent antigen-binding molecules of the invention are also referred to herein as “monovalent IgA binding molecules” and “monovalent IgA-binding molecules” since the single antigen they bind is IgA. As described elsewhere herein, monovalent IgA-binding molecules of the invention are advantageous in that they exhibit improved IgA sweeping as compared with the corresponding bivalent IgA-binding molecules.

The monovalent IgA-binding molecules of the present invention may adopt the format of any suitable antigen-binding molecule displaying immunoreactivity for IgA, provided that they comprise:

(i) an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) comprising the CDR sequences:

- HCDR3 comprising or consisting of SEQ ID NO: 3;

- HCDR2 comprising or consisting of SEQ ID NO: 2;

- HCDR1 comprising or consisting of SEQ ID NO: 1 ;

- LCDR3 comprising or consisting of SEQ ID NO: 6;

- LCDR2 comprising or consisting of SEQ ID NO: 5; and

- LCDR1 comprising or consisting of SEQ ID NO: 4; and

(ii) a variant Fc region, wherein the variant Fc region comprises a first Fc domain and a second Fc domain; and wherein the first Fc domain and the second Fc domain each comprise the amino acids Y, T, E, K, F and Y at EU positions 252, 254, 256, 433, 434 and 436, respectively; or (i) an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) comprising the CDR sequences:

- HCDR3 comprising or consisting of SEQ ID NO: 3;

- HCDR2 comprising or consisting of SEQ ID NO: 2;

- HCDR1 comprising or consisting of SEQ ID NO: 1 ;

- LCDR3 comprising or consisting of SEQ ID NO: 6;

- LCDR2 comprising or consisting of SEQ ID NO: 5; and

- LCDR1 comprising or consisting of SEQ ID NO: 4; and

(ii) a variant Fc region, wherein the variant Fc region comprises a first Fc domain and a second Fc domain; and wherein the first Fc domain and the second Fc domain each comprise the amino acids Y, P and Y at EU positions 252, 308 and 434, respectively.

The antigen-binding domain may be any domain that exhibits binding specificity for IgA. In certain embodiments, said antigen-binding domain comprises or consists of a Fab, a VH- VL domain pairing, a scFv fragment, a disulfide-linked Fv (sdFv) or a single-chain Fv (scFv). In certain embodiments, said antigen-binding domain comprises or consists of a Fab or a scFv. In preferred embodiments, the antigen-binding domain is a Fab.

The monovalent IgA-binding molecules as described herein preferably comprise a single Fab (i.e. one Fab arm).

The monovalent IgA-binding molecules of the present invention encompass modified forms of conventional heterotetrameric antibodies i.e. anti-lgA antibodies modified to be monovalent such that they are capable of binding to only one IgA antigen at any given time. Such modified anti-lgA antibodies are also referred to herein as monovalent anti-lgA antibodies or monovalent IgA antibodies.

Monovalent antigen-binding molecules, including monovalent anti-lgA antibodies, of the invention are intended for human therapeutic use. It follows that monovalent IgA-binding molecules or monovalent anti-lgA antibodies of the invention will typically be modified forms of human IgA, IgD, IgE, IgG or IgM antibodies, preferably IgG antibodies in which case they can belong to any of the four sub-classes IgG 1 , lgG2a and b, lgG3 or lgG4. In preferred embodiments, the monovalent anti-lgA antibodies of the invention are modified IgG antibodies, optionally of the IgG 1 sub-class. In further preferred embodiments, the monovalent anti-lgA antibodies of the invention are modified IgG antibodies having only one Fab arm. Most preferably, the monovalent anti-lgA antibodies of the invention are modified lgG1 antibodies having only one Fab arm.

The monovalent anti-lgA antibodies of the invention preferably derive from monoclonal antibodies since monoclonal antibodies are highly specific, being directed against a single antigenic site.

The monovalent anti-lgA antibodies described herein may exhibit high human homology. Such monovalent anti-lgA antibodies may comprise VH and VL domains of native nonhuman antibodies which exhibit sufficiently high % sequence identity to human germline sequences. In certain embodiments, the monovalent anti-lgA antibodies are humanised or germlined variants of non-human antibodies.

In some embodiments, the monovalent IgA-binding molecules can bind to: monomeric IgA and/or dimeric IgA and/or multimeric IgA. In some embodiments, the monovalent IgA- binding molecules can bind to: monomeric IgA and dimeric IgA and multimeric IgA. In some embodiments, the monovalent IgA-binding molecules can bind to secretory IgA (slgA). In preferred embodiments, the IgA that is bound is human IgA. In most preferred embodiments, the IgA that is bound is human serum IgA.

The monovalent antigen-binding molecules of the invention exhibit pH-dependent antigen binding i.e., pH-dependent binding to IgA. As used herein, “pH-dependent binding” means that the antigen-binding molecules exhibit lower IgA binding affinity at an acidic pH than at a neutral pH. pH-dependent IgA binding is advantageous since it contributes to the IgA sweeping from serum.

In more detail, once a monovalent antigen-binding molecule is bound to IgA and internalised by FcRn, the complex enters the endosomal compartment. Whilst serum and cellular pH is typically neutral, the pH of the endosomal compartments is slightly acidic. Monovalent antigen-binding molecules that are able to dissociate from IgA in the early endosome (i.e. at an acidic pH) can be recycled back to the cell surface. In contrast, antigen-binding molecules that bind with high affinity to IgA in the endosomal compartments would typically be trafficked with IgA to the lysosomes for degradation. pH-dependent monovalent antigen-binding molecules in accordance with the present invention can eliminate serum IgA (including IgA immune complexes) by binding and internalising IgA. Once internalised, the lower IgA binding affinity in the acidic endosomal compartment will facilitate release of the IgA by the monovalent antigen-binding molecules such that the IgA is trafficked to the lysosomes for degradation. The free monovalent antigen-binding molecules can be recycled to the cell surface such that they can mediate binding, internalisation and degradation of further IgA molecules. In this way, whilst a monovalent antigen-binding molecule of the invention is capable of binding only a single IgA molecule at any given time, the molecule is capable of binding a plurality of IgA molecules at different points in time via the recycling method described above.

For the monovalent antigen-binding molecules described herein, the IgA binding affinity is lower at endosomal pH as compared to the IgA binding affinity at serum pH. The endosomal pH is typically acidic pH whereas the serum pH is typically neutral pH. Accordingly, the monovalent antigen-binding molecules exhibit pH-dependent IgA binding such that their IgA binding activity is lower at acidic pH as compared to the IgA binding activity at neutral pH. Endosomal pH or “acidic pH” may be pH of from about pH 4.0 to about pH 6.5, preferably from about pH 5.5 to about pH 6.5, preferably from about pH 5.5 to about pH 6.0, preferably pH 5.5, pH 5.6, pH 5.7 or pH 5.8. Serum pH or “neutral pH” may be pH of from about pH 6.9 to about pH 8.0, preferably from about pH 7.0 to about pH 8.0, preferably from about pH 7.0 to about pH 7.4, preferably pH 7.0 or pH 7.4.

The monovalent antigen-binding molecules exhibit pH-dependent binding such that the human IgA-binding affinity at pH 6 is lower as compared with the human IgA binding affinity at pH 7.4. In preferred embodiments, the monovalent antigen-binding molecules exhibit pH-dependent binding such that the human IgA-binding affinity at pH 6 is reduced by at least 75% as compared with the human IgA binding affinity at pH 7.4.

The monovalent antigen-binding molecules also exhibit pH-dependent binding such that the human IgA-binding affinity at pH 5 is lower as compared with the human IgA binding affinity at pH 7.4. In preferred embodiments, the monovalent antigen-binding molecules exhibit pH-dependent binding such that the human IgA-binding affinity at pH 5 is reduced by 100% as compared with the human IgA binding affinity at pH 7.4. In some embodiments, the monovalent antigen-binding molecules are capable of binding to IgA at a neutral pH and are not capable of binding to IgA at an acidic pH. In said embodiments, acidic pH is about pH 5.0, about pH 5.5 or about pH 6.0. In said embodiments, neutral pH is about pH 7.4.

The monovalent IgA-binding molecules described herein can bind to IgA so as to inhibit the binding of IgA to an IgA receptor. The monovalent IgA-binding molecules of the invention can prevent IgA binding to IgA receptors such as FcoRI (CD89)

The monovalent IgA-binding molecules can target IgA autoantibodies and/or autoantibodies complexed with self-antigens (i.e. immune complexes). Such autoantibodies typically bind to activating Fc receptors, causing numerous autoimmune diseases (which occur in part because of immunologically mediated inflammation against self-tissues) (see e.g., Clarkson et al., NEJM 314(9), 1236-1239 (2013);

US20040010124A1 ; US20040047862A1 ; and US2004/0265321A1 , incorporated herein by reference in their entirety). The monovalent IgA-binding molecules can reduce IgA immune complexes. An IgA immune complex includes by way of non-limiting example, IgA-IgG complexes, IgA-antigen complexes, IgA-pathogen complexes and IgA-FcoRI (CD89) complexes.

The monovalent antigen-binding molecules of the invention are also capable of displacing IgA that is already associated (i.e. bound) with IgA receptors, such as FcoRI (CD89). The monovalent IgA-binding molecules of the invention can prevent binding of IgA to FcoRI (CD89) receptor and also displace IgA bound to FcoRI (CD89) receptor.

(II) Variant Fc regions incorporating ABDEG™ technology

The present invention provides monovalent antigen-binding molecules comprising variant Fc regions incorporating ABDEG™ technology. As reported in Vaccaro et al. (Nat. Biotechnology (2005) 23(10):1283-8), ABDEG™ antibodies (meaning “antibodies that enhance IgG degradation”) comprise an engineered or variant Fc region. This engineered or variant Fc region can bind to the neonatal Fc receptor, FcRn, with higher affinity and reduced pH dependence as compared with the Fc region of wild-type antibodies. ABDEG™ antibodies and FcRn antagonists incorporating ABDEG™ technology have been described for the treatment of antibody-mediated diseases such as autoimmune diseases (see W02006/130834 and WO2015/100299, incorporated herein by reference).

The Fc domain amino acid “signature” of ABDEG™ is well-characterised. The present invention provides monovalent antigen-binding molecules comprising a variant Fc region, wherein the first Fc domain and the second Fc domain each comprise the amino acids Y, T, E, K, F and Y at EU positions 252, 254, 256, 433, 434 and 436, respectively. The amino acid Y is the native amino acid at EU position 436 in wild-type human lgG1 Fc domain. This Fc domain amino acid signature is the ABDEG™ signature.

In certain embodiments, the first and second Fc domains of the variant Fc region are identical to their corresponding wild-type Fc domains but for the amino acids Y, T, E, K and F at EU positions 252, 254, 256, 433 and 434, respectively.

As indicated above, the positions are defined in accordance with EU numbering. EU numbering refers to the convention for the Fc region described in Edelman, G.M. et al., Proc. Natl. Acad. Sei. USA, 63: 78-85 (1969); and Kabat et al., in "Sequences of Proteins of Immunological Interest", U.S. Dept. Health and Human Services, 5th edition, 1991.

As described above, the variant Fc region of ABDEG™ monovalent antigen-binding molecules exhibits increased binding affinity for the Fc receptor FcRn, particularly human FcRn. The ABDEG™ variant Fc region binds to FcRn with increased affinity relative to a wild-type Fc region. In such embodiments, the wild-type Fc region may be the wild-type Fc region from which the variant Fc region derives. For example, if the ABDEG™ variant Fc region is derived from a human IgG 1 Fc region, the variant Fc domain may bind to FcRn with higher affinity than the human IgG 1 Fc domain.

In certain embodiments, the ABDEG™ variant Fc region binds to FcRn, preferably human FcRn, with increased affinity relative to a wild-type IgG Fc region, preferably a wild-type human IgG Fc region. In a preferred embodiment, the ABDEG™ variant Fc region binds to FcRn, preferably human FcRn, with increased affinity relative to a wild-type human IgG 1 Fc region or a wild-type human lgG3 Fc region.

The ABDEG™ variant Fc regions of the monovalent antigen-binding molecules described herein may be variant Fc regions derived from any suitable wild-type immunoglobulin Fc region. In certain embodiments, the ABDEG™ variant Fc region is a variant IgG Fc region. The wild-type IgG region may be an IgG of any sub-class including IgG 1 , lgG2, lgG3 and lgG4. The wild-type IgG region is preferably lgG1 , preferably human lgG1.

In preferred embodiments, the ABDEG™ variant Fc region is a variant IgG 1 Fc region. In such embodiments, the variant ABDEG™ Fc region has the amino acid sequence of a wildtype lgG1 Fc region, preferably a human lgG1 Fc region, having first and second Fc domains each comprising or consisting of the ABDEG™ amino acid signature described herein, specifically amino acids Y, T, E, K, F and Y at EU positions 252, 254, 256, 433, 434 and 436, respectively.

Non-limiting examples of variant first and second Fc domains for inclusion in the monovalent antigen-binding molecules described herein are set forth in Table 3A below.

In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 78. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 79. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 80. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 82. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 83. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 84. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 86. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 87. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 88.

In certain embodiments, the first or second Fc domain is linked to a heavy chain CH1 domain and the heavy chain constant region comprises or consists of the amino acid sequence set forth in SEQ ID NO: 81 . In certain embodiments, the first or second Fc domain is linked to a heavy chain CH1 domain and the heavy chain constant region comprises or consists of the amino acid sequence set forth in SEQ ID NO: 85. In certain embodiments, the first or second Fc domain is linked to a heavy chain CH1 domain and the heavy chain constant region comprises or consists of the amino acid sequence set forth in SEQ ID NO: 89.

In certain embodiments, the variant Fc region comprises: a first Fc domain that comprises or consists of the amino acid sequence set forth in any one of SEQ ID NOs: 90, 91 , 92, 98, 99, 100, 106, 107 and 108; and a second Fc domain that comprises or consists of the amino acid sequence set forth in any one of SEQ ID NOs: 94, 95, 96, 102, 103, 104, 110, 111 and 112.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 90; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 94.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 90; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 95.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 90; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 96.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 91 ; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 94.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in

SEQ ID NO: 91 ; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 95.

SEQ ID NO: 91 ; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 96.

SEQ ID NO: 92; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 94.

SEQ ID NO: 92; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 95.

SEQ ID NO: 92; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 96.

SEQ ID NO: 98; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 102. In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in

SEQ ID NO: 98; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 103.

SEQ ID NO: 98; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 104.

SEQ ID NO: 99; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 102.

SEQ ID NO: 99; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 103.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 99; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 104.

SEQ ID NO: 100; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 102. In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in

SEQ ID NO: 100; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 103.

SEQ ID NO: 100; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 104.

SEQ ID NO: 106; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 110.

SEQ ID NO: 106; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 111.

SEQ ID NO: 106; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 112.

SEQ ID NO: 107; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 110. In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in

SEQ ID NO: 107; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 111.

SEQ ID NO: 107; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 112.

SEQ ID NO: 108; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 110.

SEQ ID NO: 108; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 111.

SEQ ID NO: 108; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 112. Table 3A. Amino acid sequences of non-limiting examples of Fc domains and heavy chain constant regions incorporating Fc domains

The ABDEG™ variant Fc regions of the monovalent antigen-binding molecules of the present invention preferably comprise further alternations. Specifically, they may comprise amino acid substitutions in addition to the ABDEG™ signature. These alternations, particularly amino acid substitutions, may further improve the properties of the monovalent antigen-binding molecules. In some embodiments, the variant Fc region further comprises at least one additional amino acid substitution as compared with the corresponding wild-type Fc region in addition to the ABDEG™ signature; wherein the at least one additional substitution reduces or eliminates Fc effector function.

For such embodiments wherein the variant Fc region comprises one or more amino acid substitutions in addition to the ABDEG™ signature, the first Fc domain and second Fc domain may comprise the amino acids A, A at EU positions 234 and 235, respectively. In some embodiments, the first and second Fc domains each comprise, in addition to the ABDEG™ signature, a combination of amino acid substitutions: L234A and L235A, wherein the positions are defined in accordance with EU numbering. In further embodiments wherein the variant Fc region comprises one or more amino acid substitutions in addition to the ABDEG™ signature, the first and second Fc domains of the variant Fc region are identical to the corresponding wild-type Fc domains, for example human IgG 1 Fc domains, but for the amino acids A, A, Y, T, E, K and F at EU positions 234, 235, 252, 254, 256, 433 and 434, respectively. The so-called “LALA” mutations are known to reduce Fc binding to Fey Receptors.

In further embodiments, wherein the variant Fc region comprises one or more amino acid substitutions in addition to the ABDEG™ signature, the first Fc domain and second Fc domain may comprise the amino acids A, A and G at EU positions 234, 235 and 329, respectively. In some embodiments, the first and second Fc domains each comprise, in addition to the ABDEG™ signature, a combination of amino acid substitutions: L234A, L235A and P329G, wherein the positions are defined in accordance with EU numbering. In additional embodiments wherein the variant Fc region comprises one or more amino acid substitutions in addition to the ABDEG™ signature, the first and second Fc domains of the variant Fc region are identical to the corresponding wild-type Fc domains, for example human IgG 1 Fc domains, but for the amino acids A, A, G, Y, T, E, K and F at EU positions 234, 235, 329, 252, 254, 256, 433 and 434, respectively. The so-called “LALAPG” mutations have been reported to almost completely block Fc binding to Fey Receptors (Schlothauer et al., 2016).

As noted above, the monovalent antigen-binding molecules of the invention comprise a single antigen-binding domain. For embodiments in which the antigen-binding domain is a Fab, the heavy chain of the Fab is covalently attached to only one Fc domain (or variant Fc domain) of the variant Fc region. In other words, the molecule may be considered heterodimeric since one Fc domain (or variant Fc domain) is covalently attached to the heavy chain of a Fab and one Fc domain (or variant Fc domain) is not attached to a Fab (nor any other antigen binding domain).

As such, in order to produce the monovalent antigen-binding molecules of the invention, heterodimerisation between a first Fc domain (or variant Fc domain) and a second Fc domain (or variant Fc domain) of the Fc region is required. In other words, heterodimerisation between a first Fc domain (or variant Fc domain) that is attached to the heavy chain of a Fab with a second Fc domain (or variant Fc domain) that is not attached to any antigen binding domain is required to arrive at the monovalent antigen-binding molecules of the invention.

In preferred embodiments therefore, it is advantageous to promote heterodimerisation. Accordingly, the variant Fc region may comprise at least one alteration, particularly at least one amino acid substitution, that promotes dimerisation between a first Fc domain (or variant Fc domain) and a second Fc domain (or variant Fc domain). More specifically, the first and second Fc domains may further comprise alternations in addition to the ABDEG™ signature, for example one or more amino acid substitutions, to promote heterodimerisation.

Examples of amino acid substitutions that promote heterodimerisation are known in the literature and any suitable means to promote heterodimerisation may be adopted in the variant Fc region of the monovalent antigen-binding molecules of the present invention.

In some embodiments, the variant Fc region comprises “knob-into-hole” substitutions, which are known in the literature (see for example WO 2006/028936 which is incorporated herein by reference). This technology promotes heterodimerisation by introducing different but complementary amino acid substitutions into the first Fc domain and the second Fc domain.

As such, in a further preferred embodiment, the first and second Fc domains each comprise a set of the same amino acid substitutions such as the ABDEG™ signature and “LALAPG” (described above) as well as a set of different amino acid substitutions (such as knob-into-hole substitutions). In some embodiments, the first Fc domain comprises amino acids S, A, and V at EU positions 366, 368, and 407, respectively in addition to the ABDEG™ signature and the second Fc domain comprises amino acid W at EU position 366 in addition to the ABDEG™ signature. In preferred embodiments, the first Fc domain comprises amino acid W at EU position 366 in addition to the ABDEG™ signature and the second Fc domain comprises amino acids S, A, and V at EU positions 366, 368, and 407, respectively in addition to the ABDEG™ signature.

In certain embodiments, the first Fc domain comprises an amino acid substitution T366W in addition to the ABDEG™ signature, and the second Fc domain comprises amino acid substitutions T366S, L368A, and Y407V in addition to the ABDEG™ signature, wherein the positions are defined in accordance with EU numbering.

As noted above in preferred embodiments, the variant Fc region comprises two different variant Fc domains that form a heterodimer. By way of example, in some embodiments, the first variant Fc domain comprises the ABDEG™ signature, L234A, L235A mutations and “knob” mutations; and the second variant Fc domain also comprises the ABDEG™ signature and L234A, L235A mutations but comprises “hole” mutations instead of the “knob” mutations. In further embodiments, the second variant Fc domain comprises the ABDEG™ signature, L234A, L235A mutations and “knob” mutations; and the first variant Fc domain also comprises the ABDEG™ signature and L234A, L235A mutations but comprises “hole” mutations instead of the “knob” mutations.

In preferred embodiments, the first variant Fc domain comprises the ABDEG™ signature, L234A, L235A, P329G mutations and “knob” mutations; and the second variant Fc domain also comprises the ABDEG™ signature and L234A, L235A, P329G mutations but comprises “hole” mutations instead of the “knob” mutations. In further embodiments, the second variant Fc domain comprises the ABDEG™ signature, L234A, L235A, P329G mutations and “knob” mutations; and the first variant Fc domain also comprises the ABDEG™ signature and L234A, L235A, P329G mutations but comprises “hole” mutations instead of the “knob” mutations.

In further preferred embodiments, the first Fc domain of the variant Fc region is identical to the corresponding wild-type Fc domain, for example a human IgG 1 Fc domain, but for the amino acids A, A, G, Y, T, E, W, K and F at EU positions 234, 235, 329, 252, 254, 256, 366, 433 and 434, respectively; and the second Fc domain of the variant Fc region is identical to the corresponding wild-type Fc domain, for example a human lgG1 Fc domain, but for the amino acids A, A, G, Y, T, E, K, F, S, A and V at EU positions 234, 235, 329, 252, 254, 256, 433, 434, 366, 368 and 407, respectively. The corresponding wild-type Fc domains are preferably the Fc domains of the human IgG 1 Fc region.

In certain embodiments, the monovalent antigen-binding molecule comprises a variant Fc domain or variant Fc region comprising an N-linked glycan (e.g., at EU position 297). In this case it is possible to increase the binding affinity of the monovalent antigen-binding molecule for C89 by altering the glycan structure. Alterations of the N-linked glycan of Fc regions are well known in the art. For example, afucosylated N-linked glycans or N-glycans having a bisecting GIcNac structure have been shown to exhibit increased affinity for CD89. Accordingly, in certain embodiments, the N-linked glycan is afucosylated. Afucosylation can be achieved using any art recognized means. For example, a monovalent antigen-binding molecule can be expressed in cells lacking fucosyl transferase, such that fucose is not added to the N-linked glycan at EU position 297 of the variant Fc domain or variant Fc region (see e.g., US 8,067,232, the contents of which is incorporated by reference herein in its entirety). In certain embodiments, the N-linked glycan has a bisecting GIcNac structure. The bisecting GIcNac structure can be achieved using any art recognized means. For example, a monovalent antigen-binding molecule can be expressed in cells expressing beta1-4-N-acetylglucosaminyltransferase III (GnTIII), such that bisecting GIcNac is added to the N-linked glycan at EU position 297 of the variant Fc domain or variant Fc region (see e.g., US 8021856, the contents of which is incorporated by reference herein in its entirety). Additionally or alternatively, alterations of the N-linked glycan structure can also be achieved by enzymatic means in vitro.

To enhance the manufacturability of the monovalent antigen-binding molecules of the present invention, it is preferable that the variant Fc domains or variant Fc regions do not comprise any non-disulphide bonded cysteine residues. Accordingly, in certain embodiments the variant Fc domains or variant Fc regions do not comprise a free cysteine residue.

(Hi) Variant Fc regions incorporating YPY mutations

The present invention provides monovalent antigen-binding molecules comprising variant Fc regions incorporating YPY mutations. As reported in Yang et al., 2017 (mAbs, 9(7), 1105-1117), YPY mutations increase binding to the human neonatal Fc receptor FcRn. The YPY mutations result in amino acids Y, P and Y at EU positions 252, 308 and 434, respectively. In the wild-type human IgG 1 Fc region, the amino acids present at EU positions 252, 308 and 434 are M, V and N. By substituting these wild-type residues with “YPY”, the YPY variant Fc region can bind to FcRn with higher affinity and reduced pH dependence as compared with the Fc region of wild-type IgG antibodies.

The YPY variant Fc regions of the monovalent antigen-binding molecules of the present invention preferably comprise further alternations. Specifically, they may comprise amino acid substitutions in addition to the YPY mutations. These alternations, particularly amino acid substitutions, may further improve the properties of the monovalent antigen-binding molecules.

The YPY variant Fc regions of the monovalent antigen-binding molecules of the present invention may comprise any combination of the LALAPG mutations and/or knob-into-hole substitutions according to the embodiments already described in section (ii) above.

Accordingly, the present invention provides monovalent antigen-binding molecules comprising a variant Fc region, the variant Fc region comprising a first Fc domain and a second Fc domain, wherein the first Fc domain and the second Fc domain each comprise the amino acids A, A, G, Y, P and Y at EU positions 234, 235, 329, 252, 308, and 434, respectively.

The present invention further provides monovalent antigen-binding molecules comprising a variant Fc region, the variant Fc region comprising a first Fc domain and a second Fc domain, wherein the first Fc domain comprises the amino acids A, A, G, Y, P, Y and W at EU positions 234, 235, 329, 252, 308, 434 and 366, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, P, Y, S, A and V at EU positions 234, 235, 329, 252, 308, 434, 366, 368 and 407, respectively.

In certain embodiments, the first and second Fc domains of the variant Fc region are identical to the corresponding wild-type Fc domains, for example human IgG 1 Fc domains, but for the YPY mutations, the LALAPG mutations and the knob-into-hole mutations. Non-limiting examples of variant first and second Fc domains incorporating YPY mutations, for inclusion in the monovalent antigen-binding molecules described herein are set forth in Table 3B below.

In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 118. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 119. In certain embodiments, the first and/or second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 120.

In certain embodiments, the first or second Fc domain is linked to a heavy chain CH1 domain and the heavy chain constant region comprises or consists of the amino acid sequence set forth in any one of SEQ ID NOs: 121 , 125 and 129.

In certain embodiments, the variant Fc region comprises: a first Fc domain that comprises or consists of the amino acid sequence set forth in any one of SEQ ID NOs: 122, 123 and 124; and a second Fc domain that comprises or consists of the amino acid sequence set forth in any one of SEQ ID NOs: 126, 127 and 128.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 122; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 126.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 122; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 127.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 122; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 128.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 123; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 126.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 123; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 127.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 123; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 128.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 124; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 126.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 124; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 127.

In some embodiments, the variant Fc region comprises or consists of: a first Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 124; and a second Fc domain that comprises or consists of the amino acid sequence set forth in SEQ ID NO: 128.

Table 3B: Amino acid sequences of non-limiting examples of Fc domains and heavy chain constant regions incorporating Fc domains

(iv) One-armed antibodies

As noted elsewhere herein, the monovalent antigen-binding molecules of the invention encompass antibodies modified so as to be “one armed”. The single arm comprises an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences:

HCDR3 comprising or consisting of SEQ ID NO: 3;

HCDR2 comprising or consisting of SEQ ID NO: 2;

HCDR1 comprising or consisting of SEQ ID NO: 1 ;

LCDR3 comprising or consisting of SEQ ID NO: 6;

LCDR2 comprising or consisting of SEQ ID NO: 5; and

LCDR1 comprising or consisting of SEQ ID NO: 4.

The monovalent antigen-binding molecules of the invention are typically asymmetrical due to the presence of a single antigen-binding domain. The presence of a single antigenbinding domain means that, in certain embodiments, only one of the Fc domains of the variant Fc region is attached to the single antigen-binding domain.

In some embodiments, the antigen-binding domain is non-covalently attached to the variant Fc region. It is however preferred that the antigen-binding domain is covalently attached to the variant Fc region. The antigen-binding domain may be covalently attached to either the first Fc domain or the second Fc domain. The antigen-binding domain may be attached to the N-terminus of the first or second Fc domain. In other embodiments, the antigen-binding domain may be attached to the C- terminus of the first or second Fc domain. Alternatively, the antigen-binding domain can be attached at a position other than the N-terminus or the C-terminus of the first or second Fc domain.

In some embodiments, the C-terminus of the antigen-binding domain is attached to the N- terminus of either the first Fc domain or the second Fc domain. For embodiments in which the antigen-binding domain is a Fab, it is preferable that the C-terminus of the Fab heavy chain is attached to the N-terminus of either the first Fc domain or the second Fc domain.

The antigen-binding domain may be linked to the N-terminus of either the first Fc domain or the second Fc domain via a linker. In some embodiments, the linker is a non-cleavable linker. As used herein, the term “non-cleavable linker” refers to a linker that is not readily cleaved by one or more of a given enzyme (such as a protease), chemical agent, or photoirradiation. The linker can be a synthetic compound linker such as, for example, a chemical cross-linking agent. In some embodiments, the linker is a peptide linker. Examples of peptide linkers are well known. Any peptide linker could be used to link an antigen-binding domain to the first or second Fc domain in the monovalent antigen-binding molecules of the invention.

In preferred embodiments, the antigen binding domain is a Fab (described elsewhere herein). In such embodiments, the C-terminus of the Fab heavy chain is preferably attached to the N-terminus of either the first Fc domain or the second Fc domain.

In some embodiments, one or more additional amino acids are included between the C- terminus of the Fab heavy chain and the N-terminus of the first Fc domain or the second Fc domain.

In some embodiments, the C-terminus of the Fab heavy chain is attached to the N-terminus of the first Fc domain or second Fc domain via a hinge region as defined elsewhere herein or a portion thereof.

In some embodiments, the hinge is a naturally occurring hinge region. In some embodiments, the hinge region is an IgG hinge region. In preferred embodiments, the hinge region is a human IgG hinge region. In some embodiments, the IgG hinge region is selected from an: lgG1 , lgG2, lgG3 and lgG4 hinge region. In some embodiments, the human IgG hinge region is selected from an: lgG1 , lgG2, lgG3 and lgG4 hinge region.

In preferred embodiments, the antigen-binding domain is a Fab, and the C-terminus of the Fab heavy chain is attached to the N-terminus of the first Fc domain or the second Fc domain via an IgG hinge region. The IgG hinge region is preferably a human lgG1 hinge region.

As already noted herein, the monovalent antigen-binding molecules of the invention encompass one-armed antibodies. In some embodiments, the one-armed antibodies are one-armed IgG antibodies. In such embodiments, the one-armed IgG antibodies comprise a single Fab arm. Therefore, monovalent antigen-binding molecules of the invention encompass IgG antibodies lacking one of the two Fab arms.

In preferred embodiments, the monovalent antigen-binding molecules of the invention are modified IgG 1 antibodies having only one Fab arm. Preferably the modified antibodies are modified human lgG1 antibodies having only one Fab arm.

(v) Exemplary monovalent antigen-binding molecules that bind IgA

The antigen-binding molecules of the invention can be distinguished from the prior art on the basis that they are monovalent. This is an important distinction since the current application reports, for the first time, that monovalent IgA-binding molecules achieve improved IgA clearance in vivo as compared to bivalent IgA binding molecules (i.e. improved IgA sweeping). The improvement in IgA clearance was observed in terms of the speed of IgA removal as well as the depth of response seen.

Monovalent antigen-binding molecules of the invention comprise an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences:

HCDR3 comprising or consisting of SEQ ID NO: 3;

HCDR2 comprising or consisting of SEQ ID NO: 2;

HCDR1 comprising or consisting of SEQ ID NO: 1 ;

LCDR3 comprising or consisting of SEQ ID NO: 6; LCDR2 comprising or consisting of SEQ ID NO: 5; and LCDR1 comprising or consisting of SEQ ID NO: 4.

The monovalent antigen-binding molecules of the present invention exhibit binding specificity for human IgA. The monovalent antigen-binding molecules having the CDR, VH and/or VL amino acid sequences recited herein exhibit pH-dependent antigen binding, as described in section (i) above.

Exemplary pH-dependent monovalent antigen-binding molecules in accordance with the invention are described below with reference to specific CDR, VH and/or VL sequences.

In certain embodiments, the monovalent antigen-binding molecules comprise or consist of a variable heavy chain domain (VH) comprising or consisting of the amino acid sequence of SEQ ID NO: 7 or an amino acid sequence having at least 80%, 90%, 95%, 98% 99% identity thereto, and a variable light chain domain (VL) comprising or consisting of the amino acid sequence of SEQ ID NO: 8 or an amino acid sequence having at least 80%, 90%, 95%, 98% 99% identity thereto.

In further embodiments, the monovalent antigen-binding molecules comprise or consist of a variable heavy chain domain (VH) comprising or consisting of the amino acid sequence of SEQ ID NO: 7, and a variable light chain domain (VL) comprising or consisting of the amino acid sequence of SEQ ID NO: 8.

In preferred embodiments, the monovalent antigen-binding molecules comprise a variable heavy chain domain (VH) consisting of the amino acid sequence of SEQ ID NO: 7, and a variable light chain domain (VL) consisting of the amino acid sequence of SEQ ID NO: 8.

The exemplary monovalent antigen-binding molecules having any of the specific CDR, VH and/or VL domains recited above may comprise any of the variant Fc regions according to the embodiments described in section (ii) or section (iii) above.

In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise a variant human IgG Fc region, comprising a first variant human IgG Fc domain and a second variant human IgG Fc domain, wherein the first and second Fc domains each comprise the amino acids A, A, G, Y, T, E, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 433, 434 and 436, respectively. In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise a variant human IgG 1 Fc region, wherein the first and second Fc domains each comprise the amino acids A, A, G, Y, T, E, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 433, 434 and 436, respectively.

In preferred embodiments, the exemplary monovalent antigen-binding molecules described herein comprise a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, preferably a first and a second variant human IgG 1 Fc domain, wherein the first variant Fc domain comprises the amino acids A, A, G, Y, T, E, W, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 366, 433, 434 and 436, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, T, E, K, F, Y, S, A and V at EU positions 234, 235, 329, 252, 254, 256, 433, 434, 436, 366, 368 and 407, respectively.

In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, preferably a first and a second variant human IgG 1 Fc domain, wherein the first variant Fc domain comprises the amino acids A, A, G, Y, P, Y and W at EU positions 234, 235, 329, 252, 308, 434 and 366, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, P, Y, S, A and V at EU positions 234, 235, 329, 252, 308, 434, 366, 368 and 407, respectively.

In further preferred embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences:

HCDR3 comprising or consisting of SEQ ID NO: 3; HCDR2 comprising or consisting of SEQ ID NO: 2; HCDR1 comprising or consisting of SEQ ID NO: 1 ; LCDR3 comprising or consisting of SEQ ID NO: 6; LCDR2 comprising or consisting of SEQ ID NO: 5; and LCDR1 comprising or consisting of SEQ ID NO: 4; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first variant Fc domain comprises the amino acids A, A, G, Y, T, E, W, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 366, 433, 434 and 436, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, T, E, K, F, Y, S, A and V at EU positions 234, 235, 329, 252, 254, 256, 433, 434, 436, 366, 368 and 407, respectively; and

- wherein the antigen-binding domain is a Fab attached to the first Fc domain via an IgG hinge region.

In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences:

HCDR3 comprising or consisting of SEQ ID NO: 3; HCDR2 comprising or consisting of SEQ ID NO: 2; HCDR1 comprising or consisting of SEQ ID NO: 1 ; LCDR3 comprising or consisting of SEQ ID NO: 6; LCDR2 comprising or consisting of SEQ ID NO: 5; and LCDR1 comprising or consisting of SEQ ID NO: 4; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first variant Fc domain comprises the amino acids A, A, G, Y, P, Y and W at EU positions 234, 235, 329, 252, 308, 434 and 366, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, P, Y, S, A and V at EU positions 234, 235, 329, 252, 308, 434, 366, 368 and 407, respectively; and

Preferably, the C-terminus of the Fab heavy chain is attached to the N-terminus of the first variant Fc domain. Preferably, the variant human IgG Fc region is a variant human lgG1 Fc region. Preferably, the IgG hinge region is an lgG1 hinge region.

In further preferred embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) comprising or consisting of SEQ ID NO: 7 and a variable light chain domain (VL) comprising or consisting of SEQ ID NO: 8; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first variant Fc domain comprises the amino acids A, A, G, Y, T, E, W, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 366, 433, 434 and 436, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, T, E, K, F, Y, S, A and V at EU positions 234, 235, 329, 252, 254, 256, 433, 434, 436, 366, 368 and 407, respectively; and

In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) comprising or consisting of SEQ ID NO: 7 and a variable light chain domain (VL) comprising or consisting of SEQ ID NO: 8; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first variant Fc domain comprises the amino acids A, A, G, Y, P, Y and W at EU positions 234, 235, 329, 252, 308, 434 and 366, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, P, Y, S, A and V at EU positions 234, 235, 329, 252, 308, 434, 366, 368 and 407, respectively; and

In further preferred embodiments, the exemplary monovalent antigen-binding molecules described herein comprise a first Fc domain comprising or consisting of the amino acid sequence set forth in SEQ ID NO: 73 and a second Fc domain comprising or consisting of the amino acid sequence set forth in SEQ ID NO: 74.

In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise a first Fc domain comprising or consisting of the amino acid sequence set forth in SEQ ID NO: 114 and a second Fc domain comprising or consisting of the amino acid sequence set forth in SEQ ID NO: 115. In further preferred embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen-binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) and a variable light chain domain (VL) wherein the VH and VL domains comprise the CDR sequences:

HCDR3 comprising or consisting of SEQ ID NO: 3; HCDR2 comprising or consisting of SEQ ID NO: 2; HCDR1 comprising or consisting of SEQ ID NO: 1 ; LCDR3 comprising or consisting of SEQ ID NO: 6; LCDR2 comprising or consisting of SEQ ID NO: 5; and LCDR1 comprising or consisting of SEQ ID NO: 4; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 73 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 74; and

HCDR3 comprising or consisting of SEQ ID NO: 3; HCDR2 comprising or consisting of SEQ ID NO: 2; HCDR1 comprising or consisting of SEQ ID NO: 1 ; LCDR3 comprising or consisting of SEQ ID NO: 6; LCDR2 comprising or consisting of SEQ ID NO: 5; and LCDR1 comprising or consisting of SEQ ID NO: 4; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 114 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 115; and

- wherein the antigen-binding domain is a Fab attached to the first Fc domain via an IgG hinge region. Preferably, the C-terminus of the Fab heavy chain is attached to the N-terminus of the first variant Fc domain. Preferably, the variant human IgG Fc region is a variant human lgG1 Fc region. Preferably, the IgG hinge region is an lgG1 hinge region.

In further preferred embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) comprising or consisting of SEQ ID NO: 7 and a variable light chain domain (VL) comprising or consisting of SEQ ID NO: 8; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 73 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 74; and

In certain embodiments, the exemplary monovalent antigen-binding molecules described herein comprise: an antigen binding domain that binds to IgA, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) comprising or consisting of SEQ ID NO: 7 and a variable light chain domain (VL) comprising or consisting of SEQ ID NO: 8; a variant human IgG Fc region comprising or consisting of a first and a second variant human IgG Fc domain, wherein the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 114 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 115; and

Preferably, the C-terminus of the Fab heavy chain is attached to the N-terminus of the first variant Fc domain. Preferably, the variant human IgG Fc region is a variant human lgG1 Fc region. Preferably, the IgG hinge region is an lgG1 hinge region. In preferred embodiments, the exemplary monovalent antigen-binding molecules of the invention comprise or consist of:

(i) a first immunoglobulin heavy chain comprising or consisting of the amino acid sequence of SEQ ID NO: 75;

(ii) a second immunoglobulin heavy chain comprising or consisting of the amino acid sequence of SEQ ID NO: 76; and

(iii) an immunoglobulin light chain comprising or consisting of the amino acid sequence of SEQ ID NO: 77.

In said embodiments, preferably the first immunoglobulin heavy chain pairs with the second immunoglobulin heavy chain to form a variant IgG 1 Fc region, and the first immunoglobulin heavy chain pairs with the immunoglobulin light chain to form a Fab that binds to IgA.

In preferred embodiments, the exemplary monovalent antigen-binding molecules of the invention comprise or consist of:

(i) a first immunoglobulin heavy chain comprising or consisting of the amino acid sequence of SEQ ID NO: 116;

(ii) a second immunoglobulin heavy chain comprising or consisting of the amino acid sequence of SEQ ID NO: 117; and

In a particular embodiment, provided herein is a monovalent antigen-binding molecule that binds to IgA consisting of:

In a further embodiment, provided herein is a monovalent antigen-binding molecule that binds to IgA consisting of:

(i) a first immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 116; (ii) a second immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 117; and

Table 4: Heavy chain CDR sequences

Table 5: Light chain CDR sequences

Table 6: VH and VL sequences

Table 7: Fc region sequences

Table 8: Sequence information for gClone B HIS H with LALAPG-ABDEG OA or LALAPG-YPY OA

C. Polynucleotides encoding monovalent antigen-binding molecules

The invention also provides polynucleotide molecules encoding the monovalent antigenbinding molecules of the invention or fragments thereof. Also encompassed are expression vectors containing said nucleotide sequences of the invention operably linked to regulatory sequences which permit expression of the monovalent antigen-binding molecules or fragments thereof in a host cell or cell-free expression system, and a host cell or cell-free expression system containing this expression vector.

Polynucleotide molecules encoding the monovalent antigen-binding molecules of the invention include, for example, recombinant DNA molecules. The terms "nucleic acid", “polynucleotide” or a "polynucleotide molecule" as used herein interchangeably and refer to any DNA or RNA molecule, either single- or double-stranded and, if single-stranded, the molecule of its complementary sequence. In discussing nucleic acid molecules, a sequence or structure of a particular nucleic acid molecule may be described herein according to the normal convention of providing the sequence in the 5' to 3' direction. In some embodiments of the invention, nucleic acids or polynucleotides are "isolated". This term, when applied to a nucleic acid molecule, refers to a nucleic acid molecule that is separated from sequences with which it is immediately contiguous in the naturally occurring genome of the organism in which it originated. For example, an "isolated nucleic acid" may comprise a DNA molecule inserted into a vector, such as a plasmid or virus vector, or integrated into the genomic DNA of a prokaryotic or eukaryotic cell or non-human host organism. When applied to RNA, the term "isolated polynucleotide" refers primarily to an RNA molecule encoded by an isolated DNA molecule as defined above. Alternatively, the term may refer to an RNA molecule that has been purified/separated from other nucleic acids with which it would be associated in its natural state (i.e. , in cells or tissues). An isolated polynucleotide (either DNA or RNA) may further represent a molecule produced directly by biological or synthetic means and separated from other components present during its production.

For recombinant production of a monovalent antigen-binding molecule according to the invention, a recombinant polynucleotide encoding it may be prepared (using standard molecular biology techniques) and inserted into a replicable vector for expression in a chosen host cell, or a cell-free expression system. Suitable host cells may be prokaryote, yeast, or higher eukaryote cells, specifically mammalian cells. Examples of useful mammalian host cell lines are monkey kidney CV1 line transformed by SV40 (COS-7, ATCC CRL 1651 ); human embryonic kidney line (293 or 293 cells subcloned for growth in suspension culture, Graham et aL, J. Gen. Virol. 36:59 (1977)); baby hamster kidney cells (BHK, ATCC CCL 10); Chinese hamster ovary cells/-DHFR (CHO, Urlaub et aL, Proc. Natl. Acad. Sci. USA 77:4216 (1980)); mouse sertoli cells (TM4, Mather, Biol. Reprod. 23:243- 251 (1980)); mouse myeloma cells SP2/0-AG14 (ATCC CRL 1581 ; ATCC CRL 8287) or NSO (HPA culture collections no. 85110503); monkey kidney cells (CV1 ATCC CCL 70); African green monkey kidney cells (VERO-76, ATCC CRL-1587); human cervical carcinoma cells (HELA, ATCC CCL 2); canine kidney cells (MDCK, ATCC CCL 34); buffalo rat liver cells (BRL 3A, ATCC CRL 1442); human lung cells (W138, ATCC CCL 75); human liver cells (Hep G2, HB 8065); mouse mammary tumor (MMT 060562, ATCC CCL51); TRI cells (Mather et aL, Annals N.Y. Acad. Sci. 383:44-68 (1982)); MRC 5 cells; FS4 cells; and a human hepatoma line (Hep G2), as well as DSM’s PERC-6 cell line. Expression vectors suitable for use in each of these host cells are also generally known in the art.

It should be noted that the term "host cell" generally refers to a cultured cell line. Whole human beings into which an expression vector encoding an antigen-binding molecule according to the invention has been introduced are explicitly excluded from the definition of a “host cell”.

D. Monovalent antigen-binding molecule production

In a further aspect, the invention also provides a method of producing monovalent antigenbinding molecules of the invention which comprises culturing a host cell (or cell free expression system) containing polynucleotide (i.e. an expression vector) encoding the monovalent antigen-binding molecule under conditions which permit expression of the monovalent antigen-binding molecule, and recovering the expressed monovalent antigenbinding molecule. This recombinant expression process can be used for large scale production of the monovalent antigen-binding molecules according to the invention, including molecules intended for human therapeutic use. Suitable vectors, cell lines and production processes for large scale manufacture of monovalent antigen-binding molecules (i.e. recombinant modified antibodies) suitable for in vivo therapeutic use are generally available in the art and will be well known to the skilled person. E. Pharmaceutical compositions

The scope of the invention includes pharmaceutical compositions, containing one or a combination of monovalent antigen-binding molecules of the invention formulated with one or more pharmaceutically acceptable carriers or excipients. Such compositions may include one or a combination of (e.g., two or more different) monovalent antigen-binding molecules as described herein. Techniques for formulating monoclonal antibodies for human therapeutic use are well known in the art and are reviewed, for example, in Wang et al., Journal of Pharmaceutical Sciences, Vol.96, pp1-26, 2007, the contents of which are incorporated herein in their entirety.

Pharmaceutically acceptable excipients that may be used to formulate the compositions include, but are not limited to, ion exchangers, alumina, aluminum stearate, lecithin, serum proteins, such as human serum albumin, buffer substances such as phosphates, glycine, sorbic acid, potassium sorbate, partial glyceride mixtures of saturated vegetable fatty acids, water, salts or electrolytes, such as protamine sulfate, disodium hydrogen phosphate, potassium hydrogen phosphate, sodium chloride, zinc salts, colloidal silica, magnesium trisilicate, polyvinyl pyrrolidone, cellulose-based substances (for example sodium carboxymethylcellulose), polyethylene glycol, polyacrylates, waxes, polyethylene- polyoxypropylene- block polymers, polyethylene glycol and wool fat.

In certain embodiments, the pharmaceutical compositions are formulated for administration to a subject via any suitable route of administration including but not limited to intramuscular, intravenous, intradermal, intraperitoneal injection, subcutaneous, epidural, nasal, oral, rectal, topical, inhalational, buccal (e.g., sublingual), and transdermal administration. In preferred embodiments, the composition is formulated for intravenous or subcutaneous administration.

F. Methods of treatment

The monovalent antigen-binding molecules and pharmaceutical compositions as described herein are intended for use in methods of treatment. The present invention thus provides monovalent antigen-binding molecules in accordance with the invention (and described in section B above) or pharmaceutical compositions comprising the same for use as medicaments. Further provided are methods of treating a disorder in a subject, the methods comprising administering to a patient in need thereof a therapeutically effective amount of a monovalent antigen-binding molecule in accordance with the invention (and described in section B above) or a pharmaceutical composition comprising the same. The invention also provides monovalent antigen-binding molecules in accordance with the invention (and described in section B above) or pharmaceutical compositions comprising the same for use in the treatment of a disorder in a subject in need thereof. The invention further provides monovalent antigen-binding molecules in accordance with the invention (and described in section B above) or pharmaceutical compositions comprising the same for the manufacture of a medicament for treating a disorder in a subject in need thereof. The disorder is preferably an antibody-mediated disorder (as defined elsewhere). The subject is preferably human. All embodiments described above in relation to the monovalent antigen-binding molecules and pharmaceutical compositions of the invention are equally applicable to the methods described herein.

In certain embodiments, the disorder treated in accordance with the methods described herein is an autoantibody-mediated disorder. In certain embodiments, the disorder is an IgA-mediated disorder. In certain embodiments, the disorder is an IgA autoantibody- mediated disorder.

G. Methods of reducing IgA levels

The monovalent antigen-binding molecules and pharmaceutical compositions as described herein are also intended for use in methods of reducing IgA levels. The invention thus provides a method of reducing IgA levels in a subject, wherein the method comprises the step of: administering a monovalent antigen-binding molecule in accordance with the invention (and described in section B above) or a pharmaceutical composition comprising the same.

The invention further provides a method of reducing IgA levels in a subject, wherein the method comprises the step of: administering a monovalent antigen-binding molecule in accordance with the invention (and described in section B above) or a pharmaceutical composition comprising the same; wherein the IgA reduction is measured relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the method is for reducing serum IgA levels in the subject.

In certain embodiments, the subject has a high baseline level of IgA. In certain embodiments, the subject has a high baseline level of serum IgA. In certain embodiments, a high baseline level of serum IgA is a serum concentration of IgA that is greater than 3 mg/mL. In certain embodiments, the subject has a medium baseline level of IgA. In certain embodiments, the subject has a medium baseline level of serum IgA. In certain embodiments, a medium baseline level of serum IgA is a serum concentration of IgA in the range of 1-3 mg/mL. In certain embodiments, the subject has a low baseline level of IgA. In certain embodiments, the subject has a low baseline level of serum IgA. In certain embodiments, a low baseline level of serum IgA is a serum concentration of IgA that is less than 1 mg/mL.

The subject is preferably human.

In certain embodiments, the method comprises the step of administering a single intravenous dose of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the single intravenous dose administered is 12 mg/kg.

In certain embodiments, the method comprises the step of administering a single dose of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the single dose is 3 mg/kg. In certain embodiments, the single dose is 10 mg/kg. In certain embodiments, the single dose is 30 mg/kg.

In certain embodiments, the method comprises the step of administering multiple doses of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same, wherein the administration of each individual dose is separated by a defined dosing interval. Multiple doses may refer to at least two, at least three, at least four, at least five doses.

In certain embodiments, the subject receives multiple doses of 3 mg/kg, administered weekly. In certain embodiments, the subject receives multiple doses of 10 mg/kg, administered weekly. In certain embodiments, the subject receives multiple doses of 10 mg/kg, administered fortnightly. In certain embodiments, the subject receives multiple doses of 30 mg/kg, administered monthly.

The level of IgA reduction may be expressed as a percentage reduction relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule in accordance with the invention (and described in section B above) or a pharmaceutical composition comprising the same. By way of example, a reduction of 100% in this context means that the IgA levels were reduced to 0 after administration of the monovalent antigen-binding or the pharmaceutical composition comprising the same. A reduction of at least 90% in this context means that 90% of the baseline IgA level were eradicated after administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same.

In certain embodiments, the level of IgA is reduced by at least 50% relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the level of IgA is reduced by at least 60% relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the level of IgA is reduced by at least 70% relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the level of IgA is reduced by at least 80% relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the level of IgA is reduced by at least 85% relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the level of IgA is reduced by at least 90% relative to the baseline IgA level in the subject prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same.

In certain embodiments, the percentage IgA reduction relative to the baseline IgA level prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same is achieved within 1 month of administering the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the percentage IgA reduction relative to the baseline IgA level prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same, is achieved within 21 days of administering the first dose of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the percentage IgA reduction relative to the baseline IgA level prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same, is achieved within a fortnight of administering the first dose of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same. In certain embodiments, the percentage IgA reduction relative to the baseline IgA level prior to administration of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same, is achieved within 1 week of administering the first dose of the monovalent antigen-binding molecule or the pharmaceutical composition comprising the same.

All embodiments described above in relation to the monovalent antigen-binding molecules and pharmaceutical compositions of the invention are equally applicable to the methods described herein.

INCORPORATION BY REFERENCE

Various publications are cited in the foregoing description and throughout the following examples, each of which is incorporated by reference herein in its entirety.

EXAMPLES

The invention will be further understood with reference to the following non-limiting examples.

Example 1: Anti-lqA antibody development utilizing the SIMPLE antibody platform

The following example describes how clones A and D were generated via llama immunization. Llamas were immunized with polyclonal immunoglobulin A (IgA from human serum, Sigma). Peripheral blood lymphocytes isolated from immunized llamas were used for RNA extraction, RT-PCR and PCR-cloning of the variable domains in a single-chain variable fragment (scFv) phagemid vector. Panning phage display selections were performed for up to three rounds using human IgA Isotype Control (ThermoFisher), counter selecting with total human IgG (Sigma) and using citrate phosphate acetate buffer (CPA) elution to enrich for pH-dependent antigen binding.

Masterplates were generated upon enrichment after at least two consecutive rounds of phage display selections, both from total as well as pH-dependent elution conditions. Individual clones were grown in a 96-deep well plate and periplasmic fractions were prepared. These periplasmic extracts (P.E.) (containing scFv), were tested in an enzyme- linked immunosorbent assay (ELISA) for binding to full-length human IgA (ThermoFisher) and human IgG (Sigma). Less than 3% of the selected clones cross-reacted with human IgG, demonstrating IgA specificity.

Only scFvs that specifically bound to IgA in ELISA were further screened using surface plasmon resonance (SPR) for off-rate determination on a CM5 chip coated with full-length human IgA or human IgG. Binders with good affinity and specificity for human IgA at pH 7.4 were then sequenced. 244 valid sequences were obtained, containing 169 unique VH-gs- VL sequences with 56 different heavy chain complementarity-determining regions (HCDR3) grouped into 46 different HCDR3 families.

These unique VH-gs-VL clones were further screened using SPR to assess association and dissociation to human IgA (hulgA) and human IgG (huIgG) at pH 5.5. In addition to hulgA binding, all unique scFv variant clones were also screened for their competing properties. A CM5 chip was immobilized with recombinant human FCAR/CD89 receptor (hFcoRI) (R&D systems), followed by sequential capture of hulgA and anti-lgA P.E. clones. Anti-lgA P.E. clones that bound to captured human IgA were considered non-competitive binders. Clones that demonstrated a faster dissociation as compared to the blank control were capable of displacing IgA from its receptor hFcoRL Next to the sequential addition, a pre-mix was made of hulgA and anti-lgA P.E. to confirm the competing properties of the different scFv variants. 99 of the total 169 anti-hulgA scFv P.E. analyzed were capable of at least partially blocking the binding of IgA to its receptor FcoRL A subset of 23 scFv P.E. clones were selected based on their affinity for hulgA, the extent of pH-dependent antigen binding and/or their competitive binding properties. The 23 scFv P.E. clones included clones A and D (characterized further herein).

The 23 scFv P.E. clones were cloned into a vector containing the sequence of human IgG 1 and antibodies were produced in HEK293 cells and purified on a protein A column.

All 23 antibodies exhibited binding to human IgA. However, only 18 anti-lgA antibodies showed competition for binding of IgA to its receptor CD89, including clones A and D.

Two anti-lgA llama-derived antibody clones (A and D) were selected for further testing because they exhibited the following favourable properties - (i) competitive antigen binding; (ii) pH-dependent antigen binding; and (iii) cross-reactivity with rhesus IgA.

Example 2: Further testing and optimization of clones A and D

This example describes how the llama-derived antibody clones A and D were engineered in order to further improve their properties.

2.1 Stability testing of clones A and D

Both of the selected anti-lgA antibodies from Example 1 were subjected to stability testing.

In more detail, each clone was diluted in sterile Dulbecco’s PBS (Sigma-Aldrich) and incubated at 37°C for up to 4 weeks in a glass vial (Supelco). Intermittent sample collection was carried out weekly to determine the stability of the different anti-lgA antibodies over time. Every week, samples were subjected to a visual inspection, and after 4 weeks of stability testing, all samples were analysed via Nanodrop, Capillary electrophoresis sodium dodecyl sulfate (CE-SDS) and SPR. A CM5 chip was coated with human Mota-IgA (inhouse production) and the association of the test samples at pH 7.4 was compared with a dilution series of corresponding control samples in HBS-EP in order to assess loss of potency due to the incubation at high temperature.

After 4 weeks at 37°C, clone D had a potency loss of 63% and clone A had a potency loss of 12%. In addition to potency testing, the samples were screened for the most common post- translational modifications (PTMs) using reduced peptide map analysis. PTMs included in this analysis were oxidation, deamidation, isomerization, glycosylation, and N/C terminal processing in the VH and VL. These studies identified several potential liabilities in the peptide sequences of clones A and D. Specifically, a deamidation site at position 56 of clone D’s variable heavy chain (VH) and at position 33 of its light chain (LC). An isomerization site at position 98 of clone A’s VH. In addition, a clipping site between positions 52G and 53S of the HCDR2 of clone A was detected by mass spectrometry.

2.2 Light chain shuffling of clones A and D

Both clones were subjected to light chain shuffling in order to optimise the antigen-binding characteristics of the antibodies. After light chain shuffling, periplasmic fractions were prepared and tested for binding to human and rhesus IgA at pH 7.4, followed by dissociation at pH 7.4 or pH 5.5 using ELISA and/or SPR.

Based on these tests, ten light chain shuffled variants of clone D and eleven light chain shuffled variants of clone A were selected for analysis.

The majority of the clones however, exhibited reduced pH-dependent antigen binding.

2.3 Clone A optimization and engineering

As the light chain shuffled variants of clone A exhibited almost no pH-dependent antigen binding, their characterisation was discontinued.

The sequence of clone A was optimised by grafting the CDRs into the closest human germline sequence (X59315|IGKV1-39*01 , Z12342|IGHV3-53*02). This clone is referred to herein as germlined Clone A (gClone A). Screening of human IgA affinity, pH-dependency and thermostability was performed using SPR. Additionally, ELISA was used to test competition and displacement of IgA from its receptor CD89. Germlining drastically increased the pH-dependent binding to IgA of the antibody.

Separately, further optimization of the pH-dependent antigen binding characteristics of clone A was achieved by histidine (HIS) engineering. A total of 17 single histidine residues were introduced into the CDRs of the VK and 26 histidine residues were introduced into clone A’s VH domain. Screening for human and rhesus IgA pH dependency was performed using SPR and ELISA. These tests revealed several spots in the VH and VK which improved pH-dependent antigen binding. Each selected histidine mutation in the VK was combined with the selected histidine mutations in the VH to generate double histidine engineered clones. These further engineered clones were again screened for human and rhesus IgA pH dependency using SPR and ELISA.

Next, different histidine spots in the VH of clone A were combined into double, triple and quadruple histidine engineered variants and again screened using SPR for pH-dependent antigen binding on human and rhesus IgA. From these screening studies, a triple histidine engineered variant of clone A with a histidine residue at position 32 (VH2), position 56 (VH9) and position 59 (VH26) showed the highest pH-dependent antigen binding with both human and rhesus IgA and was therefore selected for further optimization.

The three histidine substitutions of clone A (at positions 32, 56 and 59) were then introduced into the sequence of gClone A to further increase pH-dependent IgA binding. Whilst there was only a minor increase in pH-dependent binding to human IgA, the three additional histidine substitutions to gClone A clearly increased pH-dependent binding to rhesus IgA. Therefore both variants were selected for further characterisation and comparison. They are referred to herein as “gClone A” (i.e. a germlined clone A variant without His substitutions at positions 32, 56 and 59) and “gClone A HIS” (i.e. the germlined clone A variant with His substitutions at positions 32, 56 and 59).

Since reduced peptide map analysis identified position D98 in the VH of clone A as being prone to isomerization, this position was substituted with other amino acids in gClone A and gClone A HIS. The resulting variants were screened for human, rhesus and cynomolgus monkey IgA affinity, pH dependency and lgA:CD89 competition and displacement on Biacore® and/or ELISA. Based on these studies the variant with glutamic acid (E) at position 98 was selected for further study. The engineered variants comprising the D98E substitution are referred to herein as “germlined Clone A E” or “gClone A E” and “germlined Clone A E HIS” or “gClone A E HIS”.

As noted above, a clipping site in HCDR2 of clone A was detected by mass spectrometry between position 52G and 53S. A follow-up stability study confirmed the presence of this potential liability in the parental clone A and the two derivative clones gClone A E HIS and gClone A E. In order to avoid clipping of HCDR2, positions 52G and 53S were randomly mutated in variants gClone A E HIS and gClone A E. The different variants were screened for human and cynomolgus monkey IgA affinity and pH-dependent antigen binding using SPR. lgA:CD89 competition and displacement studies were performed using ELISA. The S53K variant demonstrated similar pH-dependent antigen binding and affinity to both human and cynomolgus monkey IgA relative to gClone A E HIS and gClone A E variants. Moreover, a similar CD89 blocking capacity was observed for the S53K variant as compared to the variants gClone A E HIS and gClone A E. Given the similarities to the parental clone, the mutation S53K was selected to avoid clipping of HCDR2 in the gClone A E HIS and gClone A E variants. The removal of the clipping site was confirmed via an additional stability study followed by mass spectrometry analysis. The resulting two engineered variants are referred to herein as “germlined Clone A EK” or “gClone A EK” and “germlined Clone A EK HIS” or “gClone A EK HIS”.

The different clone A variants are summarized in Table 9.1 below.

Table 9.1 : Summary of Clone A variants 2.4 Optimization and engineering of clone D variants

As reported in section 2.2 above, the majority of the clone D variants generated by light chain shuffling exhibited reduced pH-dependent antigen binding as compared to original clone D.

As with clone A, the pH-dependent antigen binding characteristics were optimised via histidine engineering.

For clone D light chain shuffled variants, ten of the resulting clones were selected for re- introduction of a histidine at position 93 in the LCDR3, similar to the parental clone, since it was postulated that such a mutation could restore pH-dependent antigen binding. Two of the ten light chain shuffled variants were clone B and clone C. The clone B variant containing histidine at position 93 in the LCDR3 is referred to herein as “Clone B histidine” (clone B HIS). Re-introduction of a histidine residue at position 93 (LCDR3), resulted in enhanced pH-dependent antigen binding in the majority of the ten light chain shuffled variants of clone D (except for some variants such as clone C).

Four of the ten light chain shuffled variants did not exhibit any pH-dependent antigen binding (with or without reintroduction of histidine at position 93) and as such, these variants were excluded from further characterization.

The six remaining clone D variants (including clone B, clone B HIS and clone C) were then screened for affinity and pH-dependent antigen binding to human and rhesus IgA using SPR. They were also tested for their ability to compete and displace IgA from the CD89 receptor. Based on the results of these tests, four variants including clone B HIS and clone C (i.e. light chain shuffled variant with no reintroduction of 93HIS) were selected for further optimization.

Whilst all selected clone D variants already exhibited high human identity, this was further improved by grafting the CDRs of the variants onto the closest human germline sequence (M99660|IGHV3-23*01 , M94116|IGLV1 -40*01 ). The same human germline sequence was used for all variant light chains. Whilst the CDR grafting altered a serine at position 29 to a phenylalanine in the heavy chain (HC), the serine appeared to be crucial for binding to human IgA. As such, all germlined clone D variants were produced with the endogenous serine at position 29 in the HC. The resulting clone B variant was called “germlined Clone B histidine” or “gClone B HIS”. The resulting clone C variant was called “germlined Clone C” or “gClone C”.

The germlined variants were further screened for affinity to human and cynomolgus monkey IgA, pH-dependent antigen binding as well as lgA:CD89 competition and displacement. These tests revealed that one of the germlined variants (not clone B HIS or clone C) had a reduced affinity for cynomolgus monkey IgA and this variant was therefore excluded from further analysis.

A deamidation site was discovered at position 33 (LCDR1 ) of clone D’s VL domain. This potential liability site was also present in all selected light chain shuffled variants with the exception of clone B. To remove this potential liability, different positions were mutated in LCDR1 (Y32, N33 or Y34) of clone C and a second clone, and screened for human, rhesus and cynomolgus monkey IgA affinity, pH-dependent antigen binding and lgA:CD89 competition/displacement using SPR and/or ELISA. For the second clone, introduction of the aforementioned mutations resulted in drastic loss of affinity for human IgA and therefore this variant was not selected for further analysis.

In contrast, screening of the different variants of clone C resulted in the selection of threonine (T) at position 33, which no longer contained a deamidation site but did not significantly affect binding to human IgA, rhesus IgA or cynomolgus monkey IgA.

In addition, a potential liability was also identified in the VH of clone D at position 56 (HCDR2). This position was randomly mutated to find the most optimal amino acid as replacement for the asparagine (N) residue. All variants of clone C and clone B HIS were screened for human and cynomolgus monkey IgA affinity, pH-dependent antigen binding and lgA:CD89 competition and displacement using SPR and/or ELISA.

For gClone C, this resulted in the selection of N56A (germlined Clone C TA or gClone C TA) and N56Q (germlined Clone C TQ or gClone C TQ) variants that showed the highest pH-dependent antigen binding and CD89 blocking capacity, respectively.

For gclone B HIS, three variants were selected. The N56H variant (germlined clone B histidine H (gClone B HIS H)) was selected since it demonstrated the highest pH dependency for human IgA. The N56A variant (germlined clone B HIS A (gClone B HIS A)) was selected because it showed better cynomolgus monkey IgA cross-reactivity. The N56L variant (germlined clone B HIS L (gClone B HIS L)) was selected since it had similar properties to the N56H variant.

To summarise the above, llama-derived clone D was engineered to optimise its characteristics. After engineering, five different variants of clone D were selected for further characterisation. These clones are referred to herein as “germlined Clone C TA” (gClone C TA), “germlined Clone C TQ” (gClone C TQ), “germlined clone B histidine H” (gClone B HIS H), “germlined clone B histidine A” (g Clone B HIS A) and “germlined clone B histidine L” (gClone B HIS L).

These five clone D variants are summarized in Table 9.2 below.

Table 9.2: Summary of clone D variants.

Example 3: Anti-lqA Fab selection

This example describes the experiments that led to selection of variant “gClone B HIS H” for further testing. 3.1 pH-dependent binding to human and cynomolqus monkey IgA using SPR

A crucial property of an IgA sweeping antibody is that it possesses pH-dependent binding to its target antigen (i.e. IgA), such that the antigen is released in the endosome due to the acidic environment. To maximize the lysosomal degradation of the antigen, re-binding of the target in the endosome should also be avoided. Therefore, affinity as well as antigenbinding at acidic pH of the anti-lgA antibodies described in Example 2 to human and cynomolgus monkey serum IgA was analysed using SPR. The clones tested were: gClone A E; gClone A E HIS; gClone C TA; gClone C TQ; gClone B HIS L; gClone B HIS A; and gClone B HIS H.

In more detail, affinity for human IgA was determined using a CM5 chip coated with human Mota-IgA (in-house production) by association of the test antibody at pH 7.4 and dissociation at pH 7.4 or 5.5. More in-depth characterization of the affinity and pH- dependency was done on a CM5 chip coated with anti-human Fc IgG (Jackson ImmunoResearch) followed by capturing the anti-lgA variants and association of serum human IgA (Abeam) or serum cynomolgus money IgA (Life diagnostics) at pH 7.4 and dissociation at pH 7.4 or pH 5.5.

Upon comparison of the two gClone A variants, a slight improvement in pH-dependent binding of gClone A E HIS to human IgA compared to gClone A E was observed, but this difference was more pronounced with cynomolgus monkey IgA. Whilst there was only a minor decrease in affinity with the HIS-engineered variant (gClone A E; KD 5,02E-09 and gClone A E HIS; KD 5,51 E-09), gClone A E reached higher RU levels for both human and cynomolgus monkey IgA.

For the variants of clone D; gClone C TQ had the highest affinity for human IgA (KD 1 ,90E- 09). However, this variant exhibited reduced pH-dependent antigen binding as compared to gClone C TA.

All gClone B HIS variants exhibited strong pH-dependent antigen binding to human IgA. The variant with the greatest pH-dependent human IgA binding was gClone B HIS H. The three gClone B HIS variants differed in their affinity to cynomolgus monkey IgA - gClone B HIS A (KD 9,17E-08) demonstrated the strongest binding, followed by gClone B HIS H (KD 1 ,13E-08) and gClone B HIS L (KD 1 ,29E-07). The three gClone B HIS variants had comparable affinity for human IgA - gClone B HIS A (KD 2,16E-08), gClone B HIS H (KD 2.43E-08) and g Clone B HIS L (KD 2.02E-08). Amongst the gClone B HIS variants, the gClone B HIS H variant demonstrated the highest pH-dependent binding for both human and cynomolgus monkey IgA.

The same SPR experimental method as described above was also used to investigate association and dissociation of variants to human and cynomolgus monkey IgA at different pHs - pH 7.4, pH 6 or pH 5. These experiments were performed in order to accommodate for differences in the pH of the endosome and lysosome. The percentage binding reduction at acidic pH compared to pH 7.4 was calculated based on Rmax values and the results are summarised in Table 10:

Table 10: Percentage (%) binding reduction of the anti-lgA antibodies to human or cynomolgus monkey IgA at pH 6 and 5 compared to pH 7.4.

In general, reduced binding to human IgA was observed at pH 5 as compared to binding at pH 6, with only variants gClone A E and gClone C TQ showing some residual binding of 25% and approximately 15% at pH 5, respectively. All other variants did not bind to human IgA at pH 5.

Whilst only a 27% reduction in human IgA binding of variant gClone A E was observed at pH 6, the introduction of the three histidine residues in the HC resulted in 50% reduced binding at the same acidic pH. Variant gClone A E showed only a minimal reduction in binding to cynomolgus monkey IgA at pH 6 (15%), while the HIS engineered variant gClone A E HIS demonstrated a 60% reduction.

In addition to variant gClone A E, a minimal reduction in human IgA binding at pH 6 was observed for variant gClone C TQ (30%). Variant gClone C TA showed less residual binding to human IgA at pH 6 than variant gClone C TQ, but this was still only a 42% reduction. Variant gClone C TQ showed a minimal reduction in binding to cynomolgus monkey IgA at pH 6 (34%) and variant gClone C TA showed a 48% reduction.

In contrast to the gClone A and gClone C variants, all gClone B HIS variants showed at least -50% reduction in binding to human IgA at pH 6, with gClone B HIS H demonstrating the highest reduction in binding to human IgA at pH 6 of -80%. All gClone B HIS variants also showed the lowest levels of cynomolgus monkey IgA binding at acidic pHs and two of the variants (gClone B HIS H and gClone B HIS L) even had 100% reduced IgA binding at pH 6.

3.2 CD89 receptor blocking with anti-lqA mAbs

It is described in the art that monomeric IgA binds CD89 receptor with relatively low affinity (Ka of ± 10⁶ M'¹), while IgA-immune complexes bind with high avidity and can crosslink the receptor (Wines et al. 1999; Wines et al. 2001). Therefore, the anti-lgA variants were tested for their ability to block the binding of human monomeric IgA as well as human IgA immune complexes to the receptor CD89.

3.2.1 Competition of anti-lqA mAbs for human monomeric lqA:CD89 interaction using ELISA

Monomeric human lgA:CD89 competition was analyzed by coating a 96-well half area high binding microplate (Greiner) with recombinant human CD89 His-tag (Elabscience). After washing and blocking the plate, a dilution series of the anti-lgA antibodies in 0.1% casein- PBS, pre-incubated with a fixed concentration of serum human IgA (Abeam) was added to the plate and allowed to bind for 1 hour at room temperature. Binding of IgA was detected with mouse anti-human IgA (Abeam) and peroxidase-conjugated Donkey Anti-Mouse IgG (H+L) (Jackson Immunoresearch) and developed with s(SH)TMB (SDT Reagents for life). OD was measured at 450 nm (ref 620 nm) with a 96-well ELISA plate reader (Tecan Sunrise). lgA:CD89 displacement was analyzed with a similar set-up, but IgA was incubated 1 hour before the addition of the dilution series of anti-lgA antibodies, with a washing step in between.

All variants that were tested competed with IgA for CD89 binding and were also able to displace human monomeric IgA from its receptor CD89 (Table 11). Table 11 : EC50 values of the hlgA:hCD89 competition and displacement ELISA

3.2.2 Competition of anti-lqA mAbs for human IgA immune complexes:CD89 interaction on RBL cells

To complement the competition and displacement data generated by ELISA and to evaluate the competitive properties of the anti-lgA antibodies with regard to IgA immune complexes, an in vitro assay using rat basophilic leukemia (RBL) cells, stably transfected with human CD89, was performed.

For these experiments, RBL cells were seeded in a 96 well V-bottom plate. A preincubation mix of the anti-lgA antibodies and Mota-IgA: RSV-F complex was added to the cells. After washing with FACS buffer, cell-bound IgA was stained at 4°C with goat antihuman IgA-PE (Southern Biotech); in a mix also containing Fixable Viability Dye eFluor 780 (eBioscience), and Purified Mouse Anti-Rat CD32 (Pharmigen). Data acquisition was performed by flow cytometry, using the BD LSRFortessa™X-20 Cell Analyzer.

The results indicated that all variants were able to compete with human IgA immune complexes (Mota-IgA: RSV-F) for binding to human CD89 in a cell-based assay. The highest degree of blocking was observed for gClone A EK HIS, followed by g Clone B HIS H and gClone B HIS L, respectively (Table 12).

Table 12: EC50 values of the hlgA IC:hCD89 competition assay on RBL-hCD89-GFP cells using flow cytometry

3.2.2 Anti-lqA mAbs block IgA-mediated phagocytosis of CD89-expressinq primary PMNs and monocytes In order to translate the blocking capacity of the anti-lgA antibodies in a functional readout, the ability of these mAbs to inhibit IgA (coated to yellow/green-fluorescent latex beads)- mediated phagocytosis by surface CD89-expressing primary human neutrophils and monocytes was determined.

IgA coated and fluorescently-labeled latex beads were pre-incubated with a dilution series of anti-lgA Fabs and subsequently added to isolated CD89-expressing polymorphonuclear leukocytes (PMN) or monocytes. As a positive control, a commercial mouse anti-hCD89 blocking antibody clone Mip8A (BioRad) was used and as negative control, beads coated with BSA instead of IgA were added. Finally, with flow cytometry, the phagocytic index was measured i.e. the percentage of cells that phagocytosed, multiplied by the geometric mean of fluorescent cells.

The commercial CD89 blocking antibody Mip8A showed the expected blocking and the signal observed with BSA-coated beads was taken as a background reading. Overall the results showed that the Fabs of variants gClone A E HIS and gClone C TQ block IgA- mediated phagocytosis similarly in a dose-dependent manner. All Fabs of the gClone B HIS variants had a lower blocking capacity and only blocked IgA-mediated phagocytosis at the highest concentrations on PMNs (Table 13) and monocytes (Table 14). This is in line with the reduced competing and displacing properties observed in the ELISA and RBL cellbased assay for the gClone B HIS variants.

Table 13: Percentages (%) of remaining IgA-mediated phagocytosis on primary CD89 expressing PMNs after blocking with anti-lgA antibodies. Values were normalized to the PMN + IgA condition. Table 14: Percentages (%) of remaining IgA-mediated phagocytosis on primary CD89 expressing monocytes after blocking with anti-lgA antibodies. Values were normalized to the monocytes + IgA condition.

3.3 FcRn degradation assay with anti-lgA mAbs using FcRn overexpressinq HEK cells

FcRn degradation was assessed in the presence of anti-lgA mAbs in complex with their target human IgA.

HEK hFcRn-GFP cells, overexpressing FcRn coupled to a GFP tag, were seeded in growth medium and incubated overnight at 37°C. After washing with treatment buffer, 500 nM of anti-lgA mAbs alone or in complex with recombinant human Mota-IgA were added to the cells and allowed to incubate at 37°C. All tested anti-lgA antibodies shared the same Fc ‘LALA ABDEG hlgGT to allow for FcRn-mediated internalization. After washing and viability staining, data acquisition was performed by flow cytometry, using the BD LSRFortessa™ X-20 Cell Analyzer.

The positive control (anti-FcRn antibody) and the negative control (untreated cells) showed the expected results (Figure 1). FcRn degradation was observed for all of the tested antiIgA mAbs when added in immune complex with IgA (Figure 1). However, a difference in the level of FcRn degradation was observed between the different group of variants.

All gClone A based variants (gClone A E, gClone A E HIS and gClone A EK HIS) in complex with IgA demonstrated marked FcRn degradation, with an average reduction of 66%. The two gClone C variants (gClone C TQ and gClone C TA) showed an average reduction in FcRn levels of 38%. In contrast, the three gClone B HIS variants (gClone B HIS H, gClone B HIS L and gClone B HIS A) induced much less FcRn degradation when present in an immune complex with IgA, with an average reduction of 20%.

3.4 Off-target binding to immunoglobulins IqE, IqM & IqGs on SPR

Further SPR experiments were performed to investigate whether any of the anti-lgA antibodies exhibited off-target binding to other immunoglobulins.

A CM5 chip was coated with the dedicated immunoglobulins: Motavizumab-lgE (in-house production), Motavizumab-lgG1 (in-house production), IgM (Sigma), Motavizumab-lgG4 (in-house production), IVIG (Privigen, contains 69% lgG1 , 26% lgG2, 3% lgG3 and 2% lgG4) and Motavizumab-lgA (in-house production) as a reference. As IVIG could contain traces of IgA, it was depleted for IgA using IgA capture select beads (Thermo scientific). Afterwards, every anti-lgA antibody was injected for 120s at pH 7.4, followed by dissociation at pH 7.4.

The experiments confirmed that for all variants tested (gClone A E, gClone A EK, gClone A E HIS, gClone C TA, gClone C TQ, gClone B HIS L), there was no off-target binding to non-lgA immunoglobulins (specifically IgE, IgG, IgM or IVIG depleted for IgA).

3.5 Epitope mapping of variants

3.5.1 SPR experiments

To evaluate if the epitopes of the seven anti-lgA clones overlap, an epitope mapping experiment was performed using SPR.

For this experiment, a CM5 chip was coated with human Mota-IgA (in-house production) and two different Fabs of the anti-lgA clones were consecutively injected for 120s at pH 7.4, followed by dissociation at pH 7.4.

When gClone A E HIS is bound to human IgA, all variants derived from the parental clone D (gClone C TA, gClone C TQ, gClone B HIS H, gClone B HIS L, gClone B HIS A) could bind to IgA as well and the other way around. In contrast, no binding was observed for any of the clone D variants when another clone D variant was already pre-bound to human IgA. These data suggest that gClone A binds to a different epitope than the different clone D variants. Minor differences in the response units (RU) reached were observed when the gClone A E HIS clone was tested in the presence of different clone D variants - lower values were obtained in the presence of gClone B HIS H and g Clone B HIS L Fabs. This finding suggests that there could be a minor difference in epitopes or the conformational binding between the different variants derived from clone D.

3.5.2 Immune complex analysis with negative stain electron microscopy lgA:anti-lgA immune complexes were analysed with negative stain electron microscopy. For this analysis two approaches were used to generate and evaluate immune complex formation.

In the first approach, gClone A E HIS and gClone C TA variants were incubated with human Mota-IgA, after which size exclusion chromatography (Akta, Superdex 200 10/300 GL) was performed to purify immune complexes. In the second approach, both compounds (anti-lgA variant and human Mota-IgA) were incubated for 1 minute directly on a grid for negative stain electron microscopy.

Both approaches for negative stain microscopy had similar results. Variant gClone A E HIS was observed to form wire-like shaped immune complexes with IgA (Figure 2A). Circular immune complexes were observed for gClone C TA visualized as small clumps in negative stain microscopy (Figure 2B). Whilst no data was generated with the gClone B HIS variants, given that the variants comprise the same heavy chain as the gClone C TA variant, it is expected that gClone B HIS variants would form immune complexes that are similar to those observed for gClone C TA.

Overall, these experiments indicated that all variants originating from the parental clone D (gClone C TA, gClone C TQ, gClone B HIS A, gClone B HIS H, gClone B HIS L) generally bind a similar epitope of IgA, whilst variant gClone A E HIS binds a different epitope. The distinct manner of binding by the two anti-lgA mAbs leads to the formation of different immune complexes.

3.6 In vivo human IgA sweeping in mice

3.6.1 IgA sweeping in (hlgA) a1 KI mice

In order to explore the IgA sweeping efficacy of the anti-lgA antibodies in vivo, the anti-lgA clones were tested over several in vivo experiments in a1KI mice, which transgenically express endogenous human IgA by knocking in the human Co Ig gene in place of the Sp region (Duchez et al. 2010).

In the first study the variants gClone A and gClone A HIS hlgG 1 WT (i.e. both still containing D98) were compared and Mota hlgG 1 WT was included as negative control. As expected, no IgA sweeping was observed with the negative control, Mota hlgG 1 WT.

For both of the anti-lgA antibodies, initially a steady PK profile was observed but from day 7 onwards there was faster clearance of the clone gClone A h IgG 1 WT as compared to the histidine engineered variant (Figure 3A). The PD data (represented as % of hlgA normalized to baseline values) showed that both clones were able to clear endogenous human IgA from the circulation but that there was no real difference in the sweeping efficacy of both clones. Also, the duration of IgA sweeping was relatively prolonged, and IgA levels were still reduced by 75% 14 days post anti-lgA injection (Figure 3B).

In a follow-up study, the variants gClone A E and gClone A E HIS were compared with variants gClone C TA and gClone C TQ. The Fc regions of the tested anti-lgA antibodies were also modified to contain LALA HN substitutions (LALA: L234A, L235A; HN: H433K/N434F) instead of a WT human lgG1 (hlgG1 ) Fc region.

The PK data showed fast clearance of the anti-lgA mAbs from the circulation due to the difference in Fc region backbone. This was expected since HN mutations (H433K/N434F) are known to induce increased affinity for mouse FcRn at physiological and acidic pH as compared with the WT hlgG 1 Fc region. No difference in PK profile was observed for the different anti-lgA antibodies, and the negative control, Mota hlgG 1 LALA HN, demonstrated a longer lasting PK profile due to the absence of IgA binding (Figure 4A).

The PD data showed that all tested anti-lgA variants were able to clear endogenous human IgA from circulation and remove more than 80% of IgA within 24 hours post anti-lgA injection. Although the differences were minor, a deeper reduction in IgA levels was observed for gClone C TA and gClone A E. Afterwards, IgA levels increased again, but it was 5-7 days before IgA levels were above 50% of the baseline values (Figure 4B).

The sweeping efficacy of variant gClone A EK HIS was also compared to variant gClone B HIS H in a LALA HN hlgG1 Fc region backbone in o1 KI mice. Similar to previous experiments, the antibodies showed a very fast and similar o-phase but a slightly reduced clearance was observed for gClone B HIS H from day 3 onwards. Nonetheless, both antiIgA antibodies could no longer be detected 7 days post injection (Figure 5A).

Comparing the IgA sweeping efficacy, more than 80% of the IgA levels were reduced within 24 hours post anti-lgA antibody injection for both variants. However, a deeper reduction in IgA was observed with the gClone B HIS H variant. IgA levels did eventually increase but it took more than 7 days for IgA levels to reach 50% of the baseline levels. The recovery kinetics of IgA for the two anti-lgA test mAbs were similar (Figure 5B).

3.6.2 Human IgA sweeping in hFcRn/HSA mice

Several studies were performed in Albumus™ mice (which express human FcRn and albumin) to continue to evaluate the human IgA sweeping efficacy of the anti-lgA mAbs in vivo.

In the first study, 5 of the 7 anti-lgA antibodies were tested (gClone C TA, gClone C TQ, gClone B HIS L, gClone A E and gClone A E HIS) and all shared the same Fc format ‘LALA ABDEG hlgGT, to facilitate FcRn-mediated sweeping.

All anti-lgA antibodies demonstrated a similar PK profile over time until day 4 post-mAb administration, except gClone B HIS L, which showed reduced clearance from the circulation during the first days (Figure 6A). From D7 onwards, gClone C TA and gClone C TQ antibodies diverged from the other 3 anti-lgA antibodies, showing an accelerated - phase and faster clearance from the circulation (Figure 6A).

A clear difference could be observed in the IgA sweeping efficacy of the tested anti-lgA antibodies. After the first human Mota-IgA injection, gClone B HIS L showed the highest level of IgA sweeping, followed by gClone C TA and gClone A E HIS (Figure 6B). Upon reinjection of human Mota-IgA at day 2 (48 hours after the first injection), a similar ranking of the different antibodies was observed and gClone B HIS L was still the highest performing IgA sweeping antibody (Figure 6B). In general, the IgA sweeping efficacy seemed to correlate with the pH dependency/affinity properties of these clones. Based on all preceding data regarding the IgA sweeping capacity (PK-PD), the affinity for human and cynomolgus monkey IgA and the potential rebinding at acidic pH (cfr. endosomal pH), the variants gClone C TA and gClone C TQ were not further tested.

Given the high level of sweeping efficacy shown by gClone B HIS L, an additional in vivo study was performed to evaluate if gClone B HIS H, which showed the highest pH- dependent antigen binding of all 3 gClone B HIS variants, would further improve the sweeping efficacy. Variant gClone A EK HIS was also evaluated to see if the S53K substitution would affect the IgA sweeping efficacy. A similar experimental design as described above in Albumus™ mice was used and all test mAbs shared the same Fc format ‘LALA ABDEG hlgGT, to facilitate FcRn-mediated sweeping.

The variants gClone B HIS L and gClone B HIS H had a similar PK profile over time, whereas a faster clearance was observed for gClone A EK HIS (Figure 7A). Similar to the previous experiment, less efficient IgA sweeping was observed for variant gClone A EK HIS as compared to the gClone B HIS variants. In addition, gClone B HIS H showed a faster removal of circulating IgA compared to gClone B HIS L (Figure 7B).

3.7 Anti-IqA-mediated internalization of IgA in HEK cells overexpressinq FcRn Internalization of human IgA in the presence of the anti-lgA antibodies was assessed in HEK-FcRn cells, which overexpress FcRn at the surface.

Cells were seeded in growth medium in a V-bottom 96 well plate and 300 nM of either antiIgA mAbs alone or a pre-incubation mix of anti-lgA antibodies and a pH-sensitive labeled human Mota-IgA (in-house production, labeling was done with pHrodo iFLGreen Microscale Protein Labeling Kit according to manufacturer’s instructions) was added to the cells. All tested anti-lgA antibodies shared the same Fc ‘LALA ABDEG hlgGT to allow for FcRn-mediated internalization. After washing with FACS buffer and viability staining (Fixable Viability Dye eFluor 780, eBioscience), data acquisition was performed by flow cytometry, using the BD LSRFortessa™ X-20 Cell Analyzer.

As expected, internalization was driven by FcRn since no internalization was observed in HEK WT cells. All variants showed similar internalization levels of human Mota-IgA (Figure 8). 3.8 Cynomolqus monkey IgA cross-reactivity

3.8.1 Competition of anti-lqA mAbs for human and cynomolqus monkey lqA:CD89 interaction on MSP

As described above, all anti-lgA variants were able to compete and displace human serum IgA from its receptor hCD89. Based on the displacement EC50 values, the highest level of blocking was observed for gClone A E and gClone A E HIS, followed by the gClone C variants and gClone B HIS L, respectively (Table 11).

In order to evaluate cynomolgus monkey cross-reactivity of the lgA:CD89 blocking capacity of all variants, an MSD ELISA was developed.

After blocking and washing, the MSD GOLD 96-well Streptavidin SECTOR plate (MSD) was coated with BIOTIN-Human CD89 (in-house tagged) or BIOTIN-Cynomolgus monkey CD89 (in house tagged) and incubated at room temperature. After washing the plate, a dilution series of anti-lgA antibodies, pre-incubated with a fixed concentration of human IgA (MP Biomedicals) or a fixed concentration of cynomolgus monkey IgA (Life diagnostics) was added to the plate and allowed to bind at room temperature. After washing the plate, IgA was detected with SULFO-goat anti-human serum IgA achain specific F(ab')2 (in house tagged) and developed with Read buffer 2x (MSD) and measured using a MSD Quickplex reader (MSD). lgA:CD89 displacement was analysed with a similar set-up, but IgA was incubated 1 hour before the addition of the dilution series of anti-lgA antibodies, with a washing step in between.

In accordance with the affinity data, both variants competed for binding of human or cynomolgus monkey IgA to their respective receptors. Again, better competitive binding was observed for gClone A EK HIS as illustrated by the lower EC50 values for both IgA species (Table 15) and the same was observed in the displacement analysis (Figure 9 and Table 15).

Table 15: EC50 values corresponding to the MSD ELISA experiment

By comparing the lgA:CD89 competing properties among the two species, both clones were considered cynomolgus monkey cross-reactive as the EC50 ratio (Cyno EC50/Human EC50) was below 10 (Table 15).

3.8.2 Sweeping of human and cynomolgus monkey IgA in Albumus™ mice

To confirm the cynomolgus monkey IgA cross-reactivity data generated by SPR, an in vivo study was designed comparing the human and cynomolgus monkey serum IgA sweeping efficacy of gClone B HIS L and gClone A E HIS in Albumus™ mice. These experiments used either the hlgG1 LALA ABDEG Fc variant (Figure 10 and 11) or hlgG1 LALAPG ABDEG Fc variant (Figure 12).

All anti-lgA antibodies demonstrated a comparable and steady PK profile and no difference was observed when injecting human or cynomolgus monkey serum IgA prior to injection of the test item. Both variants, gClone A E HIS hlgG 1 LALA ABDEG and gClone B HIS L hlgG 1 LALA ABDEG, were cross-reactive to cynomolgus monkey IgA and induced efficacious IgA sweeping in Albumus™ mice. Sweeping of human IgA with gClone B HIS L hlgG 1 LALA ABDEG was slightly faster than sweeping of cynomolgus monkey IgA; the opposite was true for gClone A E HIS hlgG 1 LALA ABDEG, which demonstrated faster clearance of cynomolgus monkey IgA as compared to human IgA. Comparing the two antibodies, better human IgA sweeping was observed for gClone B HIS L hlgG1 LALA ABDEG (in-line with the experiments described above), while similar sweeping was observed for cynomolgus monkey IgA (Figure 10 and 11).

Cynomolgus cross-reactivity was confirmed for gClone B HIS H hlgG 1 LALAPG ABDEG in vivo in Albumus™ mice using the same experimental design described above (Figure 12). No difference was observed in the PK profile when mice were injected with either human serum IgA or cynomolgus monkey serum IgA. When comparing human and cynomolgus monkey serum IgA sweeping efficacy (PD), faster clearance was observed for human IgA, both after the first and second injection of IgA. gClone B HIS H hlgG 1 LALAPG ABDEG removed 81 % of cynomolgus monkey IgA and 97% of human IgA within 24 hours post antiIgA injection and upon reinjection of IgA, similar values were obtained (74% of cynomolgus monkey IgA and 96% of human IgA), confirming that gClone B HIS H hlgG 1 LALAPG ABDEG is cross-reactive with cynomolgus monkey IgA (Figure 12).

3.9 FcRn occupancy on human and cynomolgus monkey PBMCs

Anti-lgA antibody variants_gClone A HIS hlgG 1 LALA ABDEG and gClone B HIS H hlgG 1 LALAPG ABDEG were tested for their ability to bind and occupy FcRn on human and cynomolgus monkey PBMCs. The antibodies were tested alone or complexed with IgA (in a molar 1 :3 ratio). The results are shown in Figure 13 and show that upon binding of IgA reduced FcRn occupancy is observed compared to the occupancy of the antibodies alone but this difference is similar for both variants.

3.10 FcRn degradation

In similar experiments to those described in section 3.3, anti-lgA antibody variants gClone A HIS hlgG 1 LALA ABDEG and gClone B HIS H h IgG 1 LALAPG ABDEG were tested for degradation of FcRn in FcRn-overexpressing HEK cells. The results are shown in Figure 14. Similar to the results observed previously, gClone B HIS H hlgG1 LALAPG ABDEG (TA) induced less FcRn degradation as compared with gClone A HIS hlgG1 LALA ABDEG (TA) when both anti-lgA antibodies were complexed with IgA (in a molar 1 :3 ratio).

3.11 Summary

Based on the data described above, the variant gClone B HIS H” was selected as the source of a Fab for further development of an anti-lgA molecule.

This Fab showed the highest level of pH-dependent antigen binding in vitro to both human and cynomolgus monkey IgA and the lowest level of IgA binding at an acidic endosomal pH (i.e. pH 6) according to SPR. The latter property is especially important since it means that re-binding of the antibody to its target antigen when it is released into the endosome would be minimised, thereby allowing more efficacious IgA sweeping and antibody recycling.

In vivo, variant gClone B HIS H with a hlgG 1 comprising ABDEG substitutions demonstrated the highest level of IgA sweeping. Notably the IgA sweeping occurred very quickly but the variant was also capable of removing IgA after re-injection of human mota- IgA in Albumus™ mice (i.e. durable sweeping). This is in line with the PK of the antibody, which showed a better and steady PK profile compared to the other tested variants. The variant gClone B HIS H was cross-reactive to cynomolgus monkey IgA (confirmed both in vitro on ELISA/SPR and in vivo by cynomolgus IgA sweeping in Albumus™ mice) and was also capable of blocking the interaction between human or cynomolgus monkey IgA and CD89. Importantly, gClone B HIS H caused only low levels of FcRn degradation, which minimises the risk of lowering albumin levels in vivo. Moreover, this reduces the likelihood of the anti-lgA antibody detrimentally affecting the functionality of the FcRn receptor in vivo.

Example 4: gClone B HIS H binds to all in vivo molecular forms of IgA

This example demonstrates that variant gClone B HIS H is able to bind all molecular forms of IgA, including forms of IgA that are more prevalent in patients with IgA-mediated diseases.

4.1 Binding to galactose-deficient lqA1 (Gd-lqA1)

The selected Fab (gClone B HIS H) was analysed for the ability to bind IgA from IgA Nephropathy (IgAN) patients. IgAN is the most common form of primary glomerulonephritis worldwide and is defined by mesangial IgA deposition, with consequent mesangial cell proliferation, inflammation, and tubulointerstitial fibrosis. In IgAN patients, there is an increase in circulating lgA1 that is polymeric and lacks terminal galactose residues in its hinge region (termed galactose-deficient lgA1 of Gd-lgA1 ) (Scionti et al. 2022 (Figure 15A)).

In order to analyse if gClone B HIS H could bind to galactose-deficient lgA1 (which lacks the galactose residue that normally is attached to the GalNAc residue in the O-glycan chain) a lectin binding ELISA with IgA purified (using Jacalin purification) from serum of IgAN patients was performed.

Nunc immunoplates were coated with rabbit anti-human IgA (Dako), the variant gClone B HIS H hlgG1 LALAPG ABDEG or an isotype control (Mota hlgG1 LALAPG ABDEG) overnight at 4°C. After washing and blocking the plate, diluted serum samples (1 :100) were added to the plate and allowed to bind overnight at 4°C. The next day, the plate was washed and treated with Neuraminidase (New England BioLabs) to cleave terminal sialic acid. After washing, biotinylated HPA (Helix pomatia agglutinin; Sigma) was incubated overnight at 4°C. As HPA can only bind terminal GalNAc residues (that are not shielded by galactose residues), only galactose-deficient IgA will be bound. Finally, the plate was washed and N-acetylgalactosamine (GalNAc) binding was detected with streptavidin-HRP (R&D) and OPD solution (Thermo) and reaction was stopped with 2,5M sulphuric acid. OD was measured at 492 nm.

The results demonstrated that the gClone B HIS H h IgG 1 LALAPG ABDEG variant is capable of binding to deglycosylated lgA1 in serum of IgAN patients since HPA is able to bind terminal GalNAc residues in the hinge region of lgA1 (Figure 15B). A slightly higher signal was observed for the positive control (rabbit anti-human IgA), but given that this was a polyclonal antibody, a direct comparison could not be made to the gClone B HIS H hlgG 1 LALAPG ABDEG variant. As expected, the isotype control, Mota hlgG 1 LALAPG ABDEG, did not bind galactose-deficient lgA1 in serum of IgAN patients (Figure 15B).

4.2 Binding to deposited IgA immune complexes in the kidney

To evaluate if variant gClone B HIS H can also bind to IgA immune complex deposits in the kidney of IgAN patients, immunofluorescent kidney staining was performed.

A frozen kidney biopsy from an IgAN patient was stained with gClone B HIS H mlgG1 FcD (N297A), followed by detection with the secondary anti-mouse AF568. DAPI was also used to counterstain the nuclei.

Ig A1 immune complex deposition typically occurs in the glomerular mesangium of the kidney of IgAN patients (Lai et al. 2016), and indeed staining with gClone B HIS H mlgG1 FcD resulted in a fluorescent signal which coincided with the glomeruli of the kidney (Figure 16). This data indicates that gClone B HIS H can bind to lgA1 immune complex deposits in the kidneys of patients with IgAN.

4.3 Binding to different molecular IgA formats

To analyse if gClone B HIS H can bind to different molecular forms of IgA or if it selectively binds to any particular form of IgA, a Western blot experiment was performed with different samples. With the exception of secretory IgA (Sigma), all IgA samples were isolated from serum using Jacalin affinity chromatography. Serum from four healthy subjects (HS) and four patients with IgA nephropathy (IgAN) were used and further subdivided based on their lectin binding properties (using the lectin binding ELISA) into High Lectin Binding (HLB) or Low Lectin Binding (LLB) samples.

The IgA samples were run on an SDS-PAGE (4-20% Mini Protean TGX precast gels, BioRad) under non-reducing conditions to be able to separate monomeric and polymeric IgA so that binding to different molecular forms of IgA could be assessed. After blotting, gClone B HIS H mlgG1 FcD (N297A) was added and detected with rabbit anti-mouse immunoglobulins-HRP (DAKO). As a control, a commercial anti-lgA (Polyclonal rabbit antihuman IgA heavy chain-HRP antibody, DAKO) was used.

The resulting Western blots showed that the binding pattern of gClone B HIS H mlgG1 FcD (N297A) did not differ to the pattern observed with the commercial polyclonal anti-lgA antibody on the same serum samples. This indicates that the gClone B HIS H mlgG1 FcD (N297A) variant can bind both lower and higher molecular weight forms of IgA (monomeric - 160 kDa - and polymeric/immune complexes which are > 350 kDa) and does not show any preferential binding to a particular form of IgA. There was also no differential binding by the test antibody to IgA isolated from either healthy subjects or IgAN patients, or between lgA1 with low (HLB) or high levels of galactose (LLB) present in the hinge region. In addition, gClone B HIS H mlgG1 FcD (N297A) was able to bind slgA similarly to the positive control (Figure 17).

4.4 Summary

The variant gClone B HIS H was able to bind all in vivo molecular forms of IgA, including galactose-deficient lgA1. The galactose-deficient type of lgA1 forms pathogenic immune complexes that are found in the kidneys of IgAN patients.

Example 5: Comparison between one-armed (OA) and two-armed (TA) anti-lgA antibody formats

This example demonstrates that one-armed (OA) modified antibodies with variant Fc regions comprising LALAPG ABDEG™ and knob-into-hole substitutions exhibited improved IgA sweeping as compared to conventional two-armed (TA) antibodies. 5.1 Screening of Fc variants with the qClone B HIS H Fab

The gClone B HIS H Fab was produced in different hlgG 1 Fc formats as shown in Figure 18.

It has been observed in an in vitro assay that LALA mutations (L234A/ L235A) do not completely abrogate FcyR binding and may induce unwanted Fc effector functions. An additional PG mutation (P329G) in addition to the LALA mutations in the Fc domain has been reported to almost completely block residual binding to activating FcyR (Schlothauer et al. 2016). Therefore, gClone B HIS H was produced in OA and TA formats including LALAPG mutations (L234A/L235A/P329G) to completely abolish Fc effector functions. The theoretical pl of the different formats are listed in Table 16.

Table 16: Theoretical pl of the different Fc formats of gClone B HIS H

The OA format was generated using knob-into-hole technology (Knob VH-T366W and Hole VH-T366S/L368A/Y407V) (Atwell et al. 1997).

The OA modified antibodies with variant Fc regions comprising LALAPG ABDEG™ and knob-into-hole substitutions as well as a gClone B HIS H Fab are referred to herein as gClone B HIS H LALAPG ABDEG OA” or LALAPG ABDEG OA”.

5.2 Human and cynomolqus monkey IgA affinity of OA and TA format of qClone B HIS H LALAPG ABDEG

To compare the affinity of the OA and TA format of gClone B HIS H LALAPG ABDEG for human and cynomolgus monkey serum-purified IgA, screening was performed on the Biacore™ T200 SPR system to analyse kinetics (association rate - ka; dissociation rate - kd; affinity - KD) at pH 7.4 and pH 6.

In more detail, affinity was determined on a CM5 chip coated with anti-human Fc IgG (Jackson ImmunoResearch) followed by capturing the OA and TA format of gClone B HIS H LALAPG ABDEG and association of serum human IgA (MP Biomedicals) or serum cynomolgus monkey IgA (Life diagnostics) at pH 7.4 or pH 6 and dissociation at pH 7.4 or pH 6.

Using this set-up, no major differences were observed in the kinetics (ka - OA: 3.49E+05 M'¹s'¹; TA: 3.56E+05 M-¹s’¹, kd - OA: 1 .62E-03 s; TA: 1 .69E-03 s) and affinity (KD - OA: 4.66 nM; TA: 4.76 nM) measurements between the two formats for human IgA binding at pH 7.4.

For cynomolgus monkey IgA, also no major differences were observed in the kinetics and affinity at pH 7.4 (ka - OA: 6.87E+05 M-¹s’¹;TA: 7.42E+05 M-¹s’¹, kd - OA: 2,05E-02 s; TA: 1.80E-02 s, KD - OA: 27.3 nM; TA 24,3 nM). Both formats are cross-reactive to cynomolgus monkey IgA (ratio KD cyno/human IgA for OA is 5,85 and for TA 5,11 ) (Table 17).

To evaluate the binding in an acidic environment such as the endosome, affinity measurements were performed at pH 6. Both OA and TA variants showed no detectable binding to cynomolgus monkey IgA at pH 6 with the used concentration range, but showed some residual binding to human IgA, with reduced kinetics and affinity compared to binding at physiological pH (pH 7.4) (ka - OA: 3.43E+05 M-1 s-1 ; TA: 2.09E+05M-1 s-1 , kd - OA: 1.12E-01 s; TA: 1.08E-01 s , KD - OA: 335 nM; TA: 534 nM) (Table 18). For the gClone B HIS H LALAPG ABDEG TA and OA formats, no fitting could be obtained for serum derived cynomolgus monkey (cyno) IgA due to lack of binding at pH 6.

Table 17: Kinetics and cross-reactivity of gClone B HIS H LALAPG ABDEG TA and OA formats to serum-derived human and cynomolgus monkey (cyno) IgA at pH 7.4 using the 1 :1 binding fitting model on the Biacore™ Insight software.

SD: standard deviation.

Table 18: Kinetics of gClone B HIS H LA LA PG ABDEG TA and OA formats to serum- derived human and cynomolgus monkey (cyno) IgA at pH 6 using the 1 :1 binding fitting model on the Biacore™ Insight software.

5.3 lqA:CD89 blocking capacity of gClone B HIS H Fc variants

The gClone B HIS H LALAPG ABDEG TA and OA Fc variants were tested for cynomolgus monkey cross reactivity. In order to test for lgA:CD89 blocking capacity, a competition and displacement MSD assay was performed for human and cynomolgus monkey IgA to their respective CD89 receptors.

Both OA and TA formats competed with binding of human or cynomolgus monkey IgA to their respective CD89 receptors. For both species, better blocking was observed for the TA format (Figure 19). A difference was observed in lgA:CD89 blocking capacity between human and cynomolgus monkey (Figure 19).

Table 19: EC50 values corresponding to the MSD ELISA experiment By comparing the lgA:CD89 competing properties among the two species, both OA and TA formats were considered cynomolgus monkey cross-reactive as the EC50 ratio (Cyno EC50/Human EC50) was below 10 (Table 19).

5.4 Complex analysis using mass photometry with the OA and TA format of qClone B HIS H LALAPG ABDEG

To better understand the complexes formed between IgA and the anti-lgA mAbs, mass photometry was performed. As OA formats only have one Fab arm, it was expected that the size of these complexes would be limited to trimeric lgA:anti-lgA complexes (two antiIgA’s binding to one IgA). In contrast, TA formats could theoretically form much larger complexes as each anti-lgA construct could bind two individual IgA molecules.

For this analysis, human Mota-IgA was added to the gClone B HIS H LALAPG ABDEG TA or OA variants in a 1 :1 molar ratio but only limited complex formation was observed in the output (Figure 20A). This is likely due to dissociation of the lgA:anti-lgA complexes as mass photometry requires a high dilution of the sample to achieve good resolution (concentrations of 10-20 nM). To prevent dissociation of lgA:anti-lgA complexes, DSSO (disuccinimidyl sulfoxide) was added to induce chemical crosslinking of the formed complexes and a higher variety and number of complexes was observed. As expected, for the OA format only two types of complexes were observed (1 :1 and 2:1), with the most prevalent being the 1 :1 complex (35.4% vs 12.5% respectively). For the TA format, the most prevalent complexes were 2:2 lgA:anti-lgA complexes (14.8%), but both smaller (1 :1 and 1 :2) and larger complexes (2:3, 3:3, 3:4 and 4:4) were also observed with varying frequency. It was also found that 4:4 complexes were the largest complexes observed by this method, suggesting that there is a limit in the size of lgA:anti-lgA complexes being formed (Figure 20B).

5.5 Characterisation of FcRn-mediated cellular uptake of qClone B HIS H variants

To characterize the FcRn binding of the gClone B HIS H Fc constructs, an IBIS’ MX96 SPR imager system was used to determine the affinity of the Fc variants to human FcRn at pH 7.4 (physiological pH) and pH 6 (endosomal pH) (Figure 21). As expected, no measurable FcRn binding (reported as >2000 nM) at pH 7.4 was observed for Fc variants that didn’t contain mutations enhancing the binding to FcRn and only minor binding was observed at pH 6. Comparing the FcRn affinity of the OA and TA LALAPG ABDEG formats showed that affinities were in the same range at both pHs. To further characterize the FcRn dependency of the different gClone B HIS H Fc constructs for IgA sweeping, an FcRn occupancy assay was performed using U937 cells, which endogenously express FcRn. Better occupancy of the OA format was observed, indicating that there is less steric hindrance when the OA format binds to hFcRn (Table 20).

Table 20: EC50 values (in nM) of the FcRn occupancy of the different gClone B HIS H OA and TA Fc variants on U937 cells.

EC50 values were determined using four parameter fitting in GraphPad.

FcRn occupancy of gClone B HIS H OA and TA Fc variants was also assessed on human and cynomolgus monkey PBMCs. The results are shown in Figure 22 and also show better occupancy of the OA format compared to the TA Fc variant. For both antibodies, reduced FcRn occupancy is observed when the antibodies are complexed with IgA.

To also evaluate FcRn-mediated internalisation of the different gClone B HIS H constructs, a FACS internalisation assay using HEK-FcRn and HEK-WT cells was performed. FcRn- dependent IgA internalization could be observed for all ABDEG containing gClone B HIS H variants. A higher IgA internalisation signal was seen with the LALAPG ABDEG TA format compared to the OA variant (Figure 23). As explained elsewhere, it is believed that this is because an OA format can only form dimeric or trimeric complexes, whilst a TA format can form larger complexes. Internalisation of larger complexes will result in a higher MFI signal due to the presence of a greater number of IgA-labelled molecules. In addition, no IgA internalization was observed for any of the variants in the control HEK-WT cells, which indicated that internalization of IgA was mediated by FcRn (Figure 23).

Finally, FcRn degradation was evaluated for the different gClone B HIS H constructs using two different cell lines, HEK-FcRn-GFP and U937 cells. The HEK-FcRn-GFP cell line overexpresses FcRn and the monocytic cell line U937 expresses endogenous levels of FcRn. In general, little FcRn degradation was observed with gClone B HIS H constructs in both cell lines (Figure 24). When evaluating FcRn degradation in HEK-FcRn-GFP cells, minor degradation was observed for the TA antibody with LALAPG ABDEG mutations (8.8%) when in complex with human Mota-IgA. Surprisingly, no FcRn degradation was observed with gClone B HIS H LALAPG ABDEG OA in complex with IgA (Figure 24).

Whilst minor FcRn degradation was observed in HEK-FcRn-GFP cells with the LALAPG ABDEG TA format, no degradation was observed in U937 cells with the same construct. A possible explanation for these results could be due to the difference in FcRn expression levels between both cell types. HEK-FcRn-GFP cells overexpress FcRn, while FcRn expression in U937 cells is endogenous and expressed at much lower levels. The positive control used in this assay, behaved as expected and induced the highest amount of FcRn degradation, both on HEK-FcRn-GFP (87.5%) as on U937 (50.2%) cells (Figure 24).

5.6 In vivo IgA sweeping efficacy of gClone B HIS H variants using Albumus™ mice In order to translate the in vitro findings and better understand the sweeping efficacy of the different antibody formats, an in vivo study was performed in Albumus™ mice comparing gClone B HIS H LALAPG ABDEG OA and LALAPG ABDEG TA for both human and cynomolgus monkey serum-derived IgA sweeping efficacy. The results are depicted in Figure 25.

As expected, no difference was observed in the PK profile of the same test item when mice were injected with either human serum IgA or cynomolgus monkey serum IgA. In contrast, there was a clear difference in the PK profile for the different antibody formats - gClone B HIS H LALAPG ABDEG OA showed a faster removal from the circulation compared to its TA counterpart (Figure 25C).

When evaluating the human IgA sweeping capacity of the two formats, it was clear that the OA format demonstrated much faster IgA sweeping after the first injection of human serum IgA. Also after re-injection of human serum IgA on day 2, faster clearance was observed with the OA format. For the groups injected with cynomolgus monkey serum IgA, a similar difference was observed, showing enhanced IgA sweeping with gClone B HIS H LALA PG ABDEG OA for both IgA injections (on day 0 and day 2). In addition, comparing human versus cynomolgus monkey IgA sweeping efficacy of each test mAb, faster clearance was observed for human IgA for both compounds. For the OA format, the differences in IgA sweeping efficacy were observed shortly after IgA injection, while for the TA format differences were mainly observed 24 hours post IgA injection (Figure 25A and 25B).

5.7 B cell binding and activation

The ability of the anti-lgA variants: gClone B HIS H LALAPG ABDEG OA; and gClone B HIS H LALAPG ABDEG TA, to bind to the IgA B cell receptor (BCR) of CD19+ CD27+ memory B cells and activate the IgA BCR (measured via phosphorylation of the downstream signalling proteins PLCy2 and Syk) was assessed by flow cytometry. The results are shown in Figure 26 and show binding to the IgA BCR for both gClone B HIS H LALAPG ABDEG OA and gClone B HIS H LALAPG ABDEG TA, with a lower signal for the OA variant. Whilst gClone B HIS H LALAPG ABDEG TA shows a consistent and dosedependent activation of the IgA BCR, no concentration-dependent activation of the IgA BCR was observed for gClone B HIS H LALAPG ABDEG OA.

5.8 Binding to human and cynomolqus monkey Fc gamma receptors

The ability of anti-lgA gClone B HIS H LALAPG ABDEG OA to bind to different human and cynomolgus Fc gamma receptors was measured. This binding was compared to a corresponding OA construct lacking the “LALAPG” mutations and a control wild-type IgG 1 TA antibody construct. The results are shown in Figure 27. These results confirm that the LALAPG mutations completely abrogate the binding of the OA construct to human and cynomolgus monkey Fc gamma receptors.

5.9 Binding to human and cynomolqus monkey C1q

The inability of anti-lgA gClone B HIS H LALAPG ABDEG OA to bind to human and cynomolgus monkey C1q was assessed by ELISA. The binding activity was compared with the binding activity of a control wild-type IgG 1 TA antibody construct. The results are shown in Figure 28 and show reduced binding to human and cynomolgus monkey C1q with gClone B HIS H LALAPG ABDEG OA.

5.10 Summary

This Example describes the characterisation and IgA sweeping efficacy of different anti-lgA antibodies having one-armed (OA) and two-armed (TA) formats. The different anti-lgA antibodies were tested for IgA affinity, lgA:CD89 blocking capacity, FcRn-dependent IgA sweeping, and cynomolgus monkey cross-reactivity. When comparing the OA and TA formats of an anti-lgA antibody variant having the LALAPG ABDEG IgG 1 Fc region, it was clear that the OA format improved IgA sweeping efficacy. Further, it was found that gClone B HIS H Fab in the LALAPG ABDEG h IgG 1 backbone, when complexed with IgA, did not lead to FcRn degradation in vitro.

Taking the above into account, gClone B HIS H LALAPG ABDEG OA hlgG1 was selected for further testing as described in Example 6.

5.11 Materials and methods

5.11.1 CD89 competition flow cytometry assay using RBL-CD89-GFP cells

In this assay, RBL-WT cells or RBL cells stably transfected with human CD89-GFP were incubated with immune complexes of RSV-F (Respiratory syncytial virus fusion glycoprotein, in-house production) and human Mota-IgA (in-house production) in a 2:1 molar ratio. RBL-WT cells and RBL-huCD89-GFP cells were seeded in a 96-well V-Bottom microplate (Falcon) and a mixture of equal volumes of immune complexed human Mota- IgA (with RSV-F) with a 1/2 dilution series of the antibody was added to the cells. The cells were incubated for 1 hour, followed by two washing steps with FACS buffer. Bound IgA on the cells was stained with a mix of Goat Anti-Human IgA-PE (Southern Biotech), Fixable Viability Dye eFluor 780 (eBioscience) and Purified Mouse Anti-Rat CD32 (Pharmigen). Afterwards, cells were washed and readout was done using the FACS Fortessa. Data analysis was done using FlowJo (V10.5.3).

5.11.2 FcRn occupancy flow cytometry assay using U937 cells

U937 cells were seeded in a 96-well V-Bottom microplate (Falcon). A 7-step dilution range of the anti-lgA antibodies was prepared and added to the cells. After incubation, the plate was washed 2 times with FACS buffer and viability staining was done using the eBioscience Fixable Viability Dye eFluor 506 antibody (Thermo Scientific). Before fixation and permeabilization of the cells, 2 additional washing steps were performed with FACS buffer. For fixation and permeabilization, cells were incubated in acidic Fix/Perm buffer (eBioscience Fix/Perm concentrate in Fix/Perm diluent, Thermo Scientific). After 2 washes with permeabilization buffer (Thermo Scientific), FcRn staining was performed using a fluorescently labelled anti-FcRn Fab fragment recognizing the IgG binding site of FcRn and Fc block (BD) in acidic permeabilization buffer. After staining, the plate was washed 2 times with FACS buffer and read-out was performed on the FACS Fortessa. Data analysis was done using FlowJo (V10.5.3). 5.11 .3 FcRn occupancy flow cytometry assay using human and cynomolqus monkey PBMCs

Human and cynomolgus monkey PBMCs (BiolVT) were incubated with a 8-step dilution range of the anti-lgA antibodies (alone or complexed with IgA in a 1 :3 molar ratio) for 2 hours at 37°C, 5% CO2. The plate was washed with FACS buffer (pH 6.0) and viability staining and Fc block were applied using the eBioscience Fixable Viability Dye eFluor 780 antibody (Thermo Scientific) and Human TruStain FcX (BioLegend). Staining to gate classical monocytes was done by adding BV605 anti-human CD14 and FITC anti-human CD16 (BD Biosciences). Fixation and permeabilization was done using the BD Cytofix/Cytoperm Fixation/Permeabilization Kit (Thermo Scientific) and FcRn staining was performed using a fluorescently labelled anti-FcRn Fab fragment recognizing the IgG binding site of FcRn. After staining, the plate was washed 2 times with FACS buffer (pH6.0) and read-out was performed on the FACS Fortessa. Data analysis was done using FlowJo (V10.5.3).

5.11.4 FcRn degradation flow cytometry assay using HEK-FcRn-GFP cells or U937 cells The FcRn degradation assay was done by flow cytometry using a HEK cell line expressing hFcRn-GFP or U937 cells. Cells were seeded in a 24-well microplate (Costar) and incubated overnight at 37°C, 5% CO2. Antibodies were diluted alone or in a 1 :3 molar ratio with human Mota-IgA (in-house production) and added to the cells overnight. After incubation, the cells were stained with viability dye (eBioscience Fixable Viability Dye eFluor^TM780 antibody - Thermo Scientific). For the assay with U937 cells, cells were fixated and permeabilized afterwards using Fix/Perm buffer (eBioscience Fix/Perm concentrate in Fix/Perm diluent, Thermo Scientific). After 2 washes with permeabilization buffer (Thermo Scientific), FcRn staining was performed using a fluorescently labelled anti-FcRn Fab fragment recognizing the IgG binding site of FcRn and Fc block (BD) in acidic permeabilization buffer. After staining, the plate was washed 2 times with FACS buffer and read-out was performed on FACS Fortessa. Data analysis was done using FlowJo (V10.5.3).

5.11.5 FcRn-mediated IgA internalization flow cytometry assay using HEK cells Human Mota-IgA (in-house production) was labeled with pHrodo Green using the pHrodo i FL Green Microscale Protein labeling kit (Invitrogen) and a pre-mix of anti-lgA antibody and pHrodoGreen Mota-IgA was prepared (1 :1 ratio). HEK-FcRn and HEK-WT cells were seeded in a 96-well V-Bottom microplate (Falcon) and incubated with the antibody mix at 37°C. After centrifugation, cells were stained with a mix containing Fixable Viability Dye eFluor 780 (eBioscience) and Human BD Fc Block (BD). Finally, cells were washed and resuspended in FACS buffer prior to readout on FACS Fortessa. Data analysis was done using FlowJo (V10.5.3).

5.11.6 IgA affinity measurements using Biacore™ T200

Human and cynomolgus monkey IgA affinity was determined using a Biacore™ T200 system. A CM5 chip was coated with anti-human Fc IgG (Jackson ImmunoResearch) followed by capturing the anti-lgA variants. Association of a dilution range of serum human IgA (MP Biomedicals) or cynomolgus monkey serum IgA (Life diagnostics) at pH 7.4 or pH6 (HBS-EP buffer; Cytiva) was conducted. To analyze the results, the 1 :1 binding fitting model on the Biacore™ Insight Software was used.

5.11.7 Cynomolgus monkey and human CD89 competition and displacement MSP

To evaluate the cynomolgus monkey cross-reactivity of the lgA:CD89 blocking capacity of the anti-lgA antibodies, an MSD assay was developed. An MSD GOLD 96-well Streptavidin SECTOR plate (MSD) was blocked for 1 hour at room temperature, washed and captured with BIOTIN-human CD89 (in house biotinylated) or BIOTIN-cynomolgus monkey CD89 (in house biotinylated). For the competition MSD assay, a dilution series of anti-lgA antibodies was prepared with a fixed concentration of human IgA (MP Biomedicals) or cynomolgus monkey IgA (Life diagnostics). After washing the plate, the mixture of IgA and anti-lgA antibodies was added to the plate and allowed to bind for 1 hour at room temperature. For the displacement assay, the fixed concentration of IgA was first added to the plate for 1 hour at room temperature after which a washing step was performed and the anti-lgA dilution series was added for 30 minutes at room temperature. After washing the plate, IgA was detected with a 1/1000 dilution of SULFO-Goat anti-human serum IgA a chain specific F(ab')2 (in house tagged) for 30 minutes and developed with Read buffer 2x (MSD) to be measured directly using an MSD Quickplex reader (MSD).

5.11.8 Complex analysis using mass photometry

For this analysis, human Mota-IgA (in-house production) was added to the gClone B HIS H LALAPG ABDEG TA or OA variants in a 1 :1 ratio. To prevent dissociation of lgA:anti-lgA complexes, 1 pL of a 50mM DSSO (disuccinimidyl sulfoxide; diluted in DMSO, dimethyl sulfoxide) was added to 49 pL of the complex and incubated for 1 hour at room temperature to induce chemical crosslinking of the complexes. The reaction was guenched by adding 1 pL of 1M Tris-HCI pH7.5. The samples were diluted in PBS pH7.4 to a concentration of 20 mM for measurement on the ReFeyn instrument (TwoMP Mass Photometer).

5.11.9 FcRn binding assessment using IBIS’MX96 SPR

An IBIS’MX96 SPR imager system (IBIS Technologies) was used in these experiments. The antibodies were spotted on a SensEye G Easy2Spot (SensEye) in 10 mM Sodium acetate at pH4.5 with 0.075% (v/v) Tween (80) in 2-fold dilution series. Kinetic titration of the different receptors (human and cynomolgus monkey) was performed by injecting twelve concentrations in 1xPBS containing 0.075% (v/v) Tween at pH 7.4 or pH6. The sensor was regenerated between the cycles using two subsequent injections of 20 mM Tris-HCI, 150 mM NaCI pH8.8 and 20 mM H3P04 pH 2.4. KD values were calculated by performing an equilibrium analysis and fitting a Langmuir 1 :1 binding model using Scrubber software version 2 (BioLogic Software).

5.11.10 FcyR binding assessment using IBIS’MX96 SPR

An IBIS’MX96 SPR imager system (IBIS Technologies) was used in these experiments. An array of 4 concentrations of human and cynomolgus monkey FcyRs was captured on a SensEye G-streptavidin sensor (SensEye) in PBS containing 0.075% (v/v) Tween (80) in 3- fold dilution series. Kinetic titration of the test items was performed by injecting concentrations between 0.41 nM to 1000 nM in 1xPBS containing 0.075% (v/v) Tween at pH 7.4. The sensor was regenerated between the cycles by injection of 10 mM Glycine- HCI pH 2.1. KD values were calculated by performing an equilibrium analysis and fitting a Langmuir 1 :1 binding model using Scrubber software version 2 (BioLogic Software).

Example 6: Cynomolgus monkey study comparing OA format vs TA format

To build on the experiments set out in Example 5, an in vivo study in cynomolgus monkeys was carried out. This study was performed to determine whether the observations in the in vitro assays and in vivo mouse studies also apply in a cynomolgus monkey model. As set out in Example 5, OA and TA formats were tested (gClone B HIS H LALAPG ABDEG OA hlgG1 and gClone B HIS H LALAPG ABDEG TA hlgG1).

6.1 Results

An overview of the study design is set out in Figure 29. It was observed that the OA format greatly improved IgA sweeping efficacy, improving both speed and depth of IgA removal as compared to the TA format (Figure 30). Specifically, in the groups receiving repeated dosing, the OA anti-lgA antibody sweeps on average 88.5% of serum IgA on day 12-15, whereas the TA format only sweeps on average 60.5% on day 15-22. The maximum sweeping for the single IV dose with the OA molecule is seen at day 8 with a 91 .5% reduction in IgA levels (Figure 30B).

In addition, the OA format also reduced IgG levels to a higher extent than the TA format upon repeated dosing; the OA format resulted in -32% average reduction in IgG, whereas the TA format reduced IgG levels with an average of 19%. Additionally, a -33.5% average IgG reduction was observed for the single IV dose of the OA format (Figure 31 ). No reduction in IgM levels was observed for both the OA and TA format after single or repeated IV dosing (Figure 32).

Moreover, it was observed that the OA anti-lgA antibody was capable of sweeping IgA efficiently in cynomolgus monkeys irrespective of the baseline levels of IgA (Figure 33). Efficient IgA sweeping was observed in cynomolgus monkeys with high, medium and low baseline IgA levels that received a single IV dose of OA anti-lgA antibody (Figure 33). Multiple doses of the TA anti-lgA antibody efficiently swept IgA in cynomolgus monkeys with high baseline IgA levels; however, the speed and depth of response did not match the sweeping seen with the OA anti-lgA antibody (Figure 34).

6.2 Materials & Methods

Pharmacokinetic sample analysis

Pharmacokinetic analysis of gClone B HIS H LALAPG ABDEG TA and OA was performed using a combination of two anti-human Fabs that specifically recognize the HN mutations H433K/N434F (Fab 14H11 and Fab 13G08, argenx).

Streptavidin pre-coated microtiter plates (MSD GOLD 96-well Streptavidin SECTOR Plate) were blocked with PBS-1 % casein (Biorad) and washed before capturing Fab 14H11 bio (in house biotinylated) for 1 hour. A calibration curve and QC’s were prepared in 1 % cynomolgus monkey serum (in house pooled batch) and samples were diluted in 1 % pooled cynomolgus monkey serum to be in the quantifiable range. After washing the captured plate, the calibration curve, QCs and samples were applied in duplicate to the plate and allowed to bind for 1 hour. Next, the plate was washed and detection of gClone B HIS H LALAPG ABDEG TA or OA was done using Fab 13G08Sulfo (in house sulfotagged). Lastly, the plate was washed and MSD read buffer T with Surfactant (MSD) was applied and measured using an MSD MESO Quickplex SQ120 reader (MSD).

The concentration of gClone B HIS H LALAPG ABDEG TA and OA in serum samples was back-calculated from a calibration curve. The calibration curve was obtained by plotting the levels of electro-chemical luminescence (ECL) from a MesoScale Quickplex SQ120 plate reader. With the use of the GraphPad Prism 9 software, a polynomial, second order polynomial (quadratic) logistic fit was applied to the standard curve with weighing 1/Y².

Pharmacodynamic sample analysis to quantify cynomolqus monkey IgA in serum Streptavidin pre-coated microtiter plates (MSD GOLD 96-well Streptavidin SECTOR) were blocked with PBS-1% casein (Biorad) and washed before adding in-house biotinylated goat anti-human serum IgA a chain specific f(ab')2 (Jackson). A calibration curve and QC- samples of cynomolgus monkey IgA (Life diagnostics) were made. After washing the captured plate, the calibration curve, QCs and samples were applied to the plate in duplicate and allowed to bind for 1 hour. Next, the plate was washed and detection of the cynomolgus monkey IgA was done using in-house sulfo-tagged goat anti-human serum IgA a chain specific f(ab')2 (Jackson). Lastly, the plate was washed and MSD read buffer T with Surfactant (MSD) was applied and measured using an MSD MESO Quickplex SQ120 reader (MSD).

The concentration of cynomolgus monkey IgA in serum samples was back-calculated from a calibration curve. The calibration curve was obtained by plotting the levels of ECL from a MesoScale Quickplex SQ120 plate reader. With the use of the GraphPad Prism 9 software, a 4-parameter logistic fit was applied to the standard curve with 1/Y² weighing.

Pharmacodynamic sample analysis to quantify cynomolgus monkey IqG in serum A Maxisorp Immunoplate (VWR) was coated with mouse anti-Monkey IgG (Southern Biotech) and incubated overnight at 4°C. The plate was washed and blocked with PBS-1% casein (Biorad). After washing the plate, a calibration curve and QC-samples of monkey IgG (MyBioSource) were added for 2 hours in duplicate to the plate, together with the diluted cynomolgus monkey serum samples. Next, the plate was washed and detection of the cynomolgus monkey IgG was done using mouse anti-monkey IgG-HRP (Southern Biotech). Lastly, the plate was washed and detection was performed using TMB (Merck Millipore) and 0.5M H2SO4 (Chemlab), whereafter absorption (450nm ref. 620nm) was measured with a Tecan Infinite 200 PRO (Tecan).

The concentration of cynomolgus monkey IgG in serum samples was back-calculated from a calibration curve. The calibration curve was obtained by the optical density (OD) from an Infinite M NanoTecan plate reader (A450-A620) supported by Magellan Software v7.3. With the use of the GraphPad Prism 9 software, a 5-parameter logistic fit was applied to the standard curve with 1/Y² weighing.

Pharmacodynamic sample analysis to quantify cynomolgus monkey IgM in serum

For quantitative determination of monkey Immunoglobulin M (IgM) in serum, the Monkey IgM ELISA Kit (Alpha Diagnostics International) was used.

The concentration of cynomolgus monkey IgM in serum samples was back-calculated from a calibration curve. The calibration curve was obtained by the OD from an Infinite M NanoTecan plate reader (A450-A620) supported by Magellan Software v7.3. With the use of the GraphPad Prism 9 software, a 5-parameter logistic fit was applied to the standard curve with 1/Y² weighing.

Example 7: Predictions for human dosing with gClone B HIS H LALAPG ABDEG OA hlqG1

Predictions for human dosing with gClone B HIS H LALAPG ABDEG OA hlgG1 show fast and deep reductions in IgA levels (Figure 35). Specifically, doses of 3 mg/kg weekly, 10 mg/kg weekly, 10 mg/kg fortnightly and 30 mg/kg monthly are predicted to lead to >90% IgA reduction (Figure 35).

Example 8: In vitro and in vivo characterization of gClone B HIS H LALAPG YPY OA

An additional OA modified antibody was generated in which the ABDEG™ substitutions in the Fc region were replaced with YPY substitutions (in which amino acids Y, P and Y are present at EU positions 252, 308 and 434, respectively). The OA modified antibody was generated with a variant Fc region having LALAPG YPY and knob-into-hole substitutions as well as the gClone B HIS H Fab. This molecule is referred to herein as “gClone B HIS H LALAPG YPY OA” or LALAPG YPY OA”. 8.1 In vitro FcRn occupancy and degradation assay

In vitro FcRn occupancy studies with gClone B HIS H LALAPG YPY OA were performed in human PBMC’s in accordance with the methods described in section 5.11.3 above.

Overall, better FcRn receptor occupancy was observed for gClone B HIS H LALAPG YPY OA as compared to gClone B HIS H LALAPG ABDEG OA (Figure 36).

8.2 In vivo efficacy study in Albumus™ mice

In vivo studies were performed in Albumus™ mice (which express human FcRn and albumin) in order to evaluate the human Mota-IgA sweeping efficacy of gClone B HIS H LALAPG YPY OA. All mice in the study received a single intravenous (IV) dose of a test item on study day 0 (PBS (negative control), gClone B HIS H LALAPG ABDEG OA or gClone B HIS H LALAPG YPY OA).

The results from this study indicate that gClone B HIS H LALAPG ABDEG OA has a longer half-life than gClone B HIS H LALAPG YPY OA (Figure 37A). In addition, it was observed that gClone B HIS H LALAPG YPY OA exhibited the greatest IgG reduction of all test items (Figure 37B). Finally, it was observed that gClone B HIS H LALAPG ABDEG OA and gClone B HIS H LALAPG YPY OA have effective and comparable IgA sweeping activity (Figure 37C and 37D).

References

Aleyd, E., M. H. Heineke, and M. van Egmond. 2015. 'The era of the immunoglobulin A Fc receptor FcaRI-its function and potential as target in disease', Immunological reviews, 268: 123-38.

Breedveld, A., and M. van Egmond. 2019. 'IgA and FcalphaRI: Pathological Roles and Therapeutic Opportunities', Front Immunol, 10: 553.

Cerutti, A. 2008. 'The regulation of IgA class switching', Nat Rev Immunol, 8: 421-34. de Sousa-Pereira, P., and J. M. Woof. 2019. 'IgA: Structure, Function, and Developability', Antibodies (Basel), 8.

Hansen, I. S., D. L. P. Baeten, and J. den Dunnen. 2019. 'The inflammatory function of human IgA', Cell Mol Life Sci, 76: 1041-55.

Hostoffer, R. W., I. Krukovets, and M. Berger. 1993. 'Increased Fc alpha R expression and IgA-mediated function on neutrophils induced by chemoattractants', J Immunol, 150: 4532- 40. Igawa, T., K. Haraya, and K. Hattori. 2016. 'Sweeping antibody capable of eliminating soluble antigens from circulation', Immunological reviews, 270: 132-51.

Monteiro, R. C., H. Kubagawa, and M. D. Cooper. 1990. 'Cellular distribution, regulation, and biochemical nature of an Fc alpha receptor in humans', J Exp Med, 171 : 597-613.

Qian, K., F. Xie, A. W. Gibson, J. C. Edberg, R. P. Kimberly, and J. Wu. 2008. 'Functional expression of IgA receptor FcalphaRI on human platelets', J Leukoc Biol, 84: 1492-500. van Gool, M. M. J., and M. van Egmond. 2020. 'IgA and FcalphaRI: Versatile Players in Homeostasis, Infection, and Autoimmunity', Immunotargets Ther, 9: 351-72.

Yang, D., C. Giragossian, S. Castellano, M. Lasaro, H. Xiao, H. Saraf, C. Hess Kenny, I. Rybina, Z. F. Huang, J. Ahlberg, T. Bigwarfe, M. Myzithras, E. Waltz, S. Roberts, R. Kroe- Barrett, and S. Singh. 2017. 'Maximizing in vivo target clearance by design of pH- dependent target binding antibodies with altered affinity to FcRn', MAbs, 9: 1105-17.

Duchez, S., R. Amin, N. Cogne, L. Delpy, C. Sirac, V. Pascal, B. Corthesy, and M. Cogne. 2010. 'Premature replacement of mu with alpha immunoglobulin chains impairs lymphopoiesis and mucosal homing but promotes plasma cell maturation', Proc Natl Acad Sci U S A, 107: 3064-9.

Lai, Kar Neng, Sydney C. W. Tang, Francesco Paolo Schena, Jan Novak, Yasuhiko Tomino, Agnes B. Fogo, and Richard J. Glassock. 2016. 'IgA nephropathy', Nature Reviews Disease Primers, 2: 16001.

Scionti, K., K. Molyneux, H. Selvaskandan, J. Barratt, and C. K. Cheung. 2022. 'New Insights into the Pathogenesis and Treatment Strategies in IgA Nephropathy', Glomerular Dis, 2: 15-29.

Vlasak, J., and R. lonescu. 2011. 'Fragmentation of monoclonal antibodies', MAbs, 3: 253- 63.

Wines, B. D., M. D. Hulett, G. P. Jamieson, H. M. Trist, J. M. Spratt, and P. M. Hogarth. 1999. 'Identification of residues in the first domain of human Fc alpha receptor essential for interaction with IgA', J Immunol, 162: 2146-53.

Wines, B. D., C. T. Sardjono, H. H. Trist, C. S. Lay, and P. M. Hogarth. 2001. 'The interaction of Fc alpha Rl with IgA and its implications for ligand binding by immunoreceptors of the leukocyte receptor cluster', J Immunol, 166: 1781-9.

Atwell, S., J. B. Ridgway, J. A. Wells, and P. Carter. 1997. 'Stable heterodimers from remodeling the domain interface of a homodimer using a phage display library', J Mol Biol, 270: 26-35.

Hori, Y., K. Ohmine, H. Katada, Y. Noguchi, K. Sato, T. Nambu, L. R. Adeline, G. S. Wan, K. Haraya, K. Ozeki, M. Nanami, T. Tachibana, Z. Sampei, T. Kuramochi, J. Nezu, K. Hattori, and T. Igawa. 2022. 'Elimination of plasma soluble antigen in cynomolgus monkeys by combining pH-dependent antigen binding and novel Fc engineering', MAbs, 14: 2068213.

Ober, Raimund J., Caius G. Radu, Victor Ghetie, and E. Sally Ward. 2001. 'Differences in promiscuity for antibody-FcRn interactions across species: implications for therapeutic antibodies', International Immunology, 13: 1551-59.

Pyzik, M., K. M. K. Sand, J. J. Hubbard, J. T. Andersen, I. Sandlie, and R. S. Blumberg. 2019. 'The Neonatal Fc Receptor (FcRn): A Misnomer?', Front Immunol, 10: 1540. Schlothauer, T., S. Herter, C. F. Koller, S. Grau-Richards, V. Steinhart, C. Spick, M. Kubbies, C. Klein, P. Umana, and E. Mbssner. 2016. 'Novel human lgG1 and lgG4 Fc- engineered antibodies with completely abolished immune effector functions', Protein Eng Des Sei, 29: 457-66.

Vaccaro, C., R. Bawdon, S. Wanjie, R. J. Ober, and E. S. Ward. 2006. 'Divergent activities of an engineered antibody in murine and human systems have implications for therapeutic antibodies', Proc Natl Acad Sci U S A, 103: 18709-14.

Hezareh, M., A. J. Hessell, R. C. Jensen, J. G. van de Winkel, and P. W. Parren. 2001 . 'Effector function activities of a panel of mutants of a broadly neutralizing antibody against human immunodeficiency virus type T, J Virol, 75: 12161-8.

Claims

1. A monovalent antigen-binding molecule comprising:

2. A monovalent antigen-binding molecule comprising:

3. The monovalent antigen-binding molecule according to claim 1 or claim 2, wherein the antigen-binding domain is a Fab.

4. The monovalent antigen-binding molecule according to any one of claims 1-3, wherein the antigen-binding domain is attached to the N-terminus of either the first Fc domain or the second Fc domain.

5. The monovalent antigen-binding molecule according to any one of claims 1-4, wherein the antigen-binding domain is a Fab and the C-terminus of the Fab heavy chain is attached to the N-terminus of either the first Fc domain or the second Fc domain via an IgG hinge region.

6. The monovalent antigen-binding molecule according to any one of claims 1-5, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) comprising or consisting of the amino acid sequence of SEQ ID NO: 7 or an amino acid sequence having at least 80%, 90%, 95%, 98%, 99% identity thereto, and a variable light chain domain (VL) comprising or consisting of the amino acid sequence of SEQ ID NO: 8 or an amino acid sequence having at least 80%, 90%, 95%, 98%, 99% identity thereto.

7. The monovalent antigen-binding molecule according to claim 6, wherein the antigen-binding domain comprises a variable heavy chain domain (VH) consisting of SEQ ID NO: 7 and a variable light chain domain (VL) consisting of SEQ ID NO: 8.

8. The monovalent antigen-binding molecule according to any one of claims 1-7, wherein the variant Fc region is a variant human Fc region.

9. The monovalent antigen-binding molecule according to any one of claims 1-8, wherein the variant Fc region is a variant IgG 1 Fc region.

10. The monovalent antigen-binding molecule according to any one of claims 1-9, wherein the variant Fc region further comprises at least one additional amino acid substitution as compared with the corresponding wild-type Fc region; and wherein the at least one additional substitution reduces or eliminates Fc effector function.

11 . The monovalent antigen-binding molecule according to any one of claims 1-10, wherein the first Fc domain and second Fc domain comprise the amino acids:

(i) A and A at EU positions 234, 235; and

(ii) G at EU position 329.

12. The monovalent antigen-binding molecule according to any one of claims 1-11 , wherein the variant Fc region further comprises at least one additional amino acid substitution as compared with the corresponding wild-type Fc region; and wherein the at least one substitution promotes dimerisation between the first Fc domain and the second Fc domain.

13. The monovalent antigen-binding molecule according to claim 12, wherein the first Fc domain and the second Fc domain comprise knob-into-holes amino acid substitutions.

14. The monovalent antigen-binding molecule according to any one of claims 1-13, wherein the first Fc domain comprises the amino acid W at EU position 366; and the second Fc domain comprises the amino acids S, A and V at EU positions 366, 368 and 407, respectively.

15. The monovalent antigen-binding molecule according to any one of claims 1 or 3-14, wherein the first Fc domain comprises the amino acids A, A, G, Y, T, E, W, K, F and Y at EU positions 234, 235, 329, 252, 254, 256, 366, 433, 434 and 436, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, T, E, K, F, Y, S, A and V at EU positions 234, 235, 329, 252, 254, 256, 433, 434, 436, 366, 368 and 407, respectively.

16. The monovalent antigen-binding molecule according to any one of claims 1 or 3-15, wherein the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 73 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 74.

17. The monovalent antigen-binding molecule according to any one of claims 2-14, wherein the first Fc domain comprises the amino acids A, A, G, Y, P, Y and W at EU positions 234, 235, 329, 252, 308, 434 and 366, respectively; and the second Fc domain comprises the amino acids A, A, G, Y, P, Y, S, A and V at EU positions 234, 235, 329, 252, 308, 434, 366, 368 and 407, respectively.

18. The monovalent antigen-binding molecule according to any one of claims 2-14 or 17, wherein the first Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 114 and the second Fc domain comprises or consists of the amino acid sequence set forth in SEQ ID NO: 115.

19. The monovalent antigen-binding molecule according to any one of claims 15-18, wherein the antigen-binding domain is a Fab attached to the first Fc domain via an IgG hinge region.

20. The monovalent antigen-binding molecule according to any one of claims 1-19, wherein the first Fc domain and the second Fc domain do not comprise an N-linked glycan at EU position 297.

21 . The monovalent antigen-binding molecule according to any one of claims 1-19, wherein the first Fc domain and second Fc domain comprise an afucosylated N-linked glycan at EU position 297.

22. The monovalent antigen-binding molecule according to any one of claims 1-19, wherein the first Fc domain and second Fc domain comprise an N-linked glycan having a bisecting GIcNac at EU position 297 of the Fc domains.

23. The monovalent antigen-binding molecule according to any one of claims 1-22, wherein the monovalent antigen-binding molecule is a modified lgG1 antibody having only one Fab arm.

24. The monovalent antigen-binding molecule according to claim 23, wherein the monovalent antigen-binding molecule consists of:

25. The monovalent antigen-binding molecule according to claim 23, wherein the monovalent antigen-binding molecule consists of:

(ii) a second immunoglobulin heavy chain consisting of the amino acid sequence of SEQ ID NO: 117; and (iii) an immunoglobulin light chain consisting of the amino acid sequence of SEQ ID NO: 77.

26. A monovalent antigen-binding molecule that binds to IgA consisting of:

27. A monovalent antigen-binding molecule that binds to IgA consisting of:

28. An isolated polynucleotide or polynucleotides, which encode the monovalent antigen-binding molecule of any one of claims 1-27.

29. An expression vector comprising the polynucleotide or polynucleotides of claim 28 operably linked to regulatory sequences which permit expression of the monovalent antigen-binding molecule.

30. A host cell or cell-free expression system containing the expression vector of claim

29.

31 . A method of producing a recombinant monovalent antigen-binding molecule which comprises culturing the host cell or cell free expression system of claim 30 under conditions which permit expression of the monovalent antigen-binding molecule and recovering the expressed monovalent antigen-binding molecule.

32. A pharmaceutical composition comprising a monovalent antigen-binding molecule according to any one of claims 1-27 and at least one pharmaceutically acceptable carrier or excipient.

33. A monovalent antigen-binding molecule according to any one of claims 1-27 or a pharmaceutical composition according to claim 32 for use as a medicament.

34. A method of treating a disorder in a subject in need thereof, wherein the method comprises administering to the subject a therapeutically effective amount of a monovalent antigen-binding molecule according to any one of claims 1-27 or a pharmaceutical composition according to claim 32.

35. The method of claim 34, wherein the disorder is an IgA-mediated disorder.

36. The method of claim 34 or claim 35, wherein the disorder is an IgA autoantibody- mediated disorder.

37. A method of reducing serum IgA levels in a subject, the method comprising administering to the subject a monovalent antigen-binding molecule according to any one of claims 1-27 or a pharmaceutical composition according to claim 32.

38. The method of claim 37, wherein the serum IgA level in the subject is reduced by at least 90% relative to the baseline serum IgA level.

39. The method of any one of claims 34-38, wherein the monovalent antigen-binding molecule is administered at a dose of 3 mg/kg.

40. The method of any one of claims 34-38, wherein the monovalent antigen-binding molecule is administered at a dose of 10 mg/kg.

41 . The method of any one of claims 34-38, wherein the monovalent antigen-binding molecule is administered at a dose of 30 mg/kg.

42. The method of claim 39 or claim 40, wherein multiple doses are administered at a weekly dosing interval.

43. The method of claim 40, wherein multiple doses are administered at a fortnightly dosing interval.

44. The method of claim 41 , wherein multiple doses are administered at a monthly dosing interval.