The present invention relates to polypeptides comprising a mutant human lgG ₄ , which mutant human lgG is capable of increasing the binding to and activation of immunoreceptor tyrosine-based inhibitory motif (ITIM)-containing FcYRIIb/c (CD32b), but not FcyRlla (CD32a). More specifically, the invention relates to polypeptides comprising at least one human lgG ₄ with a lysine at position 409 (409K), using the EU index according to Kabat et al., which lgG is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild-type human Igd and than a wild-type human lgG ₄ , for use in the prevention and/or treatment of an autoimmune disease or allergy, as further defined in the claims.

BACKGROUND OF THE I NVENTI ON

For the proper selection of an antibody isotype, several points have to be considered for the later usage of the antibody as a therapeutic drug (Salfeld, 2007). All currently approved therapeutic antibodies are G-type immunoglobulins or derivatives of mouse, human or mixed origin. The four isotypes are characterized by different features (Table 1 ).

Table 1 : Key features of the four IgG isotypes

IgG l lgG2 lgG3 lgG4

Functional form in vivo Monomeric Dimeric Monomeric Half-lg

bivalent tetravalent ³ bivalent monovalent

Biological role in host response Protein Carbohydrate Protein Response to chronic antigens antigens antigens stimulation,

anti-inf lammatory

Percentage of all IgG in humans ^b 60% 25% 10% 5%

Half-l ife (range in days) ^c 36.3 ± 9.2 37.1 ± 13.9 28.6 ± 10.4 15.6 ± 4.5

( 17.6-56.2) (22.9-62.5) ( 13.0-50.2) (7.1-24.7)

Allotypes' ¹ 4 1 13 0

Fc n ^e + + + +

Hinge length (number of amino acids) 15 12 62 12

Potential (actual) inter-heavy chain 2 (2) 4 (4? ^f ) 1 1 (11) 2 (2)

disulfide bonds in hinge region

Effector functions

Cl ^e ++ - +++ -

FcgR l ^e +++ - +++ ++

FcgR l l ^e + ± + 7

FcgR l l la/b ^e + - + ± ( ^a Homodinners of lgG ₂ are tetravalent for a given antigen, but heterodimers are also expected). lgG3 is characterized by a longer and more flexible hinge domain and the presence of 1 1 inter-heavy chain disulfide bridges (vs. 2 for Igds and lgG ₄ s; 4 for lgG ₂ S). Despite a high antibody-dependent cellular cytotoxicity potential (ADCC), lgG ₃ is generally not selected for therapeutic antibody development, probably due to a short half-life, susceptibility of the extended hinge region to proteolyses, and extensive allotypic polymorphism. Igd is frequently selected for killing pathogenic cells (with an over-expressed target antigen) or viruses. To date, most of the current therapeutic chimeric, humanized and human antibodies are based on an Igd/kappa backbone. lgG ₂ and lgG ₄ show specific structural and functional features like dynamic structural rearrangements, which are not observed for Igd . The advantage of the lgG - over the IgGHsotype is that lgG does not mediate any effector functions, for example complement-mediated lysis (CML) and antibody-dependent cytotoxicity.

One of the disadvantages of lgG ₄ is the formation of half-molecules. In vivo, lgG ₄ antibodies are secreted as both, disulfide-linked tetramers and half-molecules, linked by strong non-covalent interactions. Non-covalently associated lgG ₄ molecules dissociate into half-molecules on SDS-PAGE under non-reducing conditions. However, incubation with a reducing agent, e.g. dithiotreitol is needed (Bloom et al., 1997). In vitro and in vivo, half-molecules can bind the epitope, but blocks it without cross-linking. So, formation of half-molecules leads both to less functional antibodies and to a decreased therapeutic effect. In general, the formation of half-molecules is detected either by SDS-PAGE-analysis or chip- based gel electrophoresis in vitro (Forrer et al., 2004).

Moreover, lgG ₄ can undergo Fab arm exchange. During Fab arm exchange, two lgG ₄ tetramers associate and exchange one half-molecule, consisting of heavy and light chain resulting in bispecific antibodies (van der Neut Kolfschoten et al., 2007; Rispens et al., 2009). Fab arm exchange might be a problem in therapy with lgG ₄ antibodies as they can undergo Fab arm exchange not only with each other but also with endogenous lgG ₄ . The mechanism by which lgG ₄ Fab arm exchange occurs in vivo demands the reducing environment of blood, caused by 0.8+0.2 mM GSH (Michelet et al., 1995), or cell surfaces to facilitate the breaking of the inter- heavy chain disulfide bond within the hinge-region (van der Neut Kolfschoten et al., 2007). For example, natalizumab (marketed under the trademark Tysabri ^® ), a clinical non-stabilized lgG ₄ , undergoes Fab arm exchange in vitro and in vivo (Labrijn et al., 2009). Fab arm exchange can lead to bispecific and monovalent antibodies. So, cross-linking is no longer possible. Due to the above disadvantages, lgG ₄ isotype is usually not the first choice in the development of therapeutic antibodies.

It was shown that mutation of Ser228 to Pro stabilizes lgG ₄ , as an intrachain- disulfide bond cannot be formed easily (Schuurman et al., 2001 ). For example, Mylotarg is a therapeutic lgG ₄ -antibody with S228P-mutation that blocks Fab arm exchange both in w ^' fro and in a mouse model (Labrijn et al., 2009).

Moreover, current literature data show that amino acid differences in the CH3 domain of the lgG Fc-part as compared to the IgGi Fc-part are critical for Fab arm exchange in addition to the lgG ₄ hinge-region (van der Neut Kolfschoten et al., 2007). The Igd and lgG ₄ CH3 domains differ in five amino acid residues, of which only one amino acid is within the CH3 inter-domain interface. This amino acid is located on position 409 and it is lysine (K409) and arginine (R409) within the Igd and lgG ₄ isotype, respectively. Therefore, K409 seems to play a significant role in preventing Igd molecules from building half-molecules.

Approaches for immunosuppression via increasing binding to CD32b or its cross- linking have been patented and described, e.g. an receptor-attracting monoclonal antibody (WO 2009/083009), bispecific monoclonal antibodies (e.g. anti- CD79B/anti-CD32B by Veri et al., 2010) and "Afunctional" monoclonal antibodies (e.g. anti-CD19 and modified Fc region of lgG1 - S267E/L328F by Xencor, Horton et al (2012) and PMID:22257644 with increased affinity for both CD32a and CD32b). Another modified Fc region of an Igd was described by Mimoto et al. (2013) using lgG1 - P238D/L328E, enhancing only CD32b affinity.

WO 2014/144542 A2 discloses bioactive peptide amino acid sequences which inhibit the complement system, and further discloses that such bioactive peptide amino acid sequences may be fused to a human lgG ₄ . The thus modified antibody may be used in the treatment of psoriasis. Vidarsson et al.: IgG Subclasses and Allotypes: From Structure to Effector Functions. Frontiers in Immunology, Vol. 5 (2014), confirm that both lgG and IgGi have binding affinity to the inhibiting Fc receptor CD32b.

Chu et al. Molecular Immunology 45(15): 3926-3933 (2008) suggest that Igd-Fc has a high binding affinity to CD32b, and may have broad applicability in various atoimmune diseases.

The object of the present disclosure is to provide therapeutic lgG ₄ molecules with improved properties such as an increased binding to and activation of the inhibitory Fc receptor CD32b/c. In the context of the prevention and/or treatment of an autoimmune disease or allergy, the therapeutic lgG molecules of the invention are expected to result in an improved therapeutic effect, if administered on a regular basis.

SUMMARY OF THE I NVENTI ON

Surprisingly, it was found that point mutation R409K in the constant region of lgG ₄ increased the affinity of the lgG ₄ antibody for CD32a and CD32b/c, both are FcyRlla/b receptors. As typical lgG ₄ molecule it was expected that it does not show any affinity or functional interactions with human Fey receptors like CD16a, CD32a, CD32b/c and CD64. Further, upon functional analysis, it could be shown that only CD32b/c can be activated by such an antibody. Different wild-type human IgGi and lgG ₄ antibodies, which were used as controls, showed significantly lower affinity for CD32b/c.

CD32b/c contains an immunoreceptor tyrosine-based inhibitory motif (ITIM) within its intracellular domain and acts as a inhibitory receptor (PMID:14613290). CD32b/c is mainly present on B cells, but also on CD14 ⁺ monocytes and myeloid dendritic cells (mDCs) & plasmacytoid dendritic cells (pDCs) and plays an important role in the maintenance of tolerance (PMID:17312177). As triggering of CD32b/c dependent inhibitory signaling pathway in activated B cells leads to the inhibition of B cell proliferation and Ig secretion (PMID: 20506263), stabilized human lgG ₄ as described above could therefore be useful as therapeutic agent in inflammatory or auto-antibody mediated autoimmune diseases like e.g. SLE, psoriasis, multiple sclerosis and rheumatoid arthritis; or in allergy.

The mutant human lgG ₄ mAb with its high affinity and activation potential for the CD32b/c receptor, when used for therapeutic purpose, may induce the CD32b/c dependent signaling in vivo either after being innnnobilized on the cell surface of other immune and endothelial cells (dependent on its specificity) or after being complexed by specific (anti-drug) antibodies, which are occasionally developed upon antibody administration during antibody based therapies. In addition, antibodies consisting of the mutant human lgG ₄ backbone with specificity for cell surface receptors, which are expressed on cells expressing CD32b/c, could be used to specifically inhibit their activation. Such antibodies could bind to the given cell specific receptor (e.g. CD304, CD303, ILT7, CD123 on PDCs or CD19, CD20, BCR on B cells or CD14, CD1 1 c on monocytes) and at the same time bind and cross-link the CD32b/c on the cell surface of the very same cell or on the cell surface of the second cell, which interact with the cell, where the antibody has originally bound.

Accordingly, the above-stated object is deemed to be solved by a polypeptide comprising at least one human lgG with a lysine at position 409 (409K), using the EU index according to Kabat et al., which lgG ₄ is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild- type human IgGi and than a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of an autoimmune disease in a human subject, as further defined in the claims. Likewise, the foregoing object is also deemed to be solved by provision of a polypeptide comprising at least one human lgG ₄ with a lysine at position 409 (409K), using the EU index according to Kabat et al., which lgG is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild-type human Igd and than a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of allergy in a human subject, as further defined in the claims.

DETAI LED DESCRI PTI ON OF THE PREFERRED EM BODI MENTS

Fey receptors (FcyR), which are widely expressed amongst cells of the hematopoietic system, bind the Fc part of IgG antibodies. In the human system the family of FcyR is divided into three main classes, FcyRI (CD64), FcyRII (CD32) and FcyRIII (CD16), each one showing distinct structural and functional properties. In humans each group of receptors is encoded by 2 - 3 closely related genes designated A, B, and C. Binding of immune complexes to activating FcyR, i.e., CD16, CD32A, and CD64, induces a phosphorylation cascade that leads to cell activation characterized by calcium flux and cytokine release. On the contrary, triggering of the inhibitory FcyR (CD32b/c) expressed by B cells induces apoptosis. In contrast, concomitant stimulation of CD32b/c together with an activating receptor (such as the B cell receptor) leads to inhibition of calcium flux and proliferation, i.e., blocking of cell activation (Nimmerjahn and Ravetch, 2008). CD32b/c is expressed on B cells and on a variety of other cell types such as macrophages, neutrophils, and DC. CD32b/c as the only inhibitory receptor comprising an ITIM domain in the intracellular part of the receptor is known to be important in several regulatory processes. Recent data indicated that impairment of CD32b/c signaling is associated with autoimmune disorders such as SLE (Blank et al., 2005).

With the development of many different new monoclonal antibodies (mAb) as therapeutics, FcyR interactions came back into the research focus. Several mAb effector functions such as phagocytosis, induction of inflammatory cytokines or chemokine release and antibody-dependent cytotoxicity (ADCC) are mediated by the interaction of the Fc part with FcyR expressing immune cells, thus shaping the therapeutic potential of mAbs. For example, in a recent study FcyR interactions boosted T cell activation upon treatment with a superagonistic anti-CD28 mAb (Bartholomaeus et al., 2014), whereas in another study with an agonistic anti-CD4 mAb FcyR interactions were needed to induce CDD4 down-modulation of mAb decorated T cells (Vogel et al., 2015).

The present inventors surprisingly found that binding of human lgG ₄ to and activation of the human CD32b/c Fey receptor can be increased by introducing a mutation into the constant region of the lgG ₄ . More specifically, the present disclosure provides a polypeptide comprising at least one human lgG with a lysine at position 409 (409K), using the EU index according to Kabat et al., which lgG ₄ is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild-type human IgGi and than a wild-type human lgG , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of an autoimmune disease in a human subject, as further defined in the claims.

Examples of an autoimmune disease are Addison's disease, agammaglobulinemia alopecia areata, amyloidosis, ankylosing spondylitis, anti-GBM/anti-TBM nephritis, antiphospholipid syndrome (APS), autoimmune hepatitis, autoimmune inner ear disease (AIED), axonal & neuronal neuropathy (AMAN), Behcet's disease, bullous pemphigoid, Castleman disease (CD), Celiac disease, Chagas disease, chronic inflammatory demyelinating polyneuropathy (CIDP), chronic recurrent multifocal osteomyelitis (CRMO), Churg-Strauss, Cicatricial pemphigoid/benign mucosal pemphigoid, Cogan's syndrome, cold agglutinin disease, congenital heart block, Coxsackie myocarditis, CREST syndrome, Crohn's disease, Dermatitis herpetiformis, Dermatomyositis, Devic's disease (neuromyelitis optica), Discoid lupus, Dressler's syndrome, endometriosis, eosinophilic esophagitis (EoE), eosinophilic fasciitis, erythema nodosum, essential mixed cryoglobulinemia, Evans syndrome, fibromyalgia, Fibrosing alveolitis, giant cell arteritis (temporal arteritis), giant cell myocarditis, glomerulonephritis, Goodpasture's syndrome, granulomatosis with polyangiitis, graves' disease, Guillain-Barre syndrome, Hashimoto's thyroiditis, hemolytic anemia, Henoch-Schonlein purpura (HSP), Herpes gestationis or pemphigoid gestationis (PG), hypogammaglobulinemia, IgA nephropathy, lgG ₄ -related sclerosing disease, Inclusion body myositis (IBM), Interstitial cystitis (IC), juvenile arthritis, juvenile diabetes (Type 1 diabetes), juvenile myositis (JM), Kawasaki disease, Lambert-Eaton syndrome, leukocytoclastic vasculitis, Lichen planus, Lichen sclerosus, ligneous conjunctivitis, linear IgA disease (LAD), lupus, lyme disease chronic, Meniere's disease ^' , microscopic polyangiitis (MPA), mixed connective tissue disease (MCTD), Mooren's ulcer, Mucha-Habermann disease, multiple sclerosis (MS), myasthenia gravis, myositis, narcolepsy, neuromyelitis optica, neutropenia, ocular cicatricial pemphigoid, optic neuritis, palindromic rheumatism (PR), PANDAS (Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcus), paraneoplastic cerebellar degeneration (PCD), paroxysmal nocturnal hemoglobinuria (PNH), Parry Romberg syndrome, Pars planitis (peripheral uveitis), Parsonnage-Turner syndrome, pemphigus, peripheral neuropathy, perivenous encephalomyelitis, pernicious anemia (PA), POEMS syndrome , (polyneuropathy, organomegaly, endocrinopathy, monoclonal gammopathy, skin changes), Polyarteritis nodosa, polymyalgia rheumatic, polymyositis, postmyocardial infarction syndrome, postpericardiotomy syndrome, primary biliary cirrhosis, primary sclerosing cholangitis, progesterone dermatitis, psoriasis, psoriatic arthritis, pure red cell aplasia (PRCA), pyoderma gangrenosum, Raynaud's phenomenon, reactive arthritis, reflex sympathetic dystrophy, Reiter's syndrome, relapsing polychondritis, restless legs syndrome (RLS), retroperitoneal fibrosis, rheumatic fever, rheumatoid arthritis (RA), sarcoidosis, Schmidt syndrome, scleritis, scleroderma, Sjogren's syndrome, sperm & testicular autoimmunity, stiff person syndrome (SPS), subacute bacterial endocarditis (SBE), Susac's syndrome, sympathetic ophthalmia (SO), Takayasu's arteritis, temporal arteritis/Giant cell arteritis, thrombocytopenic purpura (TTP), Tolosa-Hunt syndrome (THS), transverse myelitis, type 1 diabetes, ulcerative colitis (UC), undifferentiated connective tissue disease (UCTD), uveitis, vasculitis, vitiligo, and Wegener's granulomatosis (now termed Granulomatosis with Polyangiitis (GPA). Moreover, particularly contemplated patient groups are those human subjects, which show a dysregulation of CD32b expression, as compared to healthy subjects. For example, the human subject may exhibit a CD32b 695T allele and/or a 2B.4 haplotype of the CD32b gene. Alternatively, the human subject may have a CD32b 695C allele and/or a 2B.1 haplotype of the CD32b gene.

Activation of CD32b/c on immune cells such as B cells, PDC or monocytes is believed to result in a down-modulation of these cells. Therefore, it is contemplated that the polypeptide comprising at least one human lgG with a lysine at position 409 (409K) can be advantageously used in the prevention or treatment of an inflammatory autoimmune disease. The autoimmune disease may be further characterized by increased plasma levels of autoantibodies as compared to healthy subjects. In further embodiments, the autoimmune disease may be characterized by an increased type I interferon production, such as increased interferon-alpha production, as compared to healthy subjects.

In a preferred embodiment, the autoimmune disease is selected from the group consisting of psoriasis, systemic lupus erythematosus, rheumatoid arthritis, and multiple sclerosis; and in a more preferred embodiment the autoimmune disease is systemic lupus erythematosus or psoriasis.

Autoantibody immune complexes activate various immune cells such as dendritic cells and B cells via binding to the activating Fc receptor CD32a and subsequent internalization and activation of pattern recognition receptors, e.g. Toll Like Receptors. This, in turn, leads to increased production of pro-inflammatory cytokines such as type-l interferon, in particular of interferon alpha, which again triggers the production of autoantibodies. As shown in the examples below, the lgG ₄ variants disclosed herein do not only bind to the inhibitory Fc receptor CD32b/c with high (or higher) affinity as compared to wild-type lgG , as probably also other immunoglobulines. As shown in the examples, mere binding of an immunoglobuline to CD32b/c does not automatically result in (improved) activation. Rather, the examples herein show that the lgG ₄ variants disclosed herein are particularly suitable for activating the inhibitory Fc receptor CD32b/c, thereby inhibiting the activation of dendritic cells and B cells via CD32a. As further demonstrated in the examples below, in particular the embodiments wherein the polypeptide comprises an lgG ₄ which recognizes an epitope of BDCA-2 (CD303) are particularly advantageous in the treatment of the above-mentioned diseases. This is because targeting BDCA-2 was demonstrated to inhibit type-l interferon production, in particular interferon-alpha production. Accordingly, it is plausibly demonstrated that the polypeptides of the present disclosure can be advantageously applied in a new therapeutic application, which involves the activation of CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of an autoimmune disease in a human subject, in particular wherein the polypeptide comprises an lgG ₄ which recognizes an epitope of BDCA-2 (CD303). As discussed in the examples, psoriasis is an autoimmune disease in which interferon alpha and autoantibodies play a key role in pathogenesis. The same holds true for the other autoimmune diseases recited above. For example, it was generally known that interferon alpha is a primary pathogenic factor in human systemic lupus erythematosus, as reviewed in Niewold, Journal of Interferon & Cytokine Research 31 (12), 887-892 (201 1 ). As shown in Figure 1 on page 889 therein, Autoantibody immune complexes binding to and internalized by Fc receptors play a pivotal role in pathogenesis.

This pivotal role was demonstrated by Means et al., J Clin Invest. 1 15(2), 407-417 (2005): Correlation between disease severity and IFN-alpha levels in SLE patients and mitigation of disease in a mouse model of SLE by disruption of IFN-alpha signaling have implicated IFN-alpha in SLE pathogenesis. High levels of IFN-alpha in the serum of SLE patients have been postulated to play a role in the loss of tolerance to autoantigens in SLE. It has been demonstrated that IFN-alpha in the serum of SLE patients drives the maturation of monocytes into dendritic cells. The ability of these mDCs to engulf apoptotic cells and activate autoreactive T cells and B cells has been hypothesized to drive the autoimmune response in SLE. It was finally demonstrated that stimulation of PDCs with DNA-containing immune complexes from SLE patients through the TLR9/CD32 pathway induced robust expression of IFN-alpha by PDC.

Notably, it could also be demonstrated that anti-BDCA-2 antibodies are capable of inhibiting IFN-alpha production in SLE patients (Blomberg et al. Arthritis & Rheumatism, 48(9), 2524-2532 (2003); Dzionek et al. J Exp. Med. 194(12), 1823- 1834 (2001 )), and that reduction of interferon-alpha activity in SLE patients has therapeutic benefit (cf. clinical trial NCT00299819 for sifalimumab, an anti-IFN- alpha monoclonal antibody). Accordingly, the therapeutic suitability of the polypeptide of the present disclosure is not only demonstrated by the experimental data shown herein below, but also in light of the known effects and the known pathogenesis in SLE.

In light of the facts and evidence provided herein with regard to psoriasis and SLE, the present inventors are of the opinion that the effects and therapeutic benefits demonstrated herein can be extrapolated to any of the above-mentioned diseases, in particular the autoimmune diseases characterized by increased plasma levels of autoantibodies as compared to healthy subjects and/or by an increased type I interferon production, such as increased interferon-alpha production, as compared to healthy subjects, such as rheumatoid arthritis and multiple sclerosis, or any other of the above-disclosed autoimmune diseases. Similar considerations apply to allergies, which are characterized by high IgE titers, resulting in IgE-antigen immune complexes which may in turn activate CD32 bearing B cells or monocytes. The polypeptide may be any polypeptide as long as it comprises at least one human lgG ₄ constant region comprising the R409K substitution. For example, the polypeptide may be a fusion protein which is able to bind to CD32b/c, and which contains the mutant lgG constant region of the present disclosure. For example, the mutant human lgG ₄ disclosed herein is fused to a ligand molecule or a soluble receptor molecule. Alternatively, the polypeptide may comprise two Fc regions, e.g. two human lgG ₄ Fc regions (e.g. fused tail-to-tail), or one human lgG ₄ Fc region fused to a human Igd Fc region.

However, as demonstrated herein, the (therapeutic) effects are obtained by the herein disclosed polypeptide as such, and not dependent on a fusion to a bioactive peptide amino acid sequences as, e.g. described in WO 2014/144542 A2. In the latter, the pharmaceutical effects are disclosed in the context of particular peptide sequences which are fused to, for example, lgG ₄ . Accordingly, in a preferred embodiment, the polypeptide is not fused to a bioactive peptide amino acid sequence which acts as the pharmaceutically active agent.

Also contemplated within the polypeptide of the present disclosure are monomeric antibodies bearing the mutant lgG constant region of the present disclosure, e.g. chimeras in which the human lgG ₄ constant region is fused to a variable region of an antibody of another isotype or species.

In a particularly preferred embodiment, the human lgG ₄ comprises a further substitution, e.g. a substitution which is known to further stabilize the human lgG molecule. Such substitutions are known in the field. Particularly contemplated is an embodiment, wherein the human lgG ₄ further comprises a proline at position 241 (241 P), using the EU index according to Kabat et al. In another embodiment, the human lgG further comprises a proline at position 228 (228P), using the EU index according to Kabat et al.. In still another embodiment, the human lgG ₄ further comprises a proline at position 228 (228P) and a proline at position 241 (241 P), using the EU index according to Kabat et al..

In certain embodiments, the mutant lgG constant region of the present disclosure has a sequence identity of at least 95% over the whole length of SEQ ID NO: 1 , more preferably a sequence identity of at least 96% over the whole length of SEQ ID NO: 1 , more preferably a sequence identity of at least 97% over the whole length of SEQ ID NO: 1 , more preferably a sequence identity of at least 98% over the whole length of SEQ ID NO: 1 , and most preferably a sequence identity of at least 99% over the whole length of SEQ ID NO: 1 .

In a special embodiment, the human lgG ₄ comprises the amino acid sequence of SEQ ID NO: 1 (hlgG4) with a lysine at the position corresponding to position 409 (409K) and a proline at the position corresponding to position 241 (241 P), using the EU index according to Kabat et al.

Generally, an amino acid sequence has "at least x % identity" with SEQ ID NO: 1 , when the sequence identity between the aligned sequences is at least x % over the full length of SEQ ID NO: 1 . Such an alignment can be performed using publicly available computer homology programs, e.g., the "BLAST" program provided at the NCBI homepage at http://www.ncbi.nlm.nih.gov/blast/blast.cgi, using the default settings provided therein. The human lgG ₄ Fc bearing the R409K substitution is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild- type human IgGi and than a wild-type human lgG ₄ . The expression "binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild-type human Igd and than a wild-type human lgG ₄ " as used herein means that the polypeptide comprising the at least one human lgG Fc bearing the R409K substitution shows a statistically significant (p=0.05) higher binding affinity as compared to a non-relevant wild-type human lgG ₄ and a wild type human Igd (or a polypeptide comprising said wild-type human lgG ₄ and a wild type human Igd instead of the human lgG bearing the R409K mutation), when subjecting the polypeptide and the wild type antibodies to the following assay:

ELISA plates are coated with recombinant Fey receptor Mb (e.g., R&D Systems, 1875-CD-050) with 5 g/mL in coating buffer overnight at 4°C. Coated plates are washed 3-fold with 1 x PBS 0.05% Tween 20 and blocked with 5% BSA in 1x PBS. The polypeptide and antibodies are applied to ELISA plates with 50 L/well (or at the same molar concentration corresponding to 50 g human lgG ₄ per well) at serial 2-fold dilution of antibody solution and incubated at room temperature for 1 h. As blank controls, respective antibody concentrations are also applied to wells lacking coated Fey receptors, in order to subtract unspecific binding from resulting signals. After washing (see above), the detection antibody (e.g., anti-human kappa-HRP, FE 1 10316.07, at 1 :4000; or another commercially available anti- lgG -HRP conjugated antibody or anti-lgGi-HRP conjugated antibody applied in a dilution in accordance with the manufacturers instruction) in 1xPBS is applied with 50 L/well and incubated again at RT for 1 h. After additional washing as described above, for assay development 50 L/well TMB (CH1 1 1027.37) are applied and stopped after approx. 5 minutes with 50 L/well 10% H ₂ SO ₄ . The developed plate is then analyzed via ELISA reader at 450 nm wavelength. In order to further improve the comparability, one may also prepare a standard curve applying different known amounts of lgG ₄ and/or Igd which are then correlated to the absorbance at 450 nm wavelength.

Alternatively, the human lgG ₄ with a lysine at position 409 (409K using the EU index according to Kabat et al.) described herein is capable of binding to human CD32b/c with a higher binding affinity of at least 2-fold, preferably at least 5-fold and most preferably at least 10-fold than the binding affinity of a wild-type human IgGi and / or a wild-type human lgG , as determined by using the above assay. As shown in the examples, the human lgG ₄ 409K shows an increased activation of human CD32b/c as compared to wt human lgG ₄ and/or wt human Igd . The expression "increased activation of human CD32b/c as compared to wt human lgG and/or wt human IgGT as used herein is intended to mean that the polypeptide of the invention activates CD32b/c stronger than a wild type human lgG ₄ and/or wild type human Igd (or a corresponding polypeptide comprising a wild type human lgG ₄ and/or wild type human Igd instead of the human lgG ₄ 409K). This feature can be tested by using the following activation assay. As apparent from this definition, binding to CD32b/c does not automatically result in an increased activation of human CD32b/c. Indeed, the surprising aspect of the present invention is that the polypeptide of the present disclosure triggers increased activation of CD32b/c upon binding to the receptor.

To test the ability of different mAbs to bind to and to cross-link the individual FcyRs, mAbs are coated to plastics and a murine thymoma cell line BW5147 transfected with a fusion protein consisting of the extracellular domain of CD32b and the transmembrane and intracellular ζ-chain of the murine T cell receptor is added. If a mAb interacts with a tested FcyR, murine IL-2 production is induced. Absolute IL-2 levels are taken as readout of FcY-CD32b/c-interaction. Tests with control mAbs of the wild type lgG ₄ and/or Igd (or a corresponding polypeptide comprising a wild type human lgG and/or wild type human Igd instead of the human lgG ₄ 409K) allow comparative evaluation of interaction strength. This experimental setting is suited to study conditions of Fc-FcR-interactions that are also relevant in vivo, when a mAb is bound to a cell-surface target and interacts with Fc-receptors on different immune cells. The production of the reporter cell line is further described in Example 2 below.

The polypeptide of the disclosure, wild type human lgG ₄ , wild type human Igd, or a corresponding control polypeptide comprising a wild type human lgG ₄ and/or wild type human Igd instead of the human lgG 409K, are serially 3-fold diluted in 1 :10 (total log3 dilution) in binding buffer (10 mM Bis-Tris, pH 6) starting with a concentration of 10 g/ml. 100 μΙ of each dilution is transferred to a single well and the coating is performed over night at 4°C. After removal of the coating reagent, 200 μΙ per well blocking buffer (PBS + FCS 10%) are added and incubated for 1 h at RT. After blocking, the wells are washed 3 times with 300 μΙ PBS. CD32b- expressing BW5147 reporter cells are transferred into the wells at a concentration of 2x10 ⁵ cells per 200 μΙ cell culture medium. Cells are incubated for 24h at 37°C, 5% CO2. Cell-free supernatant is then collected and analyzed for murine IL-2 using a mouse IL-2 ELISA kit (Bender MedSystems, #BMS601 ) following the manufacturers' instructions. IL-2 levels directly correlate with the degree of activation.

In a preferred embodiment, the polypeptide of the present disclosure is a monoclonal antibody. "Monoclonal antibody" as used herein means that the antibody comprises identical CDRs, thereby recognizing identical epitope(s). Usually, monoclonal antibodies are produced recombinantly using hybridoma technology, as generally known in the art. While the human lgG ₄ monoclonal antibody of the present disclosure binds to human CD32b/c via its heavy chain Fc part, and is capable of activating same on CD32b/c expressing cells, in a preferred embodiment said antibody does not recognize the human CD32b/c receptor via its antigen binding regions, i.e. it does not recognize an epitope of human CD32b/c. Instead, the monoclonal antibody exhibits specificity for a different antigen. In certain embodiments, the monoclonal antibody may be bispecific. This means that each of the two antigen binding region may bind to a different epitope of the same antigen or to a different epitope on a different antigen. In accordance with the foregoing disclosure, the antibody disclosed herein may be a bispecific, trifunctional antibody; or the antibody said antibody may be a bispecific, bifunctional antibody. In certain embodiments, the monoclonal antibody is unmodified, and in particular not fused to a bioactive peptide amino acid sequence which acts as the pharmaceutically active agent, as e.g. described in WO 2014/144542 A2.

However, in preferred embodiments, the monoclonal antibody is a monospecific antibody. As a monospecific monoclonal antibody, the human lgG ₄ antibody of the present disclosure will bind the same epitope with both antigen binding regions, i.e. the antibody has only one singular specificity.

In preferred embodiments, the antibody of the present disclosure binds to a cell surface receptor which is typically also present on the CD32b/c expressing cell. An antibody with such specificity may be able to cross-link CD32b/c along with said cell surface receptor, or at least co-localize the inhibitory CD32b/c receptor with said cell surface receptor in order to possibly down-modulate any activation of the cell surface receptor.

For example, the antibody bearing the mutant lgG ₄ constant region of the present disclosure could be chemically, covalently or non-covalently multimerized in a way that enables cross-linking of CD32b/c on CD32b/c expressing cells without the need for immune complex formation.

The antibody may bind to an epitope on the cell surface on CD32b/c expressing cell (e.g. PDC, monocyte, B cell, myeloid DC and others), thereby being immobilized and at the same time binding and cross-linking the CD32b/c receptor through its mutant lgG ₄ constant region as described herein, thereby inhibiting activation of the given cell and/or inducing regulatory phenotype. Cross-linking of any antibody bearing the human lgG ₄ constant region of the present disclosure and formation of immune complexes by anti-drug antibodies may introduce a beneficial effect in autoimmune setting by cross-linking CD32b/c by the means of the mutant lgG ₄ constant region disclosed herein, thereby inhibiting activation of B cells or other CD32b/c expressing cells (e.g. monocytes, myeloid or plasmacytoid DCs).

Alternatively, the polypeptide or monoclonal antibody of the invention may be designed such that it interacts with another cell which normally interacts with the CD32b/c expressing cell. An antibody bearing the human lgG ₄ constant region of the present disclosure may be immobilized upon binding to its target epitope on the surface of any kind of cell (e.g. hematopoietic cell subsets, endothelial cells, epithelial cells and others). Then, upon cell-cell contact between cell expressing the epitope of the lgG ₄ antibody and the cell expressing CD32b/c, the antibody may cross-link the CD32b/c on the surface of the CD32b/c expressing cell, thereby inhibiting activation of the CD32b/c expressing cell.

In a preferred embodiment, the antibody recognizes an epitope of a cell surface receptor selected from CD303 (also known as BDCA-2), CD304, ILT7, and CD123, as they can be found on the cell surface of PDCs. In other preferred embodiments, the antibody is directed against a cell surface molecule or receptor on B cells. For example, the antibody may recognize an epitope of a cell surface receptor selected from CD19, CD20, the B cell receptor (BCR), CD79A, and CD79B. In still other preferred embodiments, the antibody recognizes an epitope of a cell surface receptor or molecule on a monocyte. In particular, the antibody of the present disclosure may recognize an epitope selected from CD14 and CD1 1 c. In still another preferred embodiment, the antibody recognizes an epitope of a cell surface receptor selected from CD141 and CD1 1 c.

In a most preferred embodiment, the antibody recognizes an epitope of BDCA-2 (CD303). An example of such an antibody is the monoclonal antibody MB101 . Said antibody is, for example, obtainable from the hybridoma deposited under the accession number DSM ACC3237 on 24 April 2014 by Miltenyi Biotec GmbH, Bergisch Gladbach at the DSMZ (German Collection of Microorganisms and Cell Cultures) in Braunschweig, Germany, and represents another preferred embodiment of the present disclosure. In accordance with the monoclonal antibody MB101 , the antibody of the present disclosure may comprises the heavy chain CDRs 1 -3 shown in SEQ ID NOs: 2-4 and light chain CDRs 1 -3 shown in SEQ ID NOs: 5-7. Even more preferably, the antibody of the present disclosure may comprise the variable heavy chain of SEQ ID NO: 8 and the variable light chain of SEQ ID NO: 9. As shown in Example 5, data of a phase I clinical trial appears to support a curative effect of MB101 in an autoimmune disease such as psoriasis.

In a further aspect, the present disclosure further pertains to a polypeptide comprising at least one human lgG ₄ with a lysine at position 409 (409K) using the EU index according to Kabat et al., which lgG ₄ is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild- type human IgGi and than a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of allergy in a human subject, as further defined in the claims. As far as it makes technical sense, preferred embodiments of the above-described first aspect with regard to the polypeptide of the present disclosure apply likewise to this second aspect. In particular, the human lgG ₄ with a lysine at position 409 (409K using the EU index according to Kabat et al.) described herein is capable of binding to human CD32b/c with a higher binding affinity of at least 2-fold, preferably at least 5-fold and most preferably at least 10-fold than the binding affinity of a wild-type human Igd and / or a wild-type human lgG ₄ , as determined by using the above assay. In a preferred embodiment, the human lgG ₄ further comprises a proline at position 241 (241 P), using the EU index according to Kabat et al.. It is believed that this additional substitution further stabilizes the lgG ₄ molecule. Accordingly, in a particularly preferred embodiment, the human lgG comprises the amino acid sequence of SEQ ID NO: 1 (hlgG4) with a lysine at the position corresponding to position 409 (409K) and a proline at the position corresponding to position 241 (241 P), using the EU index according to Kabat et al. As described above, the polypeptide may be a fusion protein, in particular a fusion protein comprising two human lgG ₄ Fc parts of the present disclosure, or a human lgG ₄ Fc in fusion with a human Igd or a human IgE. Still, in a further preferred that the polypeptide is a monoclonal antibody. While it is possible that such a monoclonal antibody is designed to bind to CD32b/c via its antigen binding regions, in preferred embodiments the monoclonal antibody described herein does not recognize an epitope of human CD32b/c via its antigen binding regions. Rather, in the context of the treatment of allergen, the antibody may recognize an epitope of IgE. For example, the antibody may exhibit the CDRs or variable binding regions of anti-lgE monoclonal antibody Xolair ^® . Alternatively, the antibody may recognize an epitope of a cell surface receptor selected from CD14 and CD1 1 c; or a cell surface receptor selected from CD141 and CD1 1 c. Hence, in certain embodiments, the antibody recognizes an epitope of a cell surface receptor selected from CD303, CD304, I LT7, CD123, CD19, CD20, the B cell receptor (BCR) such as IgE, CD79A, CD79B, CD14, CD1 1 c, and CD141 .

The present disclosure is further described by the following embodiments.

1 . A polypeptide comprising at least one human lgG ₄ with a lysine at position 409 (409K), using the EU index according to Kabat et al., which lgG ₄ is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild-type human Igd and than a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of an autoimmune disease in a human subject.

2. The polypeptide for use of embodiment 1 , wherein the autoimmune disease is an inflammatory autoimmune disease.

3. The polypeptide for use of any one of embodiment 1 or 2, wherein the autoimmune disease is further characterized by increased plasma levels of autoantibodies as compared to healthy subjects. The polypeptide for use of any one of embodiments 1 -3, wherein the autoimmune disease is selected from the group consisting of systemic lupus erythematosus, psoriasis, multiple sclerosis, and rheumatoid arthritis. The polypeptide for use of any one of embodiments 1 -4, wherein the autoimmune disease is systemic lupus erythematosus or psoriasis.

The polypeptide for use of any one of embodiments 1 -5, wherein the human lgG ₄ further comprises a proline at position 241 (241 P), using the EU index according to Kabat et al.

The polypeptide for use of any one of embodiments 1 -6, wherein the human lgG ₄ comprises the amino acid sequence of SEQ ID NO: 1 (hlgG ₄ ) with a lysine at the position corresponding to position 409 (409K) and a proline at the position corresponding to position 241 (241 P), using the EU index according to Kabat et al.

The polypeptide for use of any one of embodiments 1 -7, wherein the human lgG ₄ 409K 241 P shows an increased activation of human CD32b/c as compared to wt human lgG ₄ and/or wt human lgG1 .

The polypeptide for use of any one of embodiments 1 -8, wherein the polypeptide is a monoclonal antibody.

The polypeptide for use of embodiment 9, wherein the monoclonal antibody does not recognize an epitope of human CD32b/c via its antigen binding region.

The polypeptide for use of embodiment 9 or 10, wherein the monoclonal antibody is a monospecific antibody.

The polypeptide for use of any one of embodiments 1 -1 1 , wherein the antibody recognizes an epitope of a cell surface receptor selected from CD303, CD304, ILT7, and CD123.

The polypeptide for use of any one of embodiments 1 -1 1 , wherein the antibody recognizes an epitope of a cell surface receptor selected from CD19, CD20, the B cell receptor (BCR), CD79A, and CD79B.

The polypeptide for use of any one of embodiments 1 -1 1 , wherein the antibody recognizes an epitope of a cell surface receptor selected from CD14 and CD1 1 c.

The polypeptide for use of any one of embodiments 1 -1 1 , wherein the antibody recognizes an epitope of a cell surface receptor selected from CD141 and CD1 1 c.

The polypeptide for use of embodiment 12, wherein the antibody recognizes an epitope of BDCA-2 (CD303). The polypeptide for use of embodiment 16, wherein the antibody comprises the heavy chain CDRs 1 -3 shown in SEQ ID NOs: 2-4 and light chain CDRs 1 -3 shown in SEQ ID NOs: 5-7.

The polypeptide for use of embodiment 17, wherein the antibody comprises the variable heavy chain of SEQ ID NO: 8 and the variable light chain of SEQ ID NO: 9.

The polypeptide for use of any one of embodiments 1 -10 or 12-18, wherein the antibody is a bispecific, trifunctional antibody; or wherein the antibody is a bispecific, bifunctional antibody.

The polypeptide for use of any one of embodiments 1 -19, wherein the human subject shows a dysregulation of CD32b expression, as compared to healthy subjects.

The polypeptide for use of any one of embodiments 1 -20, wherein the human subject has a CD32b 695T allele and/or a 2B.4 haplotype of the CD32b gene.

The polypeptide for use of any one of embodiments 1 -20, wherein the human subject has a CD32b 695C allele and/or a 2B.1 haplotype of the CD32b gene.

A polypeptide comprising at least one human lgG ₄ with a lysine at position 409 (409K) using the EU index according to Kabat et al., which lgG ₄ is capable of binding to human CD32b/c with a statistically significant (p=0.05) higher binding affinity than a wild-type human Igd and than a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of allergy in a human subject.

The polypeptide for use of any one of embodiment 23, wherein the human lgG ₄ further comprises a proline at position 241 (241 P), using the EU index according to Kabat et al..

The polypeptide for use of embodiment 23 or 24, wherein the human lgG ₄ comprises the amino acid sequence of SEQ ID NO: 1 (hlgG ₄ ) with a lysine at the position corresponding to position 409 (409K) and a proline at the position corresponding to position 241 (241 P), using the EU index according to Kabat et al.

The polypeptide for use of any one of embodiments 23-25, wherein the polypeptide is a monoclonal antibody.

The polypeptide for use of embodiment 26, wherein the monoclonal antibody does not recognize an epitope of human CD32b/c via its antigen binding region. 28. The polypeptide for use of embodiment 26 or 27, wherein the antibody recognizes an epitope of IgE.

29. The polypeptide for use of embodiment 26 or 27, wherein the antibody recognizes an epitope of a cell surface receptor selected from CD14 and CD1 1 C.

30. The polypeptide for use of embodiment 26 or 27, wherein the antibody recognizes an epitope of a cell surface receptor selected from CD141 and CD1 1 C.

31 . The polypeptide for use of embodiment 26 or 27, wherein the antibody recognizes an epitope of a cell surface receptor selected from CD303, CD304, I LT7, CD1 23, CD19, CD20, the B cell receptor (BCR) such as IgE, CD79A, CD79B, CD14, CD1 1 c, and CD141 .

32. A human lgG ₄ with a lysine at position 409 (409K using the EU index according to Kabat et al.), which lgG ₄ is capable of binding to human CD32b/c with a higher binding affinity of at least 2-fold, preferably at least 5- fold and most preferably at least 10-fold than the binding affinity of a wild- type human Igd and / or a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of an autoimmune disease in a human subject.

33. The human lgG ₄ for use of embodiment 32, as further defined by above embodiments 2-22.

34. A human lgG ₄ with a lysine at position 409 (409K using the EU index according to Kabat et al.), which lgG ₄ is capable of binding to human CD32b/c with a higher binding affinity of at least 2-fold, preferably at least 5- fold and most preferably at least 10-fold than the binding affinity of a wild- type human Igd and / or a wild-type human lgG ₄ , for use in the activation of human CD32b/c on CD32b/c expressing cells in the prevention and/or treatment of allergy in a human subject.

35. The human lgG ₄ for use of embodiment 34, as further defined by above embodiments 24-31 .

The present invention is further illustrated in the following Examples and Figures, which are not intended to be understood as to limit the invention as defined by the claims.

DESCRIPTION OF THE FIGURES

Figure 1 shows the binding affinity of different Igd mAbs and MB101 to recombinant human FcyRI (CD64) receptor (plate bound). It can be observed, that both Igd variants have a high affinity to CD64, as expected for Igd monoclonal antibodies. MB101 has a significantly lower affinity for CD64, which is also not surprising as MB101 is a humanized lgG molecule for which lower affinities to CD64 are reported, compared to Igd molecules

Figure 2 shows the binding affinity of a chimeric IgGi mAb and MB101 .10 to recombinant human FcyRlla (CD32a) receptor (plate bound). It can be observed that MB101 has a moderately higher affinity for CD32a compared to the chimeric Igd molecule.

Figure 3 shows the binding affinity of an chimeric Igd mAb and MB101 to recombinant human FcyRllb (CD32b) receptor (plate bound). It can be observed that MB101 has a significantly higher affinity for CD32b receptor compared to the chimeric Igd molecule.

Figure 4 shows the binding affinity of an chimeric Igd mAb and MB101 .10 to recombinant human Fey Rllla (CD16a) receptor (plate bound). It can be observed that MB101 has a significantly lower affinity for CD16a compared to the chimeric IgGi molecule. A minimal residual binding affinity to CD16a can still be found. Figure 5 shows the binding affinity of an chimeric Igd mAb and MB101 .10 to recombinant human Fey Rlllb/c (CD16b/c) receptor (plate bound). It can be observed that MB101 has no affinity for CD16b/c. In comparison, the chimeric IgGi molecule has a medium binding affinity to CD16b/c.

Figure 6: Interaction of different mAbs with transfectants expressing the human FcyR CD1 6vi58- BW5147 transfectants expressing the FcyR CD16 were incubated with coated TGN1412 (dot), Tysabri (square), Mabthera (triangle), Remicade (black diamond), or MB101 (blue diamond) at the indicated concentrations and murine IL-2 secretion was measured by an ELISA method (Bender MedSystems). Values are depicted as the mean of duplicate (MB101 ) or triplicate (TGN1412, Tysabri, Mabthera, and Remicade) measurements ± SD.

Figure 7: Interaction of different mAbs with transfectants expressing the human FcyR CD32A. BW5147 transfectants expressing the FcyR CD32A were incubated with coated TGN1412 (dot), Tysabri (square), Mabthera (triangle), Remicade (black diamond), or MB101 (blue diamond) at the indicated concentrations and murine IL-2 secretion was measured by ELISA (Bender MedSystems). Values are depicted as the mean of duplicate (MB101 ) or triplicate (TGN1412, Tysabri, Mabthera, and Remicade) measurements ± SD. Figure 8: Interaction of different mAbs with transfectants expressing the human FcyR CD32b/c. BW5147 transfectants expressing the FcyR CD32b/c were incubated with coated TGN1412 (dot), Tysabri (square), Mabthera (triangle), Remicade (black diamond), or MB101 (blue diamond) at the indicated concentrations and murine IL-2 secretion was measured by ELISA (Bender MedSystems). Values are depicted as the mean of duplicate (MB101 ) or triplicate (TGN1412, Tysabri, Mabthera, and Remicade) measurements ± SD.

Figure 9: A single amino acid exchange introduced into the constant region of MB101 (Pool 6; R409K) leads to a significantly increased binding to CD32b and CD64. MB101 (anti-CD303) expressed as a wild type lgG ₄ and the two single mutation variants Pool 3 (anti-CD303 lgG ₄ S241 P) and Pool 6 (anti-CD303 lgG ₄ R409K) were coated in a log3 serial dilution starting with 1 g to 96well plates in Na ₂ HPO binding buffer for 2 h at 37°C. As controls, antibodies from the lgG (TGN1412) and Igd (BT-061 ) subclass were used. Then transfectants expressing one of the human FcyR were incubated on the coated plates for 18 h. Afterwards, cell-free supernatants were harvested and the mouse IL-2 content was analyzed by an ELISA method. N=5 with two biological duplicates for each value. Statistical analysis was performed using a two-way Anova with Bonferroni's multiple comparison test.

Figure 10: MB101 (anti-CD303 lgG ₄ S241 P/R409K) and Pool 6 (anti-CD303 lgG ₄ R409K) induce CD32b expressing transfectants to produce significantly more IL-2 than Pool 3 (anti-CD303 lgG ₄ S241 P) and the wt lgG ₄ . To examine more precisely the CD32b binding of MB101 , the wt lgG ₄ and the two single mutation variants, linear dilutions of the mAbs were coated to plates starting with 1 g and incubated with the CD32b expressing transfectants. After 18 h supernatant was harvested and analyzed by an ELISA method. N=3 with two biological duplicates for each value. Statistical analysis was performed using a two-way Anova with Bonferroni's multiple comparison test.

Figure 11 : MB101 (anti-CD303 lgG ₄ S241 P/R409K) and Pool 6 (anti-CD303 lgG ₄ R409K) show stronger binding to CD32b/c compared to Pool 3 (anti-CD303 lgG ₄ S241 P) and the wt lgG . To examine more precisely the CD32b/c binding of MB101 , the wt lgG ₄ and the two single mutation variants, linear dilutions of the mAbs were incubated with ELISA plates previously coated with recombinant CD32b/c. FcR-bound anti-BDCA2 Ab were probed with anti-human kappa-HRP and TMB-substrate.

Figure 12: MB101 (anti-CD303 lgG ₄ S241 P/R409K) and Pool 6 (anti-CD303 lgG ₄ R409K) induce significantly more GFP expression in BDCA-2 transfectants than Pool 3 (anti-CD303 lgG ₄ S241 P) or the wt lgG ₄ . To examine whether MB101 , the wt lgG ₄ and the two single mutation variants differ in their binding to BDCA-2, the induction of GFP expression was analyzed in transfectants constitutively expressing BDCA-2. (A) mAbs were added at the indicated concentrations to medium alone and then 1 x10 ⁵ BDCA-2 transfectants were added. Percentage of GFP ⁺ cells was analyzed after 24 h of culture by using an LSRII and FlowJo software. (B) Antibodies were added at the indicated concentrations to 1 x10 ⁵ CD99 (upper row) or CD32b (lower row) transfectants before adding 1 x10 ⁵ BDCA- 2 transfectants. Percentage of GFP ⁺ cells was analyzed after 24 h of culture. One typical experiment is shown.

Figure 13: Incubation of BDCA-2 transfectants with MB101 (anti-CD303 lgG ₄ S241 P/R409K) and Pool 6 (anti-CD303 lgG ₄ R409K) induce significantly more GFP expression than Pool 3 (anti-CD303 lgG ₄ S241 P) and the wt lgG ₄ . To examine whether MB101 , the wt lgG ₄ and the two single mutation variants differ in their binding to BDCA-2, GFP induction of BDCA-2 transfectants was analyzed. (A) Antibodies were added at the indicated concentrations to 1 x10 ⁵ CD32b transfectants and then 1 x10 ⁵ BDCA-2 transfectants were added. Percentage of GFP ⁺ BDCA-2 transfectants was analyzed after 24 h of culture. Percentage of GFP ⁺ cells was analyzed after 24 h of culture by using an LSRII and FlowJo software. (B) like (A) but the mean fluorescence intensity (MFI) was analyzed. Data from five independent experiments are shown. Statistical analysis was performed using a two-way Anova with Tukey's multiple comparison test.

DESCRI PTI ON OF THE SEQUENCES

SEQ ID NO: 1 - Human lgG ₄ heavy chain constant region (CH1 -hinge-CH2-CH3; accession number UNIPROT P01861 -1 ; S241 and R409 are emphasized)

10 20 30 40 50

ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV

60 70 80 90 100

HTFPAVLQSS GLYSLSSVVT VPSSSLGTKT YTCNVDHKPS NTKVDKRVES

110 120 130 140 150 KYGPPCPSCP APEFLGGPSV FLFPPKPKDT LMI SRTPEVT CVVVDVSQED

160 170 180 190 200

PEVQFNWYVD GVEVHNAKTK PREEQFNSTY RVVSVLTVLH QDWLNGKEYK

210 220 230 240 250

CKVSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK

260 270 280 290 300

GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSRL TVDKSRWQEG

310 320

NVFSCSVMHE ALHNHYTQKS LSLSLGK

SEQ ID NO: 2 - MB101 heavy chain CDR1

SGFSLSTSGMGVG

SEQ ID NO: 3 - MB101 heavy chain CDR2

HIWWEDDKYYNPSLKS

SEQ ID NO: 4 - MB101 heavy chain CDR3

TRNWDYYTMDY SEQ ID NO: 5 - MB101 light chain CDR1

RASQEISGYLS

SEQ ID NO: 6 - MB101 light chain CDR2

YAASTLDS

SEQ ID NO: 7 - MB101 light chain CDR3

LQYASYPPT

SEQ ID NO: 8 - MB101 variable region heavy chain (CDRs emphasized)

OVTLKESGPTLVKPTOTLTLTCTFSQFSLSTSQMQVQWIROPPGKALEWLAHIWWEDDKY Y NPSLKSRLT I TKDTS KNQ WLTMTNMD PVDTATYYCTRN DYYTMDYWGOGTTVTVS S AS T

KGPSVFPLAPCSRSTSESTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSS GLYS LSSWTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYGPPCPP

SEQ ID NO: 9 - MB101 variable region light chain (CDRs emphasized)

DIOMTOS PSSVSASVGDRVTITCRASOEISQYLSWYOOKPGKAI KRLIYAASTLDSGVPSR FSGSRSGTDFTLTI SSLOSEDFATYYCLOYASYPPTFGGGTKLE I KGTVAAPSVFI FPPSD EQLKSGTASWCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSK ADYEKHKVYACEVTHQGLSS PVTKS FNRGEC

SEQ ID NO: 10 - Human Igd heavy chain constant region (CH1 -hinge-CH2-CH3; GenBank accession number AAC82527.1 )

STKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSG L YSLSSWTVPSSSLGTQTYI CNVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSV FLFPPKPKDTLMI SRTPEVTCAAA/DVSHEDPEVKFNWYVDGVEVHNAKTKPREEQYNSTYR WSVLTVLHQDWLNGKEYKCKVSNKALPAPI EKTI SKAKGQPREPQVYTLPPSRDELTKNQ VSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGS FFLYSKLTVDKSRWQQGNVF SCSVMHEALHNHYTQKSLSLS PGK

EXAM PLES

Example 1 - Fc receptor binding characteristics of MB101

In this example, different mAb were analyzed with regard to their ability to interact with five human Fey receptors CD16a CD16b/c, CD32a, CD32b and CD64. Two mutations (S241 P and R409K) were introduced into the human lgG ₄ antibody MB101 for stability purposes. MB101 is a recombinant humanized therapeutic monoclonal lgG antibody directed against CD303. MB101 was originally designed as a lgG ₄ therapeutic antibody in order to exhibit no or significantly reduced effector functions. In order to evaluate the ADCC (antibody dependent cell cytotoxicity ) and CDC (complement derived cytotoxicity)- potential via CD16 and CD64, MB101 was characterized for its binding to these receptors. Despite not relevant for ADCC and CDC, MB101 (aCD303-hlgG ₄ -S241 P/R409K) was also tested for its binding to human Fcy receptors CD32a and CD32b/c. As MB101 was intended and designed as humanized lgG ₄ molecule with stabilization, it was expected that MB101 behaves as typical lgG ₄ molecule and shows no affinity or functional interactions with human Fcy receptors like CD16A, CD32a, CD32b/c and CD64.

Fc receptor binding characteristics were analyzed in vitro in comparison to chimeric Igd antibody variants of MB101 by evaluating the binding profile of to recombinant plate bound human recombinant Fcy receptors (FcyR) CD16a, CD32a, CD32b/c and CD64.

Abbrevations

ADCC: antibody dependent cell cytotoxicity

ELISA: enzyme-linked immunosorbent assay

Fc: fragment crystallizable (of an Ab)

HRP: Horse-radish peroxidase

ITAM: immunoreceptor tyrosine-based activation motif

ITIM: immunoreceptor tyrosine-based inhibition motif

mAb: monoclonal antibody

PBS: phosphate buffered saline

RT: room temperature

SD: standard deviation

TMB: 3,3',5,5'-Tetramethylbenzidine

Material / Equipment

MB101 .10 lgG ₄ antibody (SAP 320-002-807, L/N A303P022)

recombinant chimeric anti-CD303 (BDCA-2) Igd wildtype (Cor 6.10.1 1 )

chimeric anti-BDCA2 Igd (FE 060620.02)

MB101 .10 (L/N A303P022) is a humanized lgG ₄ antibody directed against BDCA-2 (CD303) and intended for clinical applications. Beside this version of MB101 , a chimeric anti-BDCA-2 Igd antibody generated in chimeric mice (hybridoma) and a recombinant chimeric anti-CD303 (BDCA-2) lgG1 antibody (AC144 binding region with human Fc region) were used. Mb101 .10 was manufactured in CHO DG44 cells, recombinant chimeric anti-CD303 (BDCA-2) Igd was manufactured in HEK293.EBNA cells and chimeric anti-BDCA-2 Igd antibody was derived from hybridoma cells.

Fcv receptors:

Recombinant Human Fc gamma RI/CD64, R&D Systems, 1257-FC-050

Recombinant Human Fc gamma Rlla/CD32a, R&D Systems, 1330-CD-050

Recombinant Human Fc gamma Rllb/CD32b) R&D Systems, 1875-CD-050

Recombinant Human Fc gamma Rllla/CD16a, R&D Systems, 4325-FC-050 Recombinant Human Fc gamma Rlllb/CD16b, R&D Systems, 1597-FC-050

Method

ELISA plates were coated with Fc gamma receptors I, lla, Mb, Ilia and 1Mb with 5 g/mL each in coating buffer overnight at 4°C. Coated plates were washed 3-fold with 1 x PBS 0.05% Tween 20 and blocked with 5% BSA in 1 x PBS. Antibodies were applied to ELISA plates with 50 L/well at serial 2-fold dilution of antibody solution and incubated at room temperature for 1 h. As blank controls, respective antibody concentrations were also applied to wells lacking coated Fc gamma receptors, in order to substract unspecific binding from resulting signals. After washing (see above), the detection antibody anti-human kappa-HRP (FE 1 10316.07) 1 :4000 in 1xPBS was applied with 50 L/well and incubated again at RT for 1 h. After additional washing (see above), for assay development 50 L/well TMB (CH1 1 1027.37) were applied and stopped after approx. 5 minutes with 50 L/well 10% H ₂ SO ₄ . The developed plate was then analyzed via ELISA reader at 450 nm wavelength.

Results

With regard to the binding to Fey Rl (CD64, High Affinity Fc gamma Receptor), Figure 1 shows that both Igd variants have a high affinity to CD64, as expected for Igd monoclonal antibodies. MB101 has a significantly lower affinity for CD64, which is also not surprising as MB101 is a humanized lgG molecule for which lower affinities to CD64 are reported, compared to Igd molecules. It has to be noted that MB101 still has a minimal affinity to CD64. With regard to the binding to FCY Rlla (CD31 a, Low Affinity Fc gamma Receptor), Figure 2 shows that MB101 has a moderately higher affinity for CD32a compared to the chimeric Igd molecule.

With regard to the binding to Fcy Rllb (CD31 b, Low Affinity Fc gamma Receptor), Figure 3 shows that MB101 has a significantly higher affinity for CD32b receptor compared to the chimeric Igd molecule.

With regard to the binding to Fcy Rllla (CD16a, Low Affinity Fc gamma Receptor), Figure 4 shows that MB101 has a significantly lower affinity for CD16a compared to the chimeric Igd molecule. A minimal residual binding affinity to CD16a can still be found.

With regard to the binding to Fcy Rlllb/c (CD16b/c), Figure 5 shows that MB101 has no affinity for CD16b/c. In comparison, the chimeric lgG1 molecule has a medium binding affinity to CD16b/c.

Conclusions

Clear differences in FcyR interactions could be observed for the different tested receptors. The interactions of MB101 with CD16a, CD16b/c and CD64 were significantly lower than the compared chimeric anti-BDCA-2 Igd . The significantly lower affinity of MB101 for CD16a, CD16b/c and CD64 was an expected design feature of MB101 by the choice of using a human lgG ₄ backbone without effector function. This design feature should reduce antibody interactions with effector cells (e.g. macrophages, monocytes and natural killer cells). As an effect of this low affinity and activation potential, a low or absent ADCC activity potential of MB101 is assumed and the intended design goal was achieved.

Interestingly, MB101 shows a notable affinity to CD32a and CD32b, which is moderately to significantly higher than the binding affinity of the chimeric anti- BDCA-2 Igd and could not be expected from the choice of an human lgG ₄ molecule or from the introduced stabilization mutations that change the lgG ₄ molecule to a slightly Igd-ish human monoclonal antibody backbone. Surprisingly, it was found that the affinity of MB101 for the CD32a and CD32b receptors is even higher than for the Igd control.

The Fcy receptor Fcylla (CD32a) has an activating ITAM (immunoreceptor tyrosine-based activation motif) and is found on monocytes, neutrophils, platelets and dendritic cells. Binding to this receptor may lead to activated cells via the ITAM. The affinity of MB101 to this receptor may therefore induce activation of B cells and dendritic cells. The Fey receptor Fcyllb (CD32b) has an inhibiting ITIM (immunoreceptor tyrosine-based inhibition motif) and is found on B cells and myeloid dendritic cells. Binding to this receptor may lead to inhibited cells via the ITIM. The affinity of MB101 to this receptor may therefore inhibit the antibody production of B cells and inhibit the maturation and cell activation of certain dendritic cells.

[Example 2 - Cell-based in vitro assays for functionality on CD32a and CD32b

The purpose of this set of experiments is to evaluate the activation profile of antibody MB101 on low-affinity human Fey receptors (FcyR) CD16A, CD32a and CD32b/c on recombinant stable transfected reporter cell lines in in vitro assays. The aim of this study is to evaluate the potency of MB101 to activate different FcyR expressing cells upon Fc-part mediated binding. A particular interest was to analyze potential interactions of MB101 with Fc receptors regarding binding affinity and receptor activity upon binding. Both wild-type Igd and lgG ₄ antibodies are used as controls in these assays.

In the human population several different allelic forms of low-affinity FcyRs are expressed, albeit at comparably low frequency. We tested IgG interactions with the following allelic forms of Fey receptors, CD16aV158 (FCYRI I IAV158), CD32aR131 (FcyRIIAR131 ) and CD32b/c (FcyRIIB/C). In case of CD16 two important allelic forms are found, CD16aF158 and CD16aV158 (amino acid sequence numbering according to Nimmerjahn, F., and Ravetch, J. V. 2008. Fc gamma receptors as regulators of immune responses. Nature Reviews Immunology 8: 34). We studied the more abundant allelic form of CD16, CD16aV158. In case of CD32 three different receptors are found that are encoded by three separate gene loci, CD32a, b and c. Although CD32b and CD32c exhibit different biological functions, they express identical extracellular amino acid sequences in amino-acid positions 43-217. Thus, to determine Fey interactions of CD32b and CD32c we used one transfectant expressing the extracellular domains of CD32b. In the following this clone is referred to as CD32b/c. In case of CD32a we expressed the allelic form CD32aR131 , which is broadly found in the human population. For the generation of transfectants the murine thymoma cell line BW5147 was used. For surface expression of the different Fey receptors fusion proteins were engineered consisting of the extracellular domain of the respective human FcyR portion linked with the transmembrane and intracellular ζ-chain of the murine T cell receptor. As BW5147 cells carry an intracellular response element that leads to the production of murine Inteheukin (IL)-2 once a signal via the ζ-chain of the T cell receptor is conveyed, transfectants are expected to secrete murine IL-2 upon interaction of the expressed FcyRs with IgG. The amount of murine IL-2 production is supposed to be directly linked with the intensity of the interaction of IgG with the respective FcyR.

To test the ability of different mAbs to bind to and to cross-link the individual FcyRs, mAbs were coated to plastics and the transfectants were added. If a mAb interacted with a tested FcyR, murine IL-2 production was induced. Absolute IL-2 levels are taken as readout of Fcy-FcyR-interaction. Tests with control mAbs allow comparative evaluation of interaction strength. This experimental setting is suited to study conditions of Fc-FcR-interactions that are also relevant in vivo, when a mAb is bound to a cell-surface target and interacts with Fc-receptors on different immune cells.

Abbreviations

ADCC: antibody dependent cell-mediated cytotoxicity

CD: cluster of differentiation

ELISA: enzyme-linked immunosorbent assay

Fc: fragment crystallizable (of an Ab)

FCS: fetal calf serum

IL: Inteheukin

mAb: monoclonal antibody

PBS: phosphate buffered saline

RT: room temperature

SD: standard deviation FcvR transfectants

BW5147-transfectants were used expressing fusion proteins consisting of the mouse ζ-transmembrane and cytoplasmic domains and the extracellular domains of CD16a, CD32a, or CD32b/c.

Cell culture material for FcvR-expressing transfectants

Cells were passaged every 2-3 days in medium enriched with 0.5 mg/ml Zeocin. Coating of plastic with antibody and assay procedure

mAbs were serially 3-fold diluted in 1 :10 (total log3 dilution) in binding buffer (10 mM Bis-Tris, pH 6) starting with a concentration of 10 g/ml. 100 μΙ of each dilution were transferred to a single well and the coating was performed over night at 4°C. After removal of the coating reagent, 200 μΙ per well blocking buffer (PBS + FCS 10%) were added and incubated for 1 h at RT. After blocking the wells were washed 3 times with 300 μΙ PBS. FcyR-expressing transfectants were transferred into the wells at a concentration of 2x10 ⁵ cells per 200 μΙ cell culture medium. Cells were incubated for 24h at 37°C, 5% CO2. Cell-free supernatant was then collected and analyzed for murine IL-2 using a mouse IL-2 ELISA kit (Bender MedSystems, #BMS601 ) following the manufacturers' instructions. Results

In order to quantify the interaction of different mAbs with human FcyRs, the antibodies TGN1412, Tysabri, Mabthera, Remicade, and MB101 were coated in duplicates (MB101 ) or triplicates (TGN1412, Tysabri, Mabthera, and Remicade) in log3 dilutions ranging from 10 to 0.1 g/ml. TGN1412, Tysabri, and MB101 are lgG mAb, whereas Mabthera and Remicade are IgGi mAb.

As shown in Figure 6, MB101 showed minor interactions with the low affinity FcyR CD16. The interaction of MB101 is comparable with the interaction of the other lgG mAb Tysabri and the IgGi mAb Mabthera. Furthermore, interaction of MB101 with CD16-expressing transfectants is lower than those observed with the lgG ₄ mAb TGN1412.

The mAb MB101 showed very minor interactions with FcyR CD32a expressing transfectants (Fig. 7) and are similar to those found with Tysabri (lgG ₄ ), Remicade (IgGi), TGN1412 (lgG ₄ ) or Mabthera (IgGi). Activation profile data of MB101 on recombinant CD32a reporter cells shows that MB101 has no activation potential on CD32a (but binding potential, see above) and is judged only as binding, but not activating molecule on CD32a.

However, MB101 shows enhanced interaction with CD32b/c-expressing transfectants when compared with the other lgG ₄ mAbs (TGN1412 or Tysabri) and IgGi mAbs (Mabthera or Remicade) (see Fig. 8). Cells were highly activated to secrete IL-2 as indicator for receptor mediated effects. Activation profile data of MB101 on recombinant CD32b reporter cells shows that MB101 has not only a significant binding activity, which is much higher than for lgG and IgGi molecules, but also a significant activation potential on CD32b that is not found at this level for other tested lgG ₄ and Igd molecules. This activation capacity on the inhibitory ITIM receptor CD32b is expected to result in B cell and dendritic cell inhibition.

Conclusions

In this study, different mAb were analyzed with regard to their ability to interact with the low-affinity human FcyR CD16 (FcyRIIIAviss), CD32a (FCYRI IA _R I ₃ I) and CD32b/c (FcyRIIB). The typical high affinity CD64 (FcyRI) was not analyzed, as it would deliver not-comparable results due to the different nature (low vs. high- affinity receptor). Clear differences in FcyR interactions could be observed for the different tested mAb. The interactions of MB101 with CD16- and CD32a- expressing transfectants seemed to be lower than the interactions of TGN1412 and Remicade with CD16 as well as TGN1412 and Mabthera with CD32a. Compared with Igd and lgG ₄ mAbs, MB101 showed a highly enhanced interaction and activation on CD32b/c-expressing transfectants. This activation capacity on the inhibitory ITIM receptor CD32b is expected to result in B cell and dendritic cell inhibition.

Example 3 - FcvR interactions of MB101 and its single mutation variants

Murine transfectants expressing the extracellular domains of human FcyRs were used to assess the multivalent interaction of mAb variants with different FcyRs. In the used system the interaction is measured by the release of murine IL-2 that is secreted into the supernatant after FcyR engagement of the transfectants. mAbs were coated to maxisorb 96well plates at concentrations of 1 pg, 0.3 pg, 0.1 g and 0.03 pg. In addition to MB101 (aCD303-hlgG4-S241 P/R409K), two single mutation variants, Pool 3 (aCD303-hlgG4-S241 P) and Pool 6 (aCD303-hlgG4- R409K), the wild type lgG ₄ antibody (aCD303-hlgG ₄ ), and other lgG ₄ and Igd control mAbs were tested. The antibodies were coated in binding buffer for 2 h at 37°C and then blocked with 1 % BSA for one hour at room temperature. 2x10 ⁵ transfectants were added and incubated for 18 h at 37°C before the supernatant was harvested and tested for murine IL-2 content by an ELISA method.

The results are shown in Figure 9. IL-2 expression levels detected after incubation with the CD16 expressing transfectants revealed that the classical IgGi control antibody induced significant IL-2 expression, whereas the lgG ₄ control, the wt lgG ₄ antibody, and Pool 3 (aCD303-hlgG ₄ -S241 P) induce only minor amounts of IL-2 that further decreased with increased mAb dilution. The data obtained for Pool 6 (aCD303-hlgG ₄ -R409K) and MB101 (aCD303-hlgG ₄ -S241 P/R409K) suggest that there was a trend towards increased IL-2 expression, however, the data were not significant.

The data obtained with the CD32a expressing transfectants were overall rather similar for MB101 (aCD303-hlgG4-S241 P/R409K), the two variants and the wt antibody. The detected IL-2 levels were generally slightly higher than those of the lgG ₄ control but clearly weaker than those of the Igd control. Analyzing the induction of IL-2 after addition of the CD32b transfectant revealed a significant difference between MB101 (aCD303-hlgG ₄ -S241 P/R409K) and Pool 6 (aCD303- hlgG ₄ - R409K) compared to Pool 3 (aCD303-hlgG ₄ -S241 P) and the wt lgG ₄ . This difference was particularly evident at the coating concentration of 0.3 g. This argues that the inclusion of a single mutation in the hinge region (Pool 6; aCD303- hlgG ₄ - R409K) leads to enhanced interaction with CD32b and therefore increased IL-2 production, shifting the lgG ₄ FcyR binding profile towards that of an Igd . The enhanced CD32B interaction of Pool 6 (aCD303-hlgG ₄ - R409K) and MB101 (aCD303-hlgG ₄ -S241 P/R409K) was similarly observed in experiments with the CD64 transfectant.

To more closely analyze the interaction of MB101 (aCD303-hlgG ₄ -S241 P/R409K) with CD32B, the two MB101 variants and the wt antibody were coated at concentrations of 1 g, 0.8 g, 0.6 g, 0.4 g and 0.2 g. The result is shown in Figure 10. Also under such conditions MB101 (aCD303-hlgG ₄ -S241 P/R409K) and Pool 6 (aCD303-hlgG ₄ - R409K) induced significantly more IL-2 production at concentration of 0.8 and 0.6 pg than Pool 3 (aCD303-hlgG ₄ -S241 P) and the wt lgG ₄ . This further verified that a single mutation in the hinge region of the original lgG ₄ antibody does indeed change the FcyR binding profile.

These data was further confirmed using a CD32a/b binding ELISA. To this end ELISA plates were coated with recombinant human FcyRllb (CD32b) protein (R&D Reagents, Cat:1875-CD; Lot: GJZ0814031 ) and binding of MB101 (aCD303- hlgG ₄ -S241 P/R409K), the wt MB101 and the two mutated variants Pool 3 (aCD303-hlgG ₄ -S241 P) and Pool 6 (aCD303-hlgG ₄ -R409K) was probed. Similar to the data obtained from the previous experiment MB101 (aCD303-hlgG ₄ - S241 P/R409K) and Pool 6 variant (aCD303-hlgG ₄ -R409K) showed stronger binding to the recombinant human FCYRIIB/C than WT lgG ₄ mAb and Pool 3 (aCD303-hlgG ₄ -S241 P) as evident from the higher OD ₄₅₀ values (Fig.1 1 ).

To next validate whether the change in the FcyR binding profile of Pool 6 (aCD303-hlgG ₄ -R409K) also influenced the specific target binding of MB101 (aCD303-hlgG ₄ -S241 P/R409K), we made use of a transfectant showing induced GFP expression upon engagement of BDCA-2 (kindly provided by Miltenyi). This allows testing whether the increased binding of MB101 (aCD303-hlgG ₄ - S241 P/R409K) and Pool 6 (aCD303-hlgG ₄ -R409K) to CD32B also leads to more abundant BDCA-2 triggering as measured by a stronger GFP expression in the transfectants compared with the induction of Pool 3 (aCD303-hlgG -S241 P) and the wt antibody. Therefore, 1 x10 ⁵ CD32b or the control CD99 transfectant were seeded in a 96well plate in 100 μΙ medium. MB101 (aCD303-hlgG ₄ -S241 P/R409K), Pool 3 (aCD303- hlgG ₄ -S241 P), Pool 6 (aCD303-hlgG ₄ -R409K) and the wt antibody were added at concentrations of 0.5, 0.1 , 0.05, 0.04, 0.03, 0.02 and 0.01 pg. Then 1x10 ⁵ BDCA-2 transfectants were added and the cells were co-cultured for 24 h.

As shown in Figure 12A, addition of the soluble antibodies to the BDCA-2 transfectants resulted in GFP induction when high concentrations of MB101 were used, while the different antibody variants behaved overall similarly. When additionally the CD99 control transfectant was added BDCA-2 transfectants showed only background GFP expression, whereas co-culture with CD32b transfectants lead to massive GFP induction (Fig. 12B lower row). This expression varied between the antibody variants, especially when concentrations of 0.05 to 0.01 g mAb were used. There one can see that MB101 (aCD303-hlgG - S241 P/R409K) and Pool 6 (aCD303-hlgG ₄ -R409K) induce higher GFP levels than Pool 3 (aCD303-hlgG ₄ -S241 P) or the wt antibody.

To verify this data and to perform statistical analysis, experiments were as shown in Figure 13 repeated. Addition of the highest concentration of the antibodies (0.5 g) in soluble form and without subsequent crosslinking to the BDCA-2 transfectants already sufficed to trigger GFP expression of the transfectants (data not shown). Nevertheless, the extent of GFP expression wanes with lower antibody concentrations markedly. In the co-culture experiments of the CD32b transfectants and the BDCA-2 transfectant incubated with the highest antibody concentrations no significant differences of GFP induction were detected. Nevertheless, there seems to be a trend that MB101 (aCD303-hlgG - S241 P/R409K) and Pool 6 (aCD303-hlgG ₄ -R409K) induce slightly more GFP expression than the other antibodies. This trend gets significant starting at concentrations of 0.05 g of antibody used. This phenomenon can be detected by analyzing the percentage of GFP ⁺ cells (Fig. 13A) as well as the mean fluorescence intensity (MFI) of the GFP signal (Fig. 13B). These results clearly indicate that introducing stabilizing mutation R409K into the hinge region of an lgG ₄ antibody does not only change the FcyR binding profile of the antibody but may also enhances the specific antigen binding capacity by immobilizing the Ab via cell surface FcyR thereby providing functional advantage against WT lgG ₄ variants.

Example 4 - S241 P and R409K stabilize lgG ₄

Formation of half-molecules and Fab arm exchange are inherent phenomena considering lgG ₄ . Mutations in the hinge region, i.e. S241 P, and/or the CH3- domain, i.e. R409K, are analyzed in the present example for their capacity to eliminate dissociation of aCD303-lgG ₄ antibodies into half-molecules and occurrence of Fab arm exchange intermediates.

Antibodies

Buffers and dilutions

Buffer/dilutio composition

n

10x PBS 400.3 g sodium chloride

57.5 g disodium hydrogen phosphate

9.7 g potassium chloride

9.55 g potassium dihydrogen phosphate

ad 5000 mL with H ₂ O

1 x PBS 100 mL I OXPBS

ad 1000 mL with H ₂ O

DTT 1 M 0.154 g DTT

ad 1 mL with H ₂ O

GSH 0.1 M 15.5 mg GSH

ad 1 mL with H2O Abbreviations

SDS-PAGE analysis

First, samples were applied non-boiled and boiled on non-reducing gels.

3.5 g aCD303-lgG -wt, 2.5 g Tysabri, 2.5 g Remicade and 2.5 g AC144 were applied. From all other antibodies, 5 g were put on each gel. Antibodies were loaded native or boiled for 5 minutes at 94 °C. As control, aCD303-lgG -wt was incubated with β-mercaptoethanol to show the completely dissociated heavy and light chains. Samples were put on 4-12 % Tris/Glycin-SDS-gel. Each gel was run at 30 mA, 5 W, 200 V, 1 h 15 min and stained Coomassie.

For non-boiled samples, all antibody bands could be seen around the 170 kDa marker band, corresponding to the whole antibody, which is approx. 150 kDa. aCD303-hlgG ₄ -R409K showed also a high molecular weight contaminating band. This could be due to cultivation of antibody-producing cells in serum-containing media.

After boiling of samples, all antibody bands run above the 170 kDA marker band. Half-molecule formation, characterized by the band between the 72 kDa and 95 kDa marker bands, could only be seen for Tysabri and aCD303-lgG ₄ -wt, comprising an lgG ₄ -wild type hinge-region. Half-molecule formation was significantly decreased for all three lgG variants (faint bands). AC144 and Remicade showed no half-molecule band, representing a stable Igd-phenotype. The high molecular weight contaminating band of aCD303-hlgG -R409K could be detected again. In addition, a band slightly above the 55 kDa marker band could be detected. This was suspected to be a co-purified protein contamination from the serum-containing cell culture.

In a second set of experiments, the antibodies were incubated in the presence of 1 mM DTT. Antibodies were pre-incubated with 1 mM or 1 M DTT for 1 hour at 37 °C. Samples were loaded on 4-12 % Tris/Glycin-SDS-gel. Each gel was run at 30 mA, 5 W, 200 V, 1 h 15 min and stained Coomassie. 3.5 pg aCD303-lgG4-wt, 2.5 g Tysabri, 2.5 g Remicade and 2.5 g AC144 were applied. From all other antibodies, 5 g were put on each gel. As a control, aCD303-lgG ₄ -wt was incubated with β-mercaptoethanol to show the completely dissociated heavy and light chains.

After the incubation of samples with 1 mM DTT, AC144 and Remicade dissociated into H2 and light chains. In contrast, aCD303-lgG ₄ -wt and aCD303-hlgG ₄ -S241 P dissociated completely into light and heavy chains. Tysabri also showed a residual half-molecules band (upper band). lgG ₄ variants R409K and S241 P/R409K showed the same pattern of H2 and light chain bands as the Igd isotype antibodies AC144 and Remicade, demonstrating the desired stabilized phenotype. These results demonstrated, that Igd stability could be restored for the lgG ₄ variants R409K and S241 P/R409K with mutation R409K having the most profound effect.

For all antibodies, DTT concentration was set to 1 M DTT. The pattern of bands was comparable to the treatment with 1 mM DTT for all antibodies except for Tysabri. Tysabri did not show any half-molecules but dissociated completely into heavy and light chains. Altogether, results for the reduction with 1 mM DTT were verified in 4 experiments; results for the reduction with 1 M DTT were verified in 2 experiments.

In a last experimental set-up, antibodies were incubated with 1 mM GSH, according to physiological GSH concentration (Michelet et al., 1995). Hereby, Tysabri and aCD303-hlgG ₄ -wt dissociated into half-molecules, whereas all the other antibodies were stable. 3.5 g aCD303-lgG ₄ -wt, 2.5 g Tysabri, 2.5 g Remicade and 2.5 g AC144 were applied. From all other antibodies, 5 g were put on each gel. Antibodies were put on pre-incubated with 1 mM GSH for 1 hour at 37 °C. Samples were put on 4-12 % Tris/Glycin-SDS-gel and gel was run at 30 mA, 5 W, 200 V, 1 h 15 min. afterwards, gel was stained Coomassie. As a control, aCD303-lgG ₄ -wt was incubated with β-mercaptoethanol to show the completely dissociated heavy and light chains.

This also showed, that stability of aCD303-hlgG ₄ -R409K and aCD303-hlgG ₄ - S241 P/R409K was comparable to Igd antibodies AC144 and Remicade under physiological conditions in vitro.

Fc-Fc-interaction-assav with coated antibodies

To detect, whether antibodies are engaged in Fab arm exchange, assays in ELISA-format are commonly used. Van der Neut Kolfschoten et al. (2007) established an assay, in which plates are coated with one antigen and bispecific antibodies, binding to two antigens, are detected using the other antigen. Rispens et al. (2009) coupled Igd and lgG ₄ to a solid phase and determined binding of either Igd or lgG ₄ . Thereby, they found that Fc-Fc-interactions take place when CH3-domains of lgG ₄ or Igd coupled to a solid phase as well as lgG ₄ in the fluid phase dissociated slightly before binding. Binding of lgG to immobilized lgG is interpreted as an intermediate state of fab arm exchange, which cannot succeed, because one lgG ₄ partner is bound to a solid phase.

For the present example, for each ELISA-plate, 12 mL of coating solution was prepared by diluting an antibody, in general AC144 antibody, to a concentration of 5 g/mL in 1 x PBS. Microplates were coated with 100 L/well of coating solution and sealed with sealing tape on top of microplate to prevent evaporation. Incubation was performed overnight at 4°C. Plates were washed three times with 350 L/well washing buffer with 10 seconds soaking between washes. To block potential nonspecific binding, 200 L/well assay buffer were added and sealed plates were incubated for 2 h at room temperature. Plates could be stored at 4°C for up to one week with blocking solution when they were tightly sealed with sealing tape. Antibodies were diluted to 100 ng/mL using assay buffer. If samples were reduced with DTT or GSH, aliquots were pipetted into microcentrifuge tubes and appropriate volumes of reducing stock solutions were added. Reduction was performed at 37°C for 1 h either in the incubator or in a thermomixer. For the assay, 100 L/well of sample were applied and sealed plates were incubated for 2 h at room temperature. In general, samples were put on as duplicates and assay buffer was used as blank. When incubation was finished, plates were washed again three times with 350 L/well washing buffer. Mouse anti-human-K-HRP, diluted 1 :1 ,000 to 1 :2,000 in assay buffer, was applied and plates were incubated with 100 L/well for 1 h at room temperature. Afterwards, plates were washed three times with 350 L/well washing buffer. For detection, 100 L/well TMB were used and color reaction was stopped after approx. 5 minutes with 100 L/well 10 % sulfuric acid, resulting in a color change from blue to yellow. Absorbance was measured at 450 nm in a microtiter plate reader.

To find out, whether the aCD303-lgG variants could be engaged in Fc-Fc- interactions, ELISA-plates were coated with AC144, and antibodies were incubated with 1 -20 mM DTT in 1 mM-steps. Remicade was also incubated with DTT to include a stable hlgd-antibody as control. Staining for non-stabilized Tysabri was the brightest with highest OD (450 nm) values, followed by humanized CD303-lgG ₄ -wt and lgG ₄ variant S241 P, resulting from binding of applied antibodies to the Fc-parts of coated antibodies. lgG ₄ variants R409K and S241 P/R409K showed lowest staining comparable to Remicade, demonstrating the desired stabilized phenotype. Results were verified in three experiments.

In the next experimental set-up, human \gG ₄ - was coated on ELISA-plates. Again, antibodies were incubated with 1 -10 mM DTT and triplicates were applied. Again, non-stabilized Tysabri showed brightest staining, followed by humanized CD303-lgG ₄ -wt and lgG ₄ variant S241 P. lgG ₄ variants R409K and S241 P/R409K showed lowest staining comparable to Remicade and AC144. OD (450 nm) of these four antibodies was not much higher than blank measurements. Assay buffer, consisting of PBS and BSA, was used as blank. These results demonstrated again, that Igd stability could be restored for the lgG ₄ variants R409K and S241 P/R409K with mutation R409K having the most profound effect. In two orthogonal assays, non-reducing SDS-PAGE analyses and Fc-Fc- interaction-assays, aCD303-lgG ₄ variants R409K and S241 P/R409K exhibited the most stabilized phenotype comparable to the Igd isotype antibodies AC144 and Remicade.

Example 5 - Phase I clinical trial of MB101

MB101 is a humanized anti-BDCA-2 (anti-CD303) monoclonal antibody derived from the murine antibody AC144 using a germ line-based humanization strategy of the variable regions of the light and heavy chain. MB101 is a humanized lgG ₄ with substitutions 409K and 241 P. Binding of BDCA-2 receptor by MB101 suppresses induction of interferon (IFN)-alpha production in plasmacytoid dendritic cells (PDC). Production of IFN-alpha by PDCs is considered to be a major pathophysiological factor in autoimmune diseases and triggering of BDCA-2 is presumed to be a potential therapeutic strategy for blocking production of IFN- alpha in patients.

The production of IFN-alpha by PDCs seems to be the critical event in the induction of psoriatic inflammation and there is experimental evidence that inhibition of this process by ligation of the blood dendritic cell antigen (BDCA)-2 receptor on skin resident PDCs, using the murine predecessor monoclonal antibody AC144, may abrogate the development of psoriatic symptoms [Nestle et al. Plasmacytoid predendritic cells initiate psoriasis through interferon-a production. J Exp Med. 2005 Jul 4; 202(1 ): 135-43]. Current monoclonal antibody therapies for patients with psoriasis are based on the neutralization of TNF-alpha, IL-12/IL-23 and IL-17, with an intention to interrupt the activation cycle by neutralizing central inflammatory cytokines. The MB101 approach is believed to influence an earlier phase of the inflammatory cycle and may therefore be an alternative approach to treat plaque psoriasis by preventing the induction of proinflammatory cytokines downstream of the IFN-alpha induction.

The mode of action of MB101 was investigated using the murine predecessor molecule monoclonal antibody (mAb) AC144 on human PDCs in vitro. Binding of BDCA-2 antigen by MB101 potently suppresses the induction of IFN-α/β production in human PDC, presumably by a mechanism dependent on calcium mobilization and protein-tyrosine phosphorylation by src-family protein-tyrosine kinases [Dzionek et al. BDCA-2, BDCA-3, and BDCA-4: three markers for distinct subsets of dendritic cells in human peripheral blood. J Immunol. 2000 Dec 1 ; 165(1 1 ): 6037-46.; Dzionek et al. BDCA-2, a novel plasmacytoid dendritic cell- specific type II C-type lectin, mediates antigen capture and is a potent inhibitor of interferon alpha/beta induction. J Exp Med. 2001 Dec 17; 194(12): 1823-34.; Dzionek et al. Plasmacytoid dendritic cells: from specific surface markers to specific cellular functions. Hum Immunol. 2002 Dec;63(12):1 133-48.].

Using mAb AC144, the induction of IFN-α/β production was significantly inhibited by the ligation of BDCA-2, even if the IFN-alpha inducing stimulus was added up to 6 days later and was independent from the stimulus used. This indicates that the ligation of BDCA-2 by mAb AC144 may induce a long-lasting inhibition of IFN-α/β production in PDC, i.e. that there is a difference between the pharmacokinetic and pharmacodynamic half-life of mAb AC144 and therefore MB101 in vivo.

In a recent phase I clinical study (M-201 1 -255), safety and tolerability of ascending single i.v. doses of MB101 in patients with psoriasis was assessed. Overall, single doses of MB101 in doses of 0.025 to 32.4 pg/kg body weight were safe and well- tolerated. From dose levels 4.05 pg/kg to 32.4 g/kg (groups 6 to 9), MB101 concentrations could be quantified in serum. Two subjects (one in dose group 6, one in dose group 8) treated with MB101 showed a strong increase in secretion levels of IFN-alpha at 24 h post dose. With the exception of these subjects, for the dose groups 7-9 a decrease in IFN-alpha production in samples from treated subjects could be observed. IFN-alpha levels at 24 h compared to baseline were lower in the subjects treated with MB101 than in those treated with placebo. In the last dose-group (group 9) the reduction of IFN-alpha production upon CpG-A stimulation was significant and showed up to approx. 80% decrease when compared to the pre-dose control sample from the same subject.

The decrease of MFI values of BDCA-2 staining on PBMCs showed clear dose- response, the effect being most pronounced at higher dose levels. After doses of 16.2 pg/kg and 32.4 pg/kg, Mean fluorescence intensity (MFI) values of BDCA-2 staining decreased to about 10% to 20% of the baseline value. The relationship of MFI values of BDCA-2 staining to the dose showed a sigmoidal course. MFI values of BDCA-2 staining on PBMCs decreased with increasing doses. Hence, there was a clear decrease of BDCA-2 staining on PBMCs with increase of dose. The foregoing data makes it plausible that MB101 can be suitably used for the treatment of an autoimmune disease such as psoriasis. LI ST OF REFERENCES

WO 2014/144542 A2

WO 2009/083009

Aalberse and Schuurnnan: lgG4 breaking the rules. (Innnnunology 105 (2002); 9-19)

Bartholomaeus et al.: Cell contact-dependent priming and Fc interaction with CD32+ immune cells contribute to the TGN1412-triggered cytokine response. J Immunol 2014; 5: 2091 -8.

Blank, et al.: Decreased transcription of the human FCGR2B gene mediated by the -343 G/C promoter polymorphism and association with systemic lupus erythematosus. Hum Genet 2005; 2-3: 220-7.

Blomberg et al.: Expression of the Markers BDCA-2 and BDCA-4 and Production of Interferon-alpha by Plasmacytoid Dendritic Cells in Sytsemic Lupus Erythematosus. Arthritis & Rheumatism 2003; 48(9): 2524-2532.

Bloom et al.: Intrachain disulfide bond in the core hinge region of human lgG4. (Protein Science Vol 6 (1997); 407-415).

Chu et al. Inhibition of B cell receptor-mediated activation of primary human B cells by coengagement of CD19 and FcgammaRllb with Fc-engineered antibodies. Molecular Immunology 45(15): 3926-3933 (2008)

Dzionek et al.: BDCA-2, BDCA-3, and BDCA-4: three markers for distinct subsets of dendritic cells in human peripheral blood. J Immunol. 2000 Dec 1 ; 165(1 1 ): 6037-46.

Dzionek et al.: Plasmacytoid dendritic cells: from specific surface markers to specific cellular functions. Hum Immunol. 2002 Dec;63(12):1 133-48.

Dzionek et al.: BDCA-2, a novel plasmacytoid dendritic cell-specific type II C-type lectin, mediates antigen capture and is a potent inhibitor of interferon alpha/beta induction. J Exp Med. 2001 Dec 17; 194(12): 1823-34.

Forrer et al.: Chip-based gel electrophoresis method for for the quantification of half-antibody species in lgG4 and their by- and degradation products. (Anal. Biochem. 334(1 ) (2004):81 -88). Horton et al. Antibody-Mediated Coengagement of FcyRllb and B Cell Receptor Connplex Suppresses Humoral Immunity in Systemic Lupus Erythematosus. J Immunol 201 1 ; 186: 4223-4233.

Labrijn et al.: Therapeutic lgG4 antibodies engage in Fab-arm exchange with endogenous human lgG4 in vivo. (Nat Biotechnol. 27(8) (2009):767-71 ).

Means et al.: Human lupus atoantibody-DNA complexes activate DCs through cooperation of CD32 and TLR9. Journal of Clinical Investigation 2005; 1 15: 407- 417.

Michelet et al.: Blood and Plasma Glutathione Measured in Healthy Subjects by HPLC: Relation to Sex, Aging, Biological Variables, and Life Habits. (Clinical Chemistry Vol. 41 , No. 10 (1995): 1509-17).

Mimoto et al.: Engineered antibody Fc variant with selectively enhanced FcyRllb binding over both FcyRlla ^R131 and FcYRIIa ^H131 . Protein Engineering, Design & Selection 2013; 26(10): 589-598.

Nestle et al.: Plasmacytoid predendritic cells initiate psoriasis through interferon-a production. J Exp Med. 2005 Jul 4; 202(1 ): 135-43.

Niewold, 201 1 . Interferon Alpha as a Primary Pathogenic Factor in Human Lupus. Journal of Interferon & Cytokine Research 31 (12): 887-892.

Nimmerjahn, F. and J. V. Ravetch. 2008. Fey receptors as regulators of immune responses. Nature Reviews Immunology 1 :34-47.

Rispens et al.: Human lgG4 binds to lgG4 and conformationally altered lgG1 via Fc-Fc interactions. J Immunol. 2009;182(7): 4275-4281 .

Salfeld: Isotype selection in antibody engineering. (Nat. Biotechnol., Vol. 25 No. 12 (2007):1369-72).

Schuurman et al.: The inter-heavy chain disulfide bonds of lgG4 are in equilibrium with intra-chain disulfide bonds. (Molecular Immunology 38(2001 ), 1 -8)

Van der Neut Kolfschoten: Anti-inflammatory activity of human lgG4 antibodies by dynamic Fab arm exchange. (Science. 317(5844) (2007):1554-7).

Veri et al.: Therapeutic control of B cell activation via recruitment of Fcgamma receptor lib (CD32B) inhibitory function with a novel bispecific antibody scaffold. Arthritis Rheum. 2010; 62(7): 1933-1943. Vidarsson et al.: IgG Subclasses and Allotypes: From Structure to Effector Functions. Frontiers in Immunology, Vol. 5 (2014).

Vogel et al.: 2015. Antibody induced CD4 down-modulation of T cells specifically mediated by CD64(+) cells. Sci Rep 18308.

Print Out (Original in Electronic Form)

(This sheet is not part of and does not count as a sheet of the international application)

Indications are Made All designations

FOR RECEIVING OFFICE USE ONLY

0-4 This form was received with the

international application: yes

(yes or no)

0-4-1 Authorized officer

Aitken, Jock

FOR INTERNATIONAL BUREAU USE ONLY

0-5 This form was received by the

international Bureau on:

0-5-1 Authorized officer