Fusion proteins comprising a ligand-receptor pair and a biofunctional protein

Abstract

The present disclosure provides fusion proteins having multifunctional biodesigns for programmed target engagement. In certain embodiments, the fusion proteins described herein provide parallel target antigen engagement and immune checkpoint or costimulatory receptor targeting. In certain aspects, the fusion proteins are masked such that they do not exhibit any off-tissue effects associated with target engagement (i.e., toxicity). In certain embodiments, the fusion proteins provide a masked antigen-binding domain and a masked immunomodulatory target-binding domain such that activation of one binding functionality also results in activation of the other binding functionality, thereby generating a bispecific molecule. Accordingly, the present disclosure also provides methods for masking and conditionally activating an antigen-binding domain and targeting and activating an immunomodulatory target in a particular target tissue environment without substantial adverse toxic effects.

Description

Background Art

[0001] With the development of monoclonal antibodies and other biologics as drugs, highly specific and targeted therapeutic agents can be designed. However, the use of these agents is often hampered by the fact that most molecular targets that can identify diseased cells (such as cancer) can also be present in non-diseased (normal) cells in the patient's body, albeit with a certain degree of differential expression. Thus, active-targeted biomolecules may exhibit unexpected activity at locations other than where they are intended to confer therapeutic benefit when used as therapeutic agents, and this can lead to potential toxicity and unwanted side effects. This is known as off-tumor (also known as off-tissue) targeting and affects the dosing regimen as well as the balance between drug efficacy and toxicity. Off-tumor targeting can result in the accidental uptake of therapeutic agents by non-diseased cells and accelerated clearance, leading to adverse pharmacokinetic characteristics of the therapeutic agent, also known as target-mediated drug disposition (TMDD). Thus, in addition to high specificity for molecular targets, these challenges also require therapeutic designs with features that allow for conditional localization of therapeutic agents to diseased cells / tissues while avoiding drug effects on non-tumor tissues expressing the same target.

[0002] Targeting immune checkpoint pathways via positive or negative costimulatory molecules can provide durable therapeutic responses by harnessing the active engagement of the patient's immune system. Unfortunately, checkpoint pathway-targeted therapies can also encounter problems with target-mediated drug toxicity and clearance challenges. There is also a growing awareness that immune responses can be more effectively restored when more than one of these checkpoints and / or costimulatory pathways are co-targeted, or when these checkpoint targets are combined with other non-immune-related targets and therapies. Thus, there is interest in designing therapeutic strategies involving checkpoint targets, but issues related to immune-related adverse events (irAEs), namely toxicity and clearance rates, remain a challenge. Designs that provide conditional engagement of therapeutic agents can offer a less toxic and more effective solution for targeting immunomodulatory molecules. Summary of the Invention

[0003] Disclosed herein are fusion proteins that comprise a biofunctional protein, a ligand-receptor pair, a first peptide linker, and a second peptide linker; wherein the biofunctional protein comprises at least a first polypeptide and a second polypeptide; and wherein the ligand-receptor pair comprises the extracellular portion of an immunoglobulin superfamily (IgSF) receptor and its cognate ligand or a receptor-binding fragment thereof; wherein the ligand is fused to the end of the first polypeptide via the first peptide linker; the receptor is fused to the same corresponding end of the second polypeptide via the second peptide linker; and the first and second peptide linkers are of sufficient length to permit ligand and receptor pairing. In some embodiments, at least one of the first and second peptide linkers comprises a protease cleavage site. In certain embodiments, the ligand is fused to the N-terminus of the first polypeptide via the first peptide linker, and the receptor is fused to the N-terminus of the second polypeptide via the second peptide linker.

[0004] In certain embodiments, the biofunctional protein comprises an antibody or an antigen-binding antibody fragment. In certain embodiments, the biofunctional protein consists of a polypeptide scaffold. In certain embodiments, the polypeptide scaffold is a dimeric Fc region, wherein the first polypeptide consists of a first Fc polypeptide, and the second polypeptide consists of a second Fc polypeptide, and the first and second Fc polypeptides form the dimeric Fc region. In certain embodiments, the biofunctional protein comprises a polypeptide scaffold.

[0005] In certain embodiments, the polypeptide scaffold comprises a dimeric Fc region. In certain embodiments, the dimeric Fc region is a heterodimeric Fc. In certain embodiments, at least one of the ligand or the receptor in the ligand-receptor pair is capable of binding to an immunomodulatory target.

[0006] In some embodiments, the ligand-receptor pair is involved in a cellular response selected from the group consisting of regulation of immune checkpoints, regulation of immune cell activity, regulation of T cell receptor signaling, regulation of T cell-dependent cytotoxicity (TDCC), regulation of antibody-dependent cell phagocytosis (ADCP), and regulation of antibody-dependent cell cytotoxicity (ADCC). In some embodiments, the receptor comprises one or more mutations that increase or decrease the receptor's binding affinity for its cognate ligand as compared to the wild-type receptor.

[0007] In some embodiments, the ligand comprises one or more mutations that increase or decrease the binding affinity of the ligand for its cognate receptor as compared to the wild-type ligand. In certain embodiments, the ligand-receptor pair is selected from the group consisting of: PD1-PDL1, PD1-PDL2, CTLA4-CD80, CD28-CD80, CD28-CD86, CTLA4-CD86, PDL1-CD80, ICOS-ICOSL, NCRSRLG1-NKp30, and CD47-SIRPa. In certain embodiments, the ligand-receptor pair is PD1-PDL1. In certain embodiments, the ligand PDL1 comprises the amino acid sequence according to SEQ ID NO:8. In certain embodiments, the receptor PD1 comprises the amino acid sequence according to SEQ ID NO:9.

[0008] In certain embodiments, the ligand-receptor pair is CTLA4-CD80. In certain embodiments, the ligand CD80 comprises the amino acid sequence according to SEQ ID NO:25, SEQ ID NO:185, SEQ ID NO:187, or SEQ ID NO:189. In certain embodiments, the receptor CTLA4 comprises the amino acid sequence according to SEQ ID NO:26.

[0009] In certain embodiments, the receptor and the ligand are fused to the respective N-termini of the first and second polypeptides. In certain embodiments, one of the first and second peptide linkers comprises more than one protease cleavage site. In certain embodiments, one of the peptide linkers fused to the ligand or the receptor is engineered to comprise one or more additional protease cleavage sites, and wherein the one or more protease cleavage sites in the ligand or the receptor and the protease cleavage sites in the first or second peptide linker can be cleaved by the same protease or different proteases.

[0010] In certain embodiments, the protease is selected from the group consisting of: serine protease, MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, MMP17, MMP18 (collagenase 4), MMP19, MMP20, MMP21, adamalysin, serralysin, astaxanthin, caspase 1, caspase 2, caspase 3, caspase 4, caspase 5, caspase 6, caspase 7, caspase 8, caspase 9, caspase 10, caspase 11, caspase 12, caspase 13, caspase 14, cathepsin A, cathepsin B, cathepsin D, cathepsin E, cathepsin K, cathepsin S, granzyme B, guanidino benzoate enzyme (GB), hepsin, elastase, legumain, proteolytic enzyme, proteolytic enzyme 2, meprin, neurosin, MT-SP1, neprilysin, plasmin, PSA, PSMA, TACE, TMPRSS3, TMPRSS4, uPA, calpain, FAP, and KLK. In certain embodiments, the protease is uPA or proteolytic enzyme.

[0011] In certain embodiments, the length of the peptide linker is 3 - 50 or 5 - 20 amino acids. In certain embodiments, one of the first and second peptide linkers does not have a protease cleavage site. In certain embodiments, the peptide linker is a (Gly n Ser) linker, wherein the (Gly n Ser) linker comprises an amino acid sequence selected from the group consisting of: (Gly3Ser) n (Gly4Ser)1, (Gly3Ser)1(Gly4Ser) n , (Gly3Ser) n (Gly4Ser) n and (Gly4Ser) n , where n is an integer from 1 to 5. In certain embodiments, the peptide linker is (EAAAK) nA linker, where n is an integer between 1 and 5. In certain embodiments, the peptide linker comprises the amino acid sequence EAAAKEAAAK (SEQ ID.NO:38). In certain embodiments, the peptide linker is a polyproline linker, optionally PPP or PPPP. In certain embodiments, the peptide linker comprises an immunoglobulin hinge region sequence, and the immunoglobulin hinge region sequence comprises an amino acid sequence having at most 30% difference in amino acid sequence identity compared to the wild-type immunoglobulin hinge region amino acid sequence. In certain embodiments, the peptide linker comprises a protease cleavage site, and the protease cleavage site comprises the amino acid sequence MSGRSANA (SEQ ID NO:28).

[0012] The present disclosure also describes a fusion protein comprising a Fab region and an Fc region; wherein the Fab region comprises a VH polypeptide and a VL polypeptide forming an antigen-binding domain and a ligand-receptor pair comprising an extracellular portion of an immunoglobulin superfamily receptor and its cognate ligand or a receptor-binding fragment thereof; wherein the ligand is fused to the N-terminus of one of the VH or VL polypeptides via a first peptide linker, and the receptor is fused to the N-terminus of the other VH or VL via a second peptide linker; wherein the first and second peptide linkers have sufficient length to allow pairing of the ligand and the receptor; wherein at least one of the first and second peptide linkers comprises a protease cleavage site; and wherein the ligand-receptor pair sterically hinders the binding of the antigen-binding domain to its cognate antigen.

[0013] In some embodiments, at least one of the first and second polypeptides comprises a first VH polypeptide and a first VL polypeptide, and the first VH and VL polypeptides form a first antigen-binding domain of the antibody, wherein the ligand is fused to one of the first VH or VL polypeptides via the first peptide linker, and the receptor is fused to the other of the first VH or VL polypeptides via the second peptide linker, and wherein the ligand-receptor pair sterically hinders the binding of the first antigen-binding domain to its cognate antigen. In certain embodiments, the first and second polypeptides further comprise a dimeric Fc. In certain embodiments, the dimeric Fc region is a heterodimeric Fc.

[0014] In certain embodiments, the fusion protein comprises ligand-linker-VL, receptor-linker-VL, ligand-linker-VH, or receptor-linker-VH from the N-terminus to the C-terminus.

[0015] In certain embodiments, the fusion protein comprises ligand-cleavable linker-VL, receptor-cleavable linker-VL, ligand-cleavable linker-VH, or receptor-cleavable linker-VH from the N-terminus to the C-terminus.

[0016] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:114)-VL, receptor-linker (SEQ ID NO:114)-VL, ligand-linker (SEQ ID NO:14)-VH, or receptor-linker (SEQ ID NO:14)-VH.

[0017] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:145)-VL, receptor-linker (SEQ ID NO:145)-VL, ligand-linker (SEQ ID NO:145)-VH, or receptor-linker (SEQ ID NO:145)-VH.

[0018] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:147)-VL, receptor-linker (SEQ ID NO:147)-VL, ligand-linker (SEQ ID NO:147)-VH, or receptor-linker (SEQ ID NO:147)-VH.

[0019] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:154)-VL, receptor-linker (SEQ ID NO:154)-VL, ligand-linker (SEQ ID NO:154)-VH, or receptor-linker (SEQ ID NO:154)-VH.

[0020] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:203)-VL, receptor-linker (SEQ ID NO:203)-VL, ligand-linker (SEQ ID NO:203)-VH, or receptor-linker (SEQ ID NO:203)-VH.

[0021] In certain embodiments, at least one of the ligand or the receptor in the ligand-receptor pair is capable of binding to an immunomodulatory target. In certain embodiments, the ligand-receptor pair is involved in a cellular response selected from the group consisting of regulation of immune checkpoints, regulation of immune cell activity, regulation of T cell receptor signaling, regulation of T cell-dependent cytotoxicity (TDCC), regulation of antibody-dependent cell phagocytosis (ADCP), and regulation of antibody-dependent cell cytotoxicity (ADCC).

[0022] In certain embodiments, the receptor comprises one or more mutations that increase or decrease the binding affinity of the receptor for its cognate ligand as compared to the wild-type receptor. In certain embodiments, the ligand comprises one or more mutations that increase or decrease the binding affinity of the ligand for its cognate receptor as compared to the wild-type ligand. In certain embodiments, the ligand-receptor pair is selected from the group consisting of: PD1-PDL1, PD1-PDL2, CTLA4-CD80, CD28-CD80, CD28-CD86, CTLA4-CD86, PDL1-CD80, ICOS-ICOSL, NCRSRLG1-NKp30, and CD47-SIRPa. In certain embodiments, the ligand-receptor pair is PD1-PDL1. In certain embodiments, the ligand PDL1 comprises the amino acid sequence according to SEQ ID NO:8. In certain embodiments, the receptor PD1 comprises the amino acid sequence according to SEQ ID NO:9. In certain embodiments, the ligand-receptor pair is CTLA4-CD80. In certain embodiments, the ligand CD80 comprises the amino acid sequence according to SEQ ID NO:25. In certain embodiments, the receptor CTLA4 comprises the amino acid sequence according to SEQ ID NO:26.

[0023] In some embodiments, the receptor and the ligand are fused to the respective N-termini of the first and second polypeptides. In certain embodiments, one of the first and second peptide linkers comprises more than one protease cleavage site. In certain embodiments, one of the ligand or the receptor is engineered to comprise one or more additional protease cleavage sites, and wherein the one or more protease cleavage sites in the ligand or the receptor and the protease cleavage site in the first or second peptide linker are cleavable by the same protease or by different proteases.

[0024] In certain embodiments, the protease is selected from the group consisting of: serine protease, MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, MMP17, MMP18 (collagenase 4), MMP19, MMP20, MMP21, adamalysin, serralysin, astaxanthin, caspase 1, caspase 2, caspase 3, caspase 4, caspase 5, caspase 6, caspase 7, caspase 8, caspase 9, caspase 10, caspase 11, caspase 12, caspase 13, caspase 14, cathepsin A, cathepsin B, cathepsin D, cathepsin E, cathepsin K, cathepsin S, granzyme B, guanidino benzoic acid enzyme (GB), hepsin, elastase, legumain, proteolytic enzyme, proteolytic enzyme 2, meprin, neurosin, MT-SP1, neprilysin, plasmin, PSA, PSMA, TACE, TMPRSS3, TMPRSS4, uPA, calpain, FAP, and KLK. In certain embodiments, the protease is uPA or proteolytic enzyme. In certain embodiments, the length of the peptide linker is 3 - 50 or 5 - 20 amino acids. In certain embodiments, one of the first and second peptide linkers does not have a protease cleavage site. In certain embodiments, the peptide linker is a (Gly n Ser) linker, wherein the (Gly n Ser) linker comprises an amino acid sequence selected from the group consisting of: (Gly3Ser) n (Gly4Ser)1, (Gly3Ser)1(Gly4Ser) n , (Gly3Ser) n (Gly4Ser) n and (Gly4Ser) n , where n is an integer from 1 to 5. In certain embodiments, the peptide linker is (EAAAK) nA linker, where n is an integer between 1 and 5. In certain embodiments, the peptide linker without a protease cleavage site comprises the amino acid sequence EAAAKEAAAK (SEQ ID.NO:38). In certain embodiments, the peptide linker is a polyproline linker, optionally PPP or PPPP. In certain embodiments, the linker is a glycine (G) - proline (P) polypeptide linker, optionally GPPPG, GGPPPGG, GPPPPG or GGPPPGG. In certain embodiments, the peptide linker comprises an immunoglobulin hinge region sequence, and the immunoglobulin hinge region sequence comprises an amino acid sequence having at most 30% difference in amino acid sequence identity compared to the wild - type immunoglobulin hinge region amino acid sequence. In certain embodiments, the peptide linker containing a protease cleavage site comprises the amino acid sequence MSGRSANA (SEQ ID NO:28).

[0025] In certain embodiments, the binding of the first antigen - binding domain to its cognate antigen is reduced by 10 - fold or more compared to the parental antigen - binding domain not fused to the ligand - receptor pair. In certain embodiments, cleavage of the protease cleavage site in a cellular environment releases a member of the ligand - receptor pair from the fusion protein, thereby allowing the antigen - binding domain to bind its cognate antigen.

[0026] In certain embodiments, the first antigen - binding domain is a Fab. In certain embodiments, the first antigen - binding domain binds an antigen expressed on a cancer cell or an immune cell. In certain embodiments, the first antigen - binding domain binds an antigen expressed on a T cell. In certain embodiments, the first antigen - binding domain binds to a tumor - associated antigen (TAA). In certain embodiments, the first antigen - binding domain binds to an antigen selected from the group consisting of cluster of differentiation 3 (CD3), human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), mesothelin (MSLN), tissue factor (TF), cluster of differentiation 19 (CD19), tyrosine - protein kinase Met (c - Met), cluster of differentiation 40 (CD40), and cadherin 3 (CDH3).

[0027] In certain embodiments, the antibody or antibody fragment comprises a second antigen-binding domain, the second antigen-binding domain comprising a second VH polypeptide and a second VL polypeptide. In certain embodiments, the fusion protein comprises a second ligand-receptor pair, wherein the ligand of the second ligand-receptor pair is fused via a third peptide linker to one of the second VH or VL polypeptides, and the receptor of the second ligand-receptor pair is fused via a fourth peptide linker to the other of the second VH or VL polypeptides, wherein at least one of the third and fourth peptide linkers comprises a protease cleavage site, and wherein the ligand-receptor pair sterically hinders binding of the second antigen-binding domain to its cognate antigen. In certain embodiments, the fusion protein binds to two different antigens. In certain embodiments, one antigen is an antigen expressed by T cells and the other antigen is an antigen expressed by cancer cells. In certain embodiments, the fusion protein binds to CD3 and HER2.

[0028] Also described herein is a fusion protein comprising an Fc region and a ligand-receptor pair, the Fc region comprising a first Fc polypeptide and a second Fc polypeptide, and the ligand-receptor pair comprising the extracellular portion of an immunoglobulin superfamily receptor and its cognate ligand or a receptor-binding fragment thereof; wherein the ligand is fused via a first peptide linker to the terminus of the first Fc polypeptide, and the receptor is fused via a second peptide linker to the same corresponding terminus of the second Fc polypeptide; wherein the first and second peptide linkers have sufficient length to allow ligand and receptor pairing; and wherein at least one of the first and second peptide linkers comprises a protease cleavage site. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] These and other features, aspects, and advantages of the present invention will become better understood with reference to the following description and appended drawings, in which:

[0030] Figure 1(A) shows a schematic diagram of the structure of certain fusion proteins described herein. By fusing PD-1 (squares) and PD-L1 (stripes) to the N-termini of the heavy and light chains, respectively, the complementarity-determining regions (CDRs) of the Fab (gray) can be sterically blocked by the Ig superfamily heterodimer formed between the two. After removing one side of the mask via TME-specific proteolytic cleavage (bolt) from one of the linkers introduced into the linker between the masking domain and the Fab, a portion of the mask can be released and binding to the target can be restored. In addition, the portion of the mask that remains covalently attached to the Fab adds functionality by binding to its immunomodulatory partner. Figure 1(B) shows a schematic diagram of an antibody with two Fab arms masked using IgSF domain pairs attached to the N-terminus with TME protease-cleavable or non-cleavable linkers. The Fab CDRs a-TAA 1 and a-TAA 2 can be the same or different, and the IgSF pairs 1:2 and 3:4 can be the same or different. Figure 1(C) shows a schematic diagram of a Fab x scFv construct with a Fab arm specific for target 1 and a scFv arm specific for target 2. The Fab arm and its binding to target 1 are masked with IgSF domain pairs attached to the N-terminus with TME protease-cleavable or non-cleavable linkers.

[0031] Figure 2 Figure 4 shows a schematic diagram of the modified bispecific CD3 x Her2 Fab x scFv Fc fusion protein described herein. One arm of the antibody-like molecule contains an anti-CD3 Fab blocked by a PD-1 / PD-L1 mask, while the other arm contains an anti-Her2 scFv.

[0032] Figure 3 shows the UPLC-SEC chromatograms and non-reduced and reduced CE-SDS profiles of representative bispecific CD3 x Her2 Fab x scFv Fc variants. (A) UPLC-SEC chromatogram of the de-masked variant 30421, (B) non-reduced (left) and reduced (right) CE-SDS profiles of the de-masked variant 30421, (C) UPLC-SEC chromatogram of the masked non-cleavable variant 30423, (D) non-reduced (left) and reduced (right) CE-SDS profiles of the masked non-cleavable variant 30423, (E) UPLC-SEC chromatogram of the masked light-chain cleavable variant 30430, (F) non-reduced (left) and reduced (right) CE-SDS profiles of the masked light-chain cleavable variant 30430, (G) UPLC-SEC chromatogram of the masked heavy-chain cleavable variant 30436, (H) non-reduced (left) and reduced (right) CE-SDS profiles of the masked heavy-chain cleavable variant 30436.

[0033] Figure 4 Overlay of DSC thermograms showing the unmodified (30421) and PD-1:PD-L1 masked variants (30430, 30436) of the studied CD3 x Her2 Fab x scFv Fc system.

[0034] Figure 5 Reduced CE-SDS profiles of representative variants showing untreated (-uPa) and uPa-treated (+uPa) at a 1:50 uPa:variant ratio for 24 h at 37 °C. Profiles of the de-masked variant (30421), masked but non-cleavable variant (30423), and masked cleavable variants (30430, 30436, 31934) are shown.

[0035] Figure 6 Initial binding results of variants targeting CD3 to Jurkat cells as determined by ELISA. Results of the de-masked variant (30421), constructs with only the attached PD-L1 or PD-1 moiety (31929, 31931), and variants with an intact non-cleavable mask (30423) or an intact mask and a cleavable PD-L1 or PD-1 moiety (30430, 30436) are shown. For samples of variants 30423, 30430, 30436, untreated (-uPa) and uPa-treated (+uPa) samples were tested.

[0036] Figure 7 Cell killing of pan T cells against JIMT-1 tumor cells as determined in the TDCC assay after treatment with variants of engineered cross-linked T cells and tumor cells. Results of the de-masked variant (30421), variant with only the PD-1 moiety attached to the heavy chain (31929), and variants with an intact non-cleavable mask (30423) or an intact mask and a cleavable PD-L1 moiety on the light chain (30430) are shown. For variant 30430, untreated (-uPa) and uPa-treated (+uPa) samples were tested. An irrelevant anti-RSV antibody (22277) was used as a negative control.

[0037] Figure 8 shows the results of initial binding studies of selected CD3-targeting variants to (A) PD-L1-transfected CHO-S cells and (B) PD-1-transfected CHO-S cells by flow cytometry. Results are shown for the de-masked variant (30421), constructs with only the attached PD-L1 or PD-1 moiety (31929, 31931), and variants with an intact non-cleavable mask (30423, 30426) or with an intact mask and a cleavable PD-L1 or PD-1 moiety (30430, 30436). Also included is an Fc fusion of an affinity-matured PD-1 moiety (31829). For samples of variants 30423, 30426, 30430, 30436, samples untreated (-uPa) and treated with uPa (+uPa) were tested.

[0038] Figure 9 shows a schematic of a hybrid PD-1 / PD-L1 reporter assay that probes crosslinking of T cells and JIMT-1 cells and blockade of the PD-1:PD-L1 checkpoint engagement (A) and analysis of both (B). Results are shown for the de-masked variant (30421) and a combination of the same de-masked variant with an excess of anti-PD-L1 antibody (30421 + 150 nM anti-PD-L1). Also studied were constructs with only the PD-1 moiety attached to the heavy chain (31929) and variants with an intact non-cleavable mask (30423) or with an intact mask and a cleavable PD-L1 moiety on the light chain (30430). For variant 30430, samples untreated (-uPa) and treated with uPa (+uPa) were tested. An irrelevant anti-RSV antibody (22277) was used as a negative control. Measurements were made in triplicate and error bars reflecting the standard deviation are shown.

[0039] Figure 10 is a figure representing a modified monospecific bivalent fusion protein targeting a tumor-associated antigen (TAA). The paratope of the Fab is sterically blocked by a PD-1 / PD-L1 mask.

[0040] Figure 11 shows UPLC-SEC chromatograms (A-J) and non-reducing SDS-PAGE (K) or non-reducing and reducing CE-SDS profiles (L) of masked fusion proteins targeting EGFR, MSLN, TF, CD19, cMet, CDH3. For all fusion proteins, data for non-cleavable variants (31722, 31728, 31736, 31732, 28647, 28662) are shown, while for EGFR, MSLN, TF, and CD19, samples of cleavable variants (31723, 31729, 31737, 31733) are also included.

[0041] Figure 12 shows the reducing SDS-PAGE profiles of representative fusion proteins targeting (A) EGFR, (B) MSLN, (C) TF, and (D) CD19. Samples without uPa treatment (-uPa) and samples with uPa treatment (+uPa) were studied. For each system, data for uPa non-cleavable variants (31722, 31728, 31736, 31732) and variants with a u-Pa cleavable sequence between the VL and PD-L1 moieties (31723, 31729, 31737, 31733) are shown.

[0042] Figure 13 shows the initial flow cytometry binding results of selected fusion proteins targeting different antigens to the following cell lines expressing the antigen: (A) EGFR on MDA-MB-468, (B) MSLN on OVCAR3, (C) TF on MDA-MB-231, (D) CD19 on Raji, (E) cMet on EBC1, (F) CDH3 on JIMT1. For all systems, data for non-cleavable variants (31722, 31728, 31736, 31732, 28647, 28662) are shown, while for EGFR, MSLN, TF, and CD19, samples of cleavable variants (31723, 31729, 31737, 31733) were also included and tested in the absence (-uPa) and presence (+uPa) of uPa. For all systems, unmodified controls (32474, 16427, 16417, 6323, 4372, 17606, 17214) and irrelevant controls for cMet and CDH3 (22277) were also included. Available (EGFR, MSLN, TF) data from SPR were included for comparison.

[0043] Figure 14 Results of growth inhibition studies of NCI-H292 cells treated with variants targeting EGFR are shown. Data for the de-masked variant (32474) and the PD-1:PD-L masked variants are shown. Masked variants include non-cleavable forms (31722) as well as forms with a cleavable PD-L1 moiety on the light chain (31723). Irrelevant controls (22277) were also included. For all variants, samples were tested in the absence (-uPa) and presence (+uPa) of treatment. Error bars reflect the standard deviation of triplicate measurements.

[0044] Figure 15Schematic representation of the engineered bispecific CD3 x Her2 Fab x scFv Fc variants studied here. One arm of the fusion protein contains an anti-CD3 Fab blocked by a CD80 / CTLA4 mask, while the other arm contains an anti-Her2 scFv.

[0045] Figure 16 shows the UPLC-SEC chromatogram and non-reduced and reduced CE-SDS profiles of variant 30444. (A) UPLC-SEC chromatogram of the masked light-chain-cleavable variant 30444, (B) non-reduced (left) and reduced (right) CE-SDS profiles of the masked light-chain-cleavable variant 30444, (C) non-reduced (left) and reduced (right) CE-SDS profiles of the masked light-chain-cleavable variant 30444, (D-F) UPLC-SEC chromatograms of the masked light-chain-cleavable variants 33525, 33526, 33527 after protein A purification.

[0046] Figure 17 Shows the reduced CE-SDS profiles of variant 30444 in the absence (-uPa) and presence (+uPa) of uPa treatment.

[0047] Figure 18 Shows the initial binding results of CD3-targeting variants to Jurkat cells as determined by ELISA. Results are shown for the de-masked variant (30421), the variant with a mask based on intact PD-1 / PD-L1 and a cleavable PD-L1 moiety (30430), and the variant with a mask based on intact CD80 / CTLA4 and a cleavable CTLA4 moiety (30444). For samples of variants 30430 and 30444, samples without uPa treatment (-uPa) and samples with uPa treatment (+uPa) were tested.

[0048] Figure 19 Schematic representation of the IgV of an immunomodulatory pair (e.g., PD-1:PD-L1) fused to a heterodimeric IgG Fc via a hinge. Cleavage of one of the two linkers by a TME-related protease (such as uPa) releases one moiety (e.g., PD-L1) and leaves the moiety with the desired function (e.g., PD-1) still attached to the Fc and available for binding to its cellular partner. In the case of PD-1, it is able to bind PD-L1 on the target cell and inhibit checkpoint function.

[0049] Figure 20 shows the UPLC-SEC chromatograms of CD40-targeting variants (A-C) and the non-reduced and reduced CE-SDS profiles of CD40-targeting variants (D). Also shown are the (E) reduced CE-SDS, (F) flow cytometry binding data, and (G) results from the CD40 RGA assay of the same variants without uPa treatment (-uPa) and with uPa treatment (+uPa). Test articles included the de-masked variant (32477), the variant with an uncleavable PD-1 / PD-L1-based mask (32478), and the variant with a PD-1 / PD-L1-based mask where the PD-L1 moiety is removable by cleavage with uPa (32479). In the functional study (G) conducted via the RGA assay, the initial CD40 binding partner CD40L and an irrelevant control (v22277) were also included. The data from the CD40 RGA assay are summarized in the table in (H).

[0050] Figure 21 (A) PD1 and PDL1 consist of immunoglobulin domains that form a complex. In this image, the binding Fab docks to the PD1-PDL1 complex at the complementarity-determining end. Connecting PD1 and PDL1 to the VH and VL chains using an appropriate linker can block antigen binding. (B) Structures of other exemplary pairs of immunomodulators that can be used as masks: PD-1 / PD-L1 (PDB: 4ZQK), PD-1 / PD-L2 (PDB: 3BP5), CTLA4 / CD86 (PDB: 1I85), NCRSRLG1 / NKp30 (PDB: 3PV6), SIRPa / CD47 (PDB: 4KJY), CTLA4 / CD80 (PDB: 1I8L).

[0051] Figure 22 Shows the initial binding results of CD3-targeting variants to pan T cells as determined by flow cytometry. Results are shown for the de-masked variant (30421), the anti-CD3 single-arm antibody (18560), the construct with only the attached PD-1 moiety (31929), and the variant with a complete uncleavable mask (30423) or the variant with a complete mask and a cleavable PD-L1 moiety (30430, 30436). For samples of variants 30423, 30430, samples without uPa treatment (-uPa) and samples treated with upa (+uPa) were tested. Data for an irrelevant control (22277) are also shown.

[0052] Figure 23A and Figure 23BShows the cell killing of pan T cells against HCC1954, JIMT-1, HCC827, and MCF-7 tumor cells as determined in two replicates of the TDCC assay after treatment with engineered cross-linked T cell and tumor cell variants. Results are shown for the de-masked variant (30421) and the combination of the de-masked variant with a saturating amount of anti-PD-L1 antibody (30421 + 120 nM atezolizumab), the variant with only the PD-1 portion attached to the heavy chain (31929), and the variant with an intact non-cleavable mask (30423) or the variant with an intact mask and a cleavable PD-L1 portion on the light chain (30430). For variants 30430 and 30423, samples untreated (-uPa) and treated (+uPa) with uPa were tested. An irrelevant anti-RSV antibody (22277) was used as a negative control.

[0053] Figure 24 Shows the IFNγ release of pan T cells as determined in two replicates of the TDCC assay of HCC1954, JIMT-1, HCC827, and MCF-7 cancer cells after treatment with engineered cross-linked T cell and tumor cell variants. Results are shown for the de-masked variant (30421) and the combination of the de-masked variant with a saturating amount of anti-PD-L1 antibody (30421 + 120 nM atezolizumab), the variant with only the PD-1 portion attached to the heavy chain (31929), and the variant with an intact non-cleavable mask (30423) or the variant with an intact mask and a cleavable PD-L1 portion on the light chain (30430). For variants 30430 and 30423, samples untreated (-uPa) and treated (+uPa) with uPa were tested. An irrelevant anti-RSV antibody (22277) was used as a negative control.

[0054] Figure 25 Shows the receptor number of each Her2 and PD-L1 cell for the set of cancer cell lines used in the TDCC and RGA assays as determined by flow cytometry.

[0055] Figures 26A to 26DResults of a hybrid PD-1 / PD-L1 reporter assay are shown that probe T cell crosslinking and blockade of PD-1:PD-L1 checkpoint engagement in four different cancer cell lines (HCC1954, JIMT-1, HCC827, MCF-7). Results are shown for the de-masked variant (30421) and a combination of the de-masked variant with a saturating amount of anti-PD-L1 antibody (30421 + 150 nM atezolizumab), a variant having only the PD-1 portion attached to the heavy chain (31929), and variants having an intact non-cleavable mask (30423) or an intact mask and a cleavable PD-L1 portion located on the light chain (30430). For variant 30430, samples untreated (-uPa) and treated with uPa (+uPa) were tested. An irrelevant anti-RSV antibody (22277) was used as a negative control.

[0056] Figure 27 Figure represents a modified monospecific bivalent fusion protein targeting EGFR (a-EGFR). The paratope of the Fab is sterically blocked by the SIRPα / CD47 mask.

[0057] Figure 28 shows the (A) UPLC-SEC chromatogram and (B) non-reduced and reduced CE-SDS profiles of a fully cleavable SIRPa / CD47 masked variant (34164) targeting EGFR. (C) Also shown is the reduced CE-SDS of the same variant without uPa treatment (-uPa) and with uPa treatment (+uPa).

[0058] Figure 29 Results of an initial binding assay on EGFR-positive H292 cells by high content analysis are shown. Test articles included a de-masked EGFR-targeting control (v32474), a fully cleavable SIRPa / CD47 masked variant targeting EGFR without uPa treatment (-uPa) and with uPa treatment (+uPa) (34164), and an irrelevant control (v22277).

[0059] Figure 30 shows (A) data from a single titration point (1 nM) in a flow cytometry binding experiment from Her2+ / PD-L1+ JIMT-1 cells and (B) data from a bridging experiment from human pan T cells and Her2+ / PD-L1+ JIMT-1 cells. Data are shown for a trispecific variant having only the PD-1 portion attached to the heavy chain (v31929) and bispecific variants having the same format but unable to bind to PD-L1 or Her2 (v32497 and v33551, respectively). Data for an irrelevant control (v22277) are included in the bridging assay (B).

[0060] Figure 31 shows the mechanism of T cell recruitment and activation by the PD-1:PD-L1 masked CD3 x Her2 Fab x scFv Fc variant. (A) Therapeutic antibodies directed to the tumor microenvironment (TME) via TAA binding. (B) The PD-L1 portion of the masker is released via cleavage by TME-specific proteases. (C) The activated therapeutic agent binds and activates T cells via the de-masked a-CD3 paratope for tumor cell killing and inhibits checkpoint activity by binding to PD-L1 on the heavy chain cells.

[0061] Figure 32 Shows the results of the initial binding of CD3-targeted variants to pan T cells as determined by flow cytometry. Results are shown for the de-masked variant (30421), the construct with only the attached PD-1 portion (31929), and the variant with a non-functional PD-1 domain attached to the heavy chain (32497). Data for an irrelevant control (22277) are also shown.

[0062] Figure 33 Shows the cell killing of JIMT-1 tumor cells by pan T cells as determined in a TDCC assay after treatment with variants of engineered cross-linked T cells and tumor cells. Results are shown for the de-masked variant (30421), the variant with only the PD-1 portion attached to the heavy chain (31929), and the variant with a non-functional PD-1 domain attached to the heavy chain (32497).

[0063] Figure 34A Is a schematic diagram of the IgSF core Ig fold, which shows a β-sandwich composed of 7 anti-parallel β-strands arranged into two 3-strand and 4-strand β-sheets. Figure 34B Is a schematic diagram showing the IgSF IgC1 subgroup domain (upper panel) and the IgC2 subgroup domain (lower panel) that differ in chain arrangement. Figure 34C and Figure 34D Is a schematic diagram of the IgV domain, which contains 9 β-strands arranged into two 4-strand and 5-strand sheets. Detailed Description

[0064] Definition

[0065] The terms used in the claims and the specification are briefly defined herein and are defined in more detail below.

[0066] "Fusion protein" refers to a protein that contains more than one polypeptide region or domain, such as those connected to each other by peptide bonds. Thus, "fusion" as used herein refers to polypeptide sequences connected to each other by peptide bonds. Examples include antibodies or scaffolds fused to an immune regulatory ligand / receptor pair. The fusion proteins described herein are sometimes referred to as "variants" or "constructs".

[0067] "Biologically functional protein" generally refers to a polypeptide or protein with biological functions, such as an antibody, such as a dimeric Fc.

[0068] "Ligand-receptor pair" refers to a receptor polypeptide and a ligand polypeptide that specifically bind to each other. Examples include PD-1-PD-L1, CTLA4-CD80, or CD28-CD80.

[0069] "Receptor-binding fragment" refers to any polypeptide that specifically binds to the receptor of a ligand-receptor pair. The receptor-binding fragment can be naturally occurring or non-naturally occurring.

[0070] "Immune regulatory" molecule refers to a molecule that has the ability to directly or indirectly regulate an immune response (e.g., upregulation or downregulation of an immune response) and / or the activity of immune cells.

[0071] "Peptide linker" refers to a peptide that links or connects other peptides or polypeptides.

[0072] The terms "Fc region", "Fc", and "Fc domain" are used interchangeably herein and refer to the C-terminal region of an immunoglobulin heavy chain that contains at least a portion of the constant region.

[0073] "Bispecific" refers to a biologically functional protein that can specifically bind two different epitopes.

[0074] "Multispecific" refers to a biologically functional protein that can specifically bind at least two or more different target molecules or epitopes.

[0075] "Masked" means that the binding of a polypeptide domain (e.g., the antigen-binding domain of an antibody) to a target sequence is sterically hindered, or the binding of a ligand to its cognate binding partner (e.g., its receptor) is sterically hindered.

[0076] "Protease-activated" or "protease-cleaved" or "cleaved" refers to a fusion protein that contains a protease cleavage site after being cleaved by a protease.

[0077] "Protease cleavage site" refers to an amino acid sequence within a fusion protein that contains a protease recognition sequence and is cleaved by a protease.

[0078] "Immune checkpoint" refers to an immune system regulatory pathway that regulates the activation of the immune system.

[0079] When referring to the binding of a particular antigen, epitope, ligand or receptor, "specifically binds" (and grammatical variants thereof) means binding that is measurably different from non-specific interactions.

[0080] As described in more detail below, "mammal" includes both human and non-human, and includes but is not limited to humans, non-human primates, canines, felines, murine, bovines, equines, and porcines.

[0081] It must be noted that, unless the context clearly dictates otherwise, as used in the specification and the appended claims, the singular forms "a", "an", and "the" include plural referents.

[0082] Abbreviations used in this application include the following: PD-1 (programmed cell death protein 1); PDL-1 (programmed death ligand 1); CD3 (cluster of differentiation 3); CTLA4 (cytotoxic T lymphocyte-associated protein 4 or cluster of differentiation 152); CD80 (cluster of differentiation 80); CD28 (cluster of differentiation 28); CD86 (cluster of differentiation 86); ICOS (inducible T cell co-stimulator); ICOS-L (inducible T cell co-stimulator ligand); CD47 (cluster of differentiation 47); SIRPA (signal regulatory protein α), HHLA2 (human endogenous retrovirus-H long terminal repeat-associated protein 2), NKp30 (natural killer cell receptor 3), NCR3LG1 (natural killer cell cytotoxic receptor 3 ligand 1), HHLA2 (HERV-H LTR-associated protein 2), VISTA (V-domain Ig suppressor of T cell activation), VTCN1 (V-set domain-containing suppressor of T cell activation 1), CD276 (cluster of differentiation 276), human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), mesothelin (MSLN), tissue factor (TF), cluster of differentiation 19 (CD19), tyrosine-protein kinase Met (c-Met), and cadherin-3 (CDH3).

[0083] As used herein, the term "about" means a variation of approximately + / - 10% from a given value. It is to be understood that such variations are always included in any given value provided herein, whether or not specifically mentioned.

[0084] As used herein, the terms "comprising," "having," "including," and "containing" and their grammatical variants are inclusive or open-ended and do not exclude additional, unrecited elements and / or method steps. When used herein in connection with a composition, use, or method, the term "consisting essentially of" means that additional elements and / or method steps may be present, but these additions do not materially affect the manner in which the recited composition, method, or use functions. The term "consisting of," when used in connection with a composition, use, or method herein, does not include the presence of additional elements and / or method steps. A composition, use, or method described herein as comprising certain elements and / or steps may also, in certain embodiments, consist essentially of those elements and / or steps, and in other embodiments, consist of those elements and / or steps, whether or not such embodiments are specifically recited.

[0085] It is contemplated that any embodiment discussed herein can be implemented by any method, use, or composition disclosed by the present invention, and vice versa.

[0086] It should also be understood that an affirmative statement of a feature in one embodiment is a basis for excluding that feature in another embodiment. In particular, in cases where a list of options is presented for a given embodiment or claim, it should be understood that one or more of the options can be removed from the list, and the shortened list can form an alternative embodiment, whether or not such alternative embodiments are specifically recited.

[0087] The various amino acid sequences and cloned sequences mentioned herein are shown in Table AA.

[0088] Fusion protein

[0089] Disclosed herein are fusion proteins that comprise a biofunctional protein fused to a ligand-receptor pair, e.g., an antibody or a polypeptide scaffold. In the fusion proteins according to the present disclosure, the biofunctional protein comprises at least a first polypeptide and a second polypeptide, and the ligand is fused to the end of one of the polypeptides via a first peptide linker, and the receptor is fused to the same corresponding end of the other polypeptide via a second peptide linker. In some embodiments, at least one of the first and second peptide linkers comprises a protease cleavage site that is naturally present in a target cell environment, e.g., in a tumor microenvironment. Also disclosed are methods of using the fusion proteins disclosed herein.

[0090] The fusion proteins according to the present disclosure are masked to reduce any off-tissue (e.g., off-tumor) effects (i.e., toxicity) associated with target engagement. Cleavage of one or more peptide linkers containing protease cleavage sites in the target cell environment results in de-masking of the fusion protein. In certain embodiments, the fusion proteins according to the present disclosure comprise a polypeptide scaffold fused to a ligand-receptor pair. In this context, the fusion protein is masked because each of the ligand and receptor in the ligand-receptor pair is impeded from engaging its native receptor or ligand through their association with each other. Cleavage of one or more peptide linkers containing protease cleavage sites in the target cell environment results in de-masking of the fusion protein by releasing one member of the ligand-receptor pair from the fusion protein, thereby allowing the other member of the ligand-receptor pair to bind its native partner. Thus, in certain embodiments, the present disclosure provides a biological design for programmed checkpoint or co-stimulatory receptor targeting.

[0091] In certain embodiments, the fusion proteins according to the present disclosure comprise an antibody or an antigen-binding antibody fragment that comprises an antigen-binding domain fused to a ligand-receptor pair. In this context, the fusion protein is masked such that the ligand-receptor pair sterically impedes the antigen-binding domain from binding its native antigen. The fusion protein is further masked such that each of the ligand and receptor in the ligand-receptor pair is impeded from engaging its native receptor or ligand through their association with each other. Cleavage of one or more peptide linkers containing protease cleavage sites in the target cell environment results in de-masking of the fusion protein by releasing one member of the ligand-receptor pair from the fusion protein, thereby allowing both the other member of the ligand-receptor pair to bind its native partner and the antigen-binding domain to bind its native antigen. Thus, in certain embodiments, the present disclosure provides a multifunctional biological design for programmed target antigen engagement and synchronous checkpoint or co-stimulatory receptor targeting. In certain aspects, the design of the fusion proteins described herein reduces target-mediated drug disposition. In certain embodiments, the fusion protein provides a masked antigen-binding domain (e.g., a biofunctional protein) and a masked immunomodulatory target-binding domain (e.g., a ligand-receptor pair) such that the programmed activation of one binding functionality also results in the activation of the other binding functionality, thereby generating a bifunctional molecule. Thus, in certain embodiments, the present disclosure provides methods for masking and conditionally activating an antigen-binding domain in a specific target tissue environment, and targeting and activating an immunomodulatory target with reduced adverse toxic effects.

[0092] Ligand-receptor pair

[0093] Described herein are fusion proteins, each comprising a ligand-receptor pair. In some aspects, the ligand-receptor pair is an immunomodulatory pair of ligand-receptor domains belonging to the immunoglobulin superfamily (IgSF) (Natarajan, Kannan; Mage, Michael G; and Margulies, David H (April 2015) Immunoglobulin Superfamily. In: eLS. John Wiley & Sons, Ltd: Chichester., A F Williams 1, A N Barclay (1988) The Immunoglobulin Superfamily--Domains for Cell Surface Recognition Annu Rev Immunol 6:381-405).

[0094] The immunoglobulin superfamily (IgSF) classifies domains commonly found in proteins based on the core immunoglobulin (Ig) fold. This Ig fold consists of a β-sandwich composed of a total of 7 antiparallel β-strands arranged into two 3-strand and 4-strand β-sheets ( Figure 34A ). The two β-sandwiches are interconnected via a disulfide bridge between strand B and strand F. The structural motif commonly identified in the Ig fold is the "Greek key" motif. Common subgroups of the IgSF are IgV, IgC1, and IgC2 domains. Members are identified based on common structural features and the arrangement of β-strands. The IgC domain contains 7 β-strands arranged into two 3-strand and 4-strand sheets ( Figure 34B ), while the IgV domain contains 9 β-strands arranged into two 4-strand and 5-strand sheets ( Figure 34C 、 Figure 34D ). IgC1 and IgC2 differ in the structural arrangement of the strands. IgSF domains can be found in a variety of biologically significant proteins including antigen receptors, immunoglobulins, and immunomodulatory receptors. The surface-exposed residues of the core β-sandwich and the loops connecting the β-strands can serve as interaction interfaces for antigen recognition, tertiary / quaternary assemblies, or other domains in receptor / ligand pairs. Since the antigen recognition site of an immunoglobulin (the VH-VL pair in an antibody such as IgG1) contains a dimer of two IgV domains, dimers of IgSF or IgV domains are structurally compatible to form a steric mask of the antigen recognition site upon covalent attachment to the N-terminus of an antibody (Figure 21).

[0095] In certain embodiments, the ligand-receptor pair is immunomodulatory, e.g., is an immune checkpoint, causing regulation of immune cell effector function, regulation of T cell receptor signal transduction, and regulation of the interaction between antigen-presenting cells and effector cells or combinations thereof. In certain embodiments, the ligand-receptor pair comprises the extracellular portion of an IgSF receptor and its cognate ligand or a receptor-binding fragment thereof. A receptor-binding fragment refers to any polypeptide that specifically binds to the receptor of the ligand-receptor pair and can be naturally occurring or non-naturally occurring. As used herein and as applied to an object, "naturally occurring" refers to the fact that the object can be found in nature. For example, a polypeptide or polynucleotide sequence that is present in an organism isolated from a natural source and has not been deliberately modified by a person in the laboratory is naturally occurring. In certain embodiments, the ligand-receptor pair can be two interacting protein domains belonging to the immunoglobulin domain superfamily. As used herein, "non-naturally occurring" refers to an engineered polypeptide sequence that has structural similarity to the IgSF, such as a mutant of a naturally occurring protein.

[0096] In certain embodiments, the disclosure herein relates to the use of an immunomodulatory pair of ligand-receptor domains belonging to the IgSF as a mask for an antibody or antibody fragment, thereby hindering target antigen binding. Examples of immunomodulatory pairs of ligand-receptor domains belonging to the immunoglobulin superfamily include, but are not limited to, pairs of the B7 / CD28 family (such as PD1-PDL1, PD1-PDL2, CTLA4-CD80, CD28-CD80, CD28-CD86, CTLA4-CD86, PDL1-CD80, and ICOS-ICOSL, NCR3LG1-NKp30, HHLA2-CD28H, and CD47-SIRP. CD80 (also known as B7-1), CD86 (B7-2), PDL1 (B7-H1), ICOSL (B7-H2), PDL2 (B7-DC), CD276 (B7-H3), VTCN1 (B7-H4), VISTA (B7-H5), NCR3LG1 (B7-H6), HHLA2 (B7-H7) belong to the B7 family. The B7 protein family is generally considered to be ligands and pair with members of the CD28 family, which include CD28, CTLA4, CD28H, NKp30, PD1, and ICOS. (S.M. West and X.A. Deng. Considering B7-CD28 as a family through sequence and structure. Exp Biol Med (Maywood) 2019;244(17):1577-1583; doi:10.1177 / 1535370219855970).

[0097] In certain embodiments, the ligand-receptor pair comprises members of the IgSF B7 / CD28 family. In certain embodiments, the ligand and receptor comprise the extracellular portions of immunoglobulin superfamily (IgSF) polypeptides. In certain embodiments, the ligand and receptor comprise the extracellular portions of IgSF immunoglobulin variable (IgV) polypeptides. In certain embodiments, the ligand is a member of the IgSF B7 family and the receptor is a member of the IgSF CD28 family.

[0098] In certain embodiments, the ligand-receptor pair comprises a leukocyte co-stimulatory receptor. Examples of leukocyte co-stimulatory receptors belonging to the B7 / CD28 family include ICOS (also known as CD278) and CD28. Examples of co-stimulatory ligand-receptor pairs include CD80:CD28, CD86:CD28, and ICOS:ICOSL (ICOS ligand). Examples of co-inhibitory ligand-receptor pairs include PD1-PDL1, PD1-PDL2, CTLA4-CD80, CTLA4-CD86, PDL1-CD80, and CD47-SIRPα. When linked to the N-terminus of the Fab, our results described herein show that they block access to the CDR and thus block binding to the antigen ( Figure 21A ).

[0099] Other members of this large IgSF can be used in a similar manner and have immunomodulatory functions. Figure 21B A representation showing the known structures of known B7-CD28 members. The sizes and orientations of the domains of other pairs are very similar to those of PD-1 and PD-L1, and thus they can be used for binding or functional blocking similar to the PD-1 / PD-L1 receptor-ligand pair.

[0100] The concept of functional masks extends beyond members of the B7 family. For example, Figure 21BA representation of the structure of SIRPα / CD47 (another ligand-receptor pair with domains belonging to the IgSF) is shown, which exhibits good spatial compatibility, is located at the N-terminus of the Fab and blocks binding. Many therapeutic candidates are being evaluated for the use of antagonists in this axis to increase the phagocytosis of cancer cells, making it a good candidate for a functional mask. (Murata Y, Saito Y, Kotani T, Matozaki T. (2018) CD47-signal regulatory protein α signaling system and its application to cancer immunotherapy. Cancer Sci. August 2018; 109(8):2349-2357).

[0101] In certain embodiments, the affinity of the ligand-receptor domains in the ligand-receptor pair of the fusion protein is altered as compared to the wild-type ligand and receptor. In certain embodiments, one or both of the ligand-receptor domains in the masking pair are engineered such that the ligand and receptor comprise sequences different from the wild-type ligand or receptor. In certain embodiments, the ligand comprises one or more mutations that increase the binding affinity of the ligand for its cognate receptor. In certain embodiments, the relative binding affinity of the ligand of the ligand-receptor pair as compared to the wild-type ligand is greater than 1, 1.5, 2, 2.5, 3, 5, 10, 20, 30, 40, 50, 100, 500, 1000, 5,000, 10,000, 50,000, or 100,000 times the relative binding affinity of the wild-type ligand for its naturally occurring cognate receptor.

[0102] In certain embodiments, the receptor comprises one or more mutations that increase the binding affinity of the receptor for its cognate ligand. In certain embodiments, the relative binding affinity of the receptor of the ligand-receptor pair as compared to the wild-type receptor is greater than 1, 1.5, 2, 2.5, 3, 5, 10, 20, 30, 40, 50, 100, 500, 1000, 5,000, 10,000, or 100,000 times the relative binding affinity of the wild-type receptor for its naturally occurring cognate ligand.

[0103] In certain embodiments, the ligand comprises one or more mutations that reduce the binding affinity of the ligand for its cognate receptor. In certain embodiments, the relative binding affinity of the ligand of the ligand-receptor pair compared to the wild-type ligand is reduced by more than 1, 1.5, 2, 2.5, 3, 5, 10, 20, 30, 40, 50, 100, 500, 1000, 5,000, 10,000, 50,000, or 100,000-fold compared to the relative binding affinity of the wild-type ligand for its naturally occurring cognate receptor.

[0104] In certain embodiments, the receptor comprises one or more mutations that reduce the binding affinity of the receptor for its cognate ligand. In certain embodiments, the relative binding affinity of the receptor of the ligand-receptor pair compared to the wild-type receptor is reduced by more than 1, 1.5, 2, 2.5, 3, 5, 10, 20, 30, 40, 50, 100, 500, 1000, 5,000, 10,000, or 100,000-fold compared to the relative binding affinity of the wild-type receptor for its naturally occurring cognate ligand.

[0105] The ligand-receptor pair can be, for example, the IgV domains of PD-L1 (Uniprot ID Q9NZQ7, 33-146) and PD-1 (Uniprot ID Q15116, 18-132). In some embodiments, the ligand is PD-L1 and has, for example, an amino acid sequence corresponding to SEQ ID NO: 8 or SEQ ID NO: 10. In certain embodiments, the PD-L1 has an amino acid sequence substantially identical to SEQ ID NO: 8. In certain embodiments, the PD-L1 has an amino acid sequence about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 8. In certain embodiments, the PD-L1 has an amino acid sequence about 96%, 97%, 98%, or about 99% identical to SEQ ID NO: 8. Any PD-L1 variant known in the art can be used, such as high-affinity variants, for example, those provided by Z. Laing et al., High-affinity human PD-L1 variants attenuate the suppression of T cell activation; Oncotarget 8 , 88360-88375 (2017) or those provided in WO2018 / 170021A1. In certain embodiments, the receptor is a high-affinity PD-L1 variant. In some embodiments, the receptor is a high-affinity PD-L1 variant having an amino acid sequence corresponding to SEQ ID NO: 10 or an amino acid sequence substantially identical to SEQ ID NO: 10.

[0106] In some embodiments, the receptor is PD-1 and has an amino acid sequence corresponding to, for example, SEQ ID NO: 7 or 11. In certain embodiments, the PD-1 has an amino acid sequence that is substantially identical to SEQ ID NO: 7 or 11. In certain embodiments, the PD-1 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 7 or 11. In certain embodiments, the PD-1 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 7 or 11. Any PD-1 variant known in the art can be used, such as high-affinity variants, for example, those provided in R.L. Maute et al., Engineering high-affinity PD-1 variants for optimized immunotherapy and immuno-PET imaging. Proc Natl Acad Sci U S A 112, E6506-6514 (2015), WO2016 / 022994A2 or E. Lazar-Molnar et al., Structure-guided development of a high affinity human Programmed Cell Death-1: Implications for tumor immunotherapy EBIOMedicine 17. 30-44 (2017) and WO2019 / 241758A1.

[0107] In certain embodiments, the receptor is a high-affinity PD-1 variant. In some embodiments, the receptor is a high-affinity PD-1 variant having an amino acid sequence corresponding to SEQ ID NO: 9 or an amino acid sequence that is substantially identical to SEQ ID NO: 9.

[0108] In certain embodiments, the ligand is CD80 and has, for example, an amino acid sequence corresponding to SEQ ID NO:25. In certain embodiments, the CD80 has an amino acid sequence that is substantially identical to SEQ ID NO:25. In certain embodiments, the CD80 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:25. In certain embodiments, the CD80 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:25. In some embodiments, the CD80 has an amino acid sequence that is substantially identical to SEQ ID NO:185, SEQ ID NO:187, or SEQ ID NO:189. In certain embodiments, the CD80 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:185, SEQ ID NO:187, or SEQ ID NO:189. In certain embodiments, the CD80 has a mutation that increases its affinity for its receptor or decreases its tendency to form homodimers during preparation. In certain embodiments, the CD80 has an amino acid sequence corresponding to SEQ ID NO:25 with one of the following sets of mutations: (a) H18Y, A26E, E35D, M47S, I61S, and D90G; (b) E35D, M47S, N48K, I61S, K89N; (c) E35D, D46V, M47S, I61S, D90G, K93E; or (d) H18Y, A26E, E35D, M47S, I61S, V68M, A71G, D90G.

[0109] In certain embodiments, the ligand is PD-L2 and has, for example, an amino acid sequence corresponding to SEQ ID NO:250. In certain embodiments, the PD-L2 has an amino acid sequence that is substantially identical to SEQ ID NO:250. In certain embodiments, the PD-L2 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:250. In certain embodiments, the PD-L2 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:250.

[0110] In certain embodiments, the ligand is CD86 and has, for example, an amino acid sequence corresponding to SEQ ID NO:248. In certain embodiments, the CD86 has an amino acid sequence that is substantially identical to SEQ ID NO:248. In certain embodiments, the CD86 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:248. In certain embodiments, the CD86 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:248.

[0111] In certain embodiments, the ligand is ICOS-L and has, for example, an amino acid sequence corresponding to SEQ ID NO:256. In certain embodiments, the ICOS-L has an amino acid sequence that is substantially identical to SEQ ID NO:256. In certain embodiments, the ICOS-L has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:256. In certain embodiments, the ICOS-L has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:256.

[0112] In certain embodiments, the ligand is B7-H3 and has, for example, an amino acid sequence corresponding to SEQ ID NO:258. In certain embodiments, the B7-H3 has an amino acid sequence that is substantially identical to SEQ ID NO:258. In certain embodiments, the B7-H3 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:258. In certain embodiments, the B7-H3 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:258.

[0113] In certain embodiments, the ligand is B7-H4 and has, for example, an amino acid sequence corresponding to SEQ ID NO:259. In certain embodiments, the B7-H4 has an amino acid sequence that is substantially identical to SEQ ID NO:259. In certain embodiments, the B7-H4 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:259. In certain embodiments, the B7-H4 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:259.

[0114] In certain embodiments, the ligand is VISTA and has, for example, an amino acid sequence corresponding to SEQ ID NO:260. In certain embodiments, the VISTA has an amino acid sequence that is substantially identical to SEQ ID NO:260. In certain embodiments, the VISTA has an amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO:260. In certain embodiments, the VISTA has an amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO:260.

[0115] In certain embodiments, the ligand is HHLA2 and has, for example, an amino acid sequence corresponding to SEQ ID NO:262. In certain embodiments, the HHLA2 has an amino acid sequence that is substantially identical to SEQ ID NO:262. In certain embodiments, the HHLA2 has an amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO:262. In certain embodiments, the HHLA2 has an amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO:262.

[0116] In certain embodiments, the ligand is SIRPα and has, for example, an amino acid sequence corresponding to SEQ ID NO:255. In certain embodiments, the SIRPα has an amino acid sequence that is substantially identical to SEQ ID NO:255. In certain embodiments, the SIRPα has an amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO:255. In certain embodiments, the SIRPα has an amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO:255.

[0117] In some embodiments, the receptor is CTLA4 and has, for example, an amino acid sequence corresponding to SEQ ID NO:26. In certain embodiments, the CTLA4 has an amino acid sequence that is substantially identical to SEQ ID NO:26. In certain embodiments, the CTLA4 has an amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO:26. In certain embodiments, the CTLA4 has an amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO:26.

[0118] In some embodiments, the receptor is CD28 and has, for example, an amino acid sequence corresponding to SEQ ID NO:253. In certain embodiments, the CD28 has an amino acid sequence that is substantially identical to SEQ ID NO:253. In certain embodiments, the CD28 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:253. In certain embodiments, the CD28 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:253.

[0119] In some embodiments, the receptor is CD28H and has, for example, an amino acid sequence corresponding to SEQ ID NO:263. In certain embodiments, the CD28H has an amino acid sequence that is substantially identical to SEQ ID NO:263. In certain embodiments, the CD28H has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:263. In certain embodiments, the CD28H has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:263.

[0120] In some embodiments, the receptor is NKp30 and has, for example, an amino acid sequence corresponding to SEQ ID NO:264. In certain embodiments, the NKp30 has an amino acid sequence that is substantially identical to SEQ ID NO:264. In certain embodiments, the NKp30 has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:264. In certain embodiments, the NKp30 has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:264.

[0121] In some embodiments, the receptor is ICOS and has, for example, an amino acid sequence corresponding to SEQ ID NO:257. In certain embodiments, the ICOS has an amino acid sequence that is substantially identical to SEQ ID NO:257. In certain embodiments, the ICOS has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:257. In certain embodiments, the ICOS has an amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:257.

[0122] In certain embodiments, the IgSF ligand and / or receptor has an immunoglobulin variable domain (IgV)-like structure. Amino acid sequences of some exemplary naturally occurring IgV domain receptors and ligands described herein are shown in Table CC.

[0123] In certain embodiments, the engineered non-naturally occurring but paired ligand and / or receptor of the ligand-receptor pair comprises an immunoglobulin domain, wherein at least one domain in the domain has an affinity for a naturally occurring immunomodulatory receptor.

[0124] In certain embodiments, the immunomodulatory ligand-receptor pair is selected to act as an antagonist or agonist of its cognate target pair. In certain embodiments, the immunomodulatory ligand-receptor pair is selected to act as an antagonist or agonist of its cognate target pair in the tumor microenvironment. In certain embodiments, one or both of the ligand or receptor in the ligand-receptor pair is designed to function after protease cleavage activation.

[0125] Fusion protein format

[0126] The fusion proteins described herein can be in many different formats. The fusion protein can be considered to have a modular architecture that includes at least a ligand-receptor pair, wherein each of the ligand and receptor is fused to a biofunctional protein via a peptide linker. The biofunctional protein further comprises at least a first polypeptide and a second polypeptide. For example, the N-terminus or C-terminus of the ligand or receptor of the ligand-receptor pair can be fused to the first polypeptide and the second polypeptide of the biofunctional protein via a peptide linker, respectively. The ligand is fused to the first polypeptide, and the receptor is fused to the same corresponding terminus of the second polypeptide. When describing a ligand-receptor pair fused to a polypeptide, the term "same corresponding terminus" means that each of the ligand and receptor is fused to the N-terminus or the C-terminus of the first and second polypeptides. Thus, in certain embodiments, the ligand is fused to the N-terminus of the first polypeptide via a first peptide linker, and the receptor is fused to the N-terminus of the second polypeptide via a second peptide linker. In certain embodiments, the ligand is fused to the C-terminus of the first polypeptide via a first peptide linker, and the receptor is fused to the C-terminus of the second polypeptide via a second peptide linker. The ligand and receptor can be fused via their C-termini or their N-termini. Both the ligand and receptor can be fused via their N-termini or C-termini, or one of the ligand or receptor can be fused via its N-terminus and the other of the ligand or receptor can be fused via its C-terminus.

[0127] In certain embodiments, the N-terminus of the ligand is fused to the N-terminus of a first polypeptide via a first peptide linker, and the N-terminus of the receptor is fused to the N-terminus of a second polypeptide via a second peptide linker. In certain embodiments, the C-terminus of the ligand is fused to the C-terminus of a first polypeptide via a first peptide linker, and the C-terminus of the receptor is fused to a second polypeptide via a second peptide linker.

[0128] In certain embodiments, the ligand is fused to the terminus of a first polypeptide of a biofunctional protein via a first peptide linker that comprises a protease cleavage site. In certain embodiments, the receptor is fused to the terminus of a second polypeptide of a biofunctional protein via a second peptide linker that comprises a protease cleavage site. In certain embodiments, the ligand is fused to the terminus of a first polypeptide of a biofunctional protein via a first peptide linker that comprises a protease cleavage site, and the receptor is fused to the terminus of a second polypeptide of a biofunctional protein via a second peptide linker that comprises a protease cleavage site. When both the first and second peptide linkers comprise protease cleavage sites, the protease cleavage sites may be capable of being cleaved by the same protease or they may be capable of being cleaved by different proteases.

[0129] In certain embodiments, the ligand is fused to the terminus of a first polypeptide of a biofunctional protein via a first peptide linker that comprises a protease cleavage site, and the ligand is engineered to comprise an internal protease cleavage site that may be the same as or different from the cleavage site in the first peptide linker. In certain embodiments, the receptor is fused to the terminus of a second polypeptide of a biofunctional protein via a second peptide linker that comprises a protease cleavage site, and the receptor is engineered to comprise an internal protease cleavage site that may be the same as or different from the cleavage site in the first peptide linker. The inclusion of protease cleavage sites in the peptide linker and in the members of the ligand-receptor pair that are linked to the biofunctional protein via that linker allows for the cleavage and inactivation of that member of the ligand-receptor pair in the target cell environment while the member of the ligand-receptor pair that remains fused to the bioactive protein is unmasked (i.e., is conditionally activated).

[0130] In certain embodiments, the fusion protein is conjugated to another therapeutic agent and / or diagnostic moiety, such as a chemotherapeutic agent or a radioisotope.

[0131] Biofunctional protein

[0132] Biologically functional proteins can act as scaffolds and / or contain binding domains. Examples of polypeptide scaffolds include immunoglobulin Fc regions, albumin, albumin analogs and derivatives, toxins, cytokines, chemokines, growth factors, and protein pairs such as leucine zipper domains. In certain embodiments, the biologically functional protein comprises a label, a drug, or a combination thereof. Any label known in the art suitable for detecting the fusion proteins described herein can be used. The biologically functional protein can include any drug, toxin, or chemical known in the art capable of conjugating to a protein and achieving a desired biological result.

[0133] In certain embodiments, the biologically functional protein of the fusion proteins described herein comprises at least one antigen-binding domain. The binding domain can be, for example, an immunoglobulin-based binding domain or a non-immunoglobulin antibody mimetic, or other polypeptides or small molecules capable of specifically binding to its target, such as a native or engineered ligand. Non-immunoglobulin antibody mimetic formats include, for example, anticalin, Fynomer, affimer, alphabody, DARPins, and Avimer.

[0134] The fusion proteins described herein include a biologically functional protein. Examples of biologically functional proteins include, but are not limited to, antibodies, for example, polypeptides having an antigen-binding domain, and polypeptide scaffolds, for example, dimeric Fc. Thus, in certain embodiments, the first and second polypeptides of the biologically functional protein are polypeptides comprising variable and / or constant domains of an antibody or other domains conferring antigen-binding or scaffold function to the fusion protein.

[0135] Antibody

[0136] In certain embodiments, the biologically functional protein is an antibody, i.e., an immunoglobulin. Antibodies according to the present disclosure can be in a variety of formats as described herein, including antibody fragments. Thus, in certain embodiments, the biologically functional protein is an antibody fragment. The terms "antibody" and "immunoglobulin" are used interchangeably herein to refer to polypeptides encoded by one or more immunoglobulin genes or modified forms of immunoglobulin genes that specifically bind to an antigen.

[0137] Specific binding can be measured, for example, by enzyme-linked immunosorbent assay (ELISA), surface plasmon resonance (SPR) technology (using, for example, a BIAcore instrument) (Liljeblad et al., 2000, Glyco J, 17:323 - 329), or traditional binding assays (Heeley, 2002, Endocr Res, 28:217 - 229). In certain embodiments, specific binding is defined as the degree of binding to an unrelated protein being less than about 10% of the binding to the target antigen as measured, for example, by SPR. In certain embodiments, specific binding of an antibody or antibody fragment to a specific antigen or epitope is defined as a dissociation constant (K D ) ≤ 1 μM, for example, ≤ 100 nM, ≤ 10 nM, ≤ 1 nM, ≤ 0.1 nM, ≤ 0.01 nM, or ≤ 0.001 nM. In certain embodiments, specific binding of an antibody or antibody fragment to a specific antigen or epitope is defined as a dissociation constant (K D ) of 10 -6 M or less, for example 10 -7 M or less, or 10 -8 M or less. In some embodiments, specific binding of an antibody or antibody fragment to a specific antigen or epitope is defined as a dissociation constant (K D ) between 10 -6 M and 10 -13 M, for example, between 10 -7 M and 10 - 13 M, between 10 -8 M and 10 -13 M, or between 10 -9 M and 10 -13 M.

[0138] Conventional immunoglobulin structural units typically consist of two pairs of polypeptide chains, each pair having one "light" chain (about 25 kD) and one "heavy" chain (about 50 - 70 kD). Light chains are classified as κ or λ. The "class" of an immunoglobulin refers to the type of constant domain possessed by its heavy chain. There are five main classes of antibodies: IgA, IgD, IgE, IgG, and IgM, and several of these classes can be further divided into subclasses (isotypes), for example, IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2. The heavy chain constant domains corresponding to the different classes of immunoglobulins are called alpha (α), delta (δ), epsilon (ε), gamma (γ), and mu (μ), respectively.

[0139] In certain embodiments, the antibodies described herein are based on IgG-class immunoglobulins, such as IgG1, IgG2, IgG3, or IgG4 immunoglobulins. In some embodiments, the antibodies described herein are based on IgG1, IgG2, or IgG4 immunoglobulins. In some embodiments, the antibodies described herein are based on IgG1 immunoglobulins. In the context of the present disclosure, when an antibody is based on a particular immunoglobulin isotype, it means that the antibody contains all or part of the constant region of the particular immunoglobulin isotype. It should be understood that in some embodiments, the antibody may also contain hybrids of isotypes and / or subclasses.

[0140] The N-terminal domain of each polypeptide chain of an immunoglobulin defines a variable region of about 100 to 110 or more amino acids that is primarily responsible for antigen recognition. The terms "variable light chain (VL)" and "variable heavy chain (VH)" refer to these domains in the light and heavy chains, respectively.

[0141] Thus, it can be seen that immunoglobulins contain distinct domains within the heavy and light chains. Such domains can overlap and include the Fc domain (or Fc region), CH1 domain, CH2 domain, CH3 domain, hinge domain, heavy chain constant domains (CH1-hinge-Fc or CH1-hinge-CH2-CH3), heavy chain variable domain (VH), light chain variable domain (VL), and light chain constant domain (CL). The "Fc domain" includes the CH2 and CH3 domains, and optionally the hinge domain (or hinge region).

[0142] In each VH and VL domain of an immunoglobulin, there are three loops that are hypervariable in sequence and form the antigen-binding site. Each of these loops is referred to as a "hypervariable region" or "HVR". The terms hypervariable region (HVR) and complementarity-determining region (CDR) are used interchangeably herein to refer to the portions of the variable region that form the antigen-binding domain. Except for CDR1 in VH, CDRs generally contain the amino acid residues that form the hypervariable loops. The VH and VL domains are composed of relatively invariant stretches called framework regions (FRs) that are between about 15 and 30 amino acids in length, separated by the shorter CDRs, each of which is typically between about 5 and 15 amino acids in length, but can occasionally be longer or shorter. The three CDRs and four FRs that make up each VH and VL domain are arranged from the N-terminus to the C-terminus as follows: FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4.

[0143] Many different definitions of CDR regions are commonly used, including those described by Kabat et al. (1983, Sequences of Proteins of Immunological Interest, NIH Publication No. 369-847, Bethesda, MD), Chothia et al. (1987, J Mol Biol, 196:901-917), and the IMGT, AbM, and Contact definitions. These different definitions include the superposition or subsets of amino acid residues when compared to each other. For example, the CDR definitions according to Kabat, Chothia, IMGT, AbM, and Contact are provided in Table 1 below. Thus, as will be apparent to those skilled in the art, the exact numbering and placement of CDRs can vary based on the numbering system employed. However, it should be understood that the disclosure herein of the heavy chain variable domain (VH) includes the disclosure of the relevant (inherent) heavy chain cDRs (HCDRs) as defined by any known numbering system. Similarly, the disclosure herein of the variable light domain (VL) includes the disclosure of the relevant (inherent) heavy chain CDRs (HCDRs) as defined by any known numbering system.

[0144] Table 1: General CDR Definitions 1

[0145]

[0146] 1 For all definitions except Contact which uses Chothia numbering, the Kabat or Chothia numbering system can be used for HCDR2, HCDR3, and light chain CDRs

[0147] 2 Use Kabat numbering. The position that differentiates the end of the Chothia and IMGT CDR-H1 loops in the Kabat numbering scheme changes according to the length of the loop because Kabat made insertions at positions 35A and 35B outside of those CDR definitions. The IMGT and Chothia CDR-H1 loops can be defined explicitly using Chothia numbering. The CDR-H1 definition using Chothia numbering is: Kabat H31-H35, Chothia H26-H32, AbM H26-H35, IMGT H26-H33, Contact H30-H35.

[0148] One of ordinary skill in the art will recognize that a limited number of amino acid substitutions can be introduced into the CDR sequences, or VH or VL sequences, of a known antibody without the antibody losing its ability to bind to its target. Candidate amino acid substitutions can be identified by computer modeling or by techniques such as alanine scanning as described above, and the binding activity of the resulting variants can be tested by standard techniques. For example, in certain embodiments, the EGFR-binding domain of the fusion protein comprises a set of CDRs (i.e., heavy chain CDR1, CDR2, and CDR3 and light chain CDR1, CDR2, and CDR3) that have 90% or higher, 95% or higher, 98% or higher, 99% or higher, or 100% sequence identity to the set of CDRs from cetuximab or panitumumab, wherein the binding domain retains the ability to bind EGFR. In certain embodiments, the EGFR-binding domain of the fusion protein comprises variants of these CDR sequences that comprise between 1 and 10 amino acid substitutions spanning the three CDRs (i.e., the CDRs can be modified by incorporation of up to 10 amino acid substitutions in any combination with the modified CDR), for example, between 1 and 7 amino acid substitutions spanning the CDRs, between 1 and 5 amino acid substitutions, between 1 and 4 amino acids, between 1 and 3 amino acid substitutions, between 1 and 2 amino acid substitutions, or 1 amino acid substitution, wherein the variant retains the ability to bind EGFR. Typically, such amino acid substitutions will be conservative amino acid substitutions, such as those summarized in column 1 or column 2 of Table 4 below.

[0149] In certain embodiments, the antibodies described herein comprise at least one immunoglobulin domain from a mammalian immunoglobulin such as bovine immunoglobulin, human immunoglobulin, camel immunoglobulin, rat immunoglobulin, or mouse immunoglobulin. In some embodiments, the biofunctional protein can be a chimeric antibody and comprises two or more immunoglobulin domains, wherein at least one domain is from a first mammalian immunoglobulin, e.g., human immunoglobulin, and at least a second domain is from a second mammalian immunoglobulin, e.g., mouse or rat immunoglobulin. In some embodiments, the biofunctional protein comprises at least one immunoglobulin constant domain from a human immunoglobulin.

[0150] Those skilled in the art will understand that these domains can be combined in various ways to provide antibodies with different formats, including multispecific antibodies in different formats. These formats are generally based on antibody formats known in the art (see, for example, reviewed in Brinkmann & Kontermann, 2017, MABS, 9(2):182 - 212, and Müller & Kontermann, “Bispecific Antibodies” in Handbook of Therapeutic Antibodies, Wiley - VCH Verlag GmbH & Co. (2014)).

[0151] Antibodies to the biofunctional proteins described herein can have different valences. In certain embodiments, the biofunctional protein comprises a single antigen - binding domain. In certain embodiments, the biofunctional protein comprises two or more antigen - binding domains. In certain embodiments, the biofunctional protein comprises antibodies with different valences and specificities. As used herein, a “bispecific antibody” comprises two binding domains. In certain embodiments, each of the two binding domains has a unique binding specificity. As used herein, a “multispecific antibody” comprises two or more binding domains. In certain embodiments, each of the two or more binding domains has a unique binding specificity. In some embodiments, at least two of the two or more binding domains have unique binding specificities. For example, the antibody can be bivalent and bispecific, or can be bivalent and monospecific. Alternatively, the antibody can be trivalent and bispecific, i.e., the antibody comprises three binding domains. The antibody can also be bispecific and tetravalent, i.e., the antibody comprises four binding domains. Other valences are possible.

[0152] When the antibody comprises two binding domains that bind to the same target molecule, the binding domains can bind to the same epitope on the target molecule or they can bind to different epitopes on the target molecule. In some embodiments, the antibody comprises two binding domains that bind to different epitopes on the target molecule. The term “bifunctional antibody” can be used to refer to an antibody comprising two binding domains that bind to different epitopes on the same target molecule (antigen). A bifunctional antibody can bind to a single antigen molecule through two different epitopes, or it can bind to two separate antigen molecules, each through a different epitope.

[0153] In certain embodiments, the antibody is bispecific and bivalent as it comprises a first binding domain and a second binding domain, each of which binds to a different epitope on a first target molecule; and a third binding domain that binds to the second target molecule. Alternatively, a bispecific bivalent antibody can comprise a first binding domain and a second binding domain, each of which binds to a different epitope on the first target molecule; and a third binding domain and a fourth binding domain, each of which binds to a different epitope on the second target molecule.

[0154] In some embodiments, the antibody further comprises a scaffold and the binding domains are operably linked to the scaffold. As used herein, "operably linked" means that the described components are in a relationship that permits each of them to function in its intended manner. The binding domains can be linked to the scaffold either directly or indirectly. Indirect linkage means that a given binding domain is linked to the scaffold via another component (e.g., a linker or one of the other binding domains). Various formats of fusion proteins that include a scaffold are described in more detail below.

[0155] Antigen-binding domain format

[0156] In some embodiments, the fusion proteins described herein include an antibody having at least one antigen-binding domain, which is an antibody fragment such as Fab, Fab′, single-chain Fab (scFab), single-chain Fv (scFv), or single-domain antibody (sdAb).

[0157] "Fab" or "Fab fragment" contains the constant domain of the light chain (CL) and the first constant domain of the heavy chain (CH1) and the variable domains VL and VH located on the light and heavy chains, respectively, that contain the CDRs. A Fab’ or Fab’ fragment differs from a Fab fragment in that several amino acid residues are added at the C-terminus of the heavy chain CH1 domain, including one or more cysteine residues from the hinge region.

[0158] A Fab fragment can comprise two separate polypeptide chains (light and heavy chains), or it can be single-chain Fab. Single-chain Fab is a Fab molecule in which the Fab light chain and the Fab heavy chain are joined by a peptide linker to form a single peptide chain. Typically, the C-terminus of the Fab light chain is joined to the N-terminus of the Fab heavy chain in the single-chain Fab molecule, however, other formats are possible.

[0159] "scFv" contains the variable heavy domain (VH) and variable light domain (VL) of an antibody in a single polypeptide chain. The scFv may optionally contain a polypeptide linker between the VH and VL domains, which may assist the scFv in forming the desired structure for antigen binding. The scFv may contain a VL that is linked to the N-terminus of the VH from its C-terminus via a linker (i.e., VL-linker-VH), or alternatively, the scFv may contain a VH that is linked to the N-terminus of the VL via its C-terminus (i.e., VH-linker-VL). For a review of scFv, see Pluckthun in The Pharmacology of Monoclonal Antibodies, Volume 113, edited by Rosenberg and Moore, Springer-Verlag, New York, pages 269-315 (1994).

[0160] The term "sdAb" refers to a single immunoglobulin domain. The sdAb can be, for example, of camel origin. Camel antibodies lack a light chain, and their antigen-binding site consists of a single domain called "VHH". The sdAb contains three CDRs / hypervariable loops that form the antigen-binding site: CDR1, CDR2, and CDR3. The sdAb is rather stable and easy to express, for example, as a fusion with an antibody Fc chain (see, e.g., Harmsen & De Haard, 2007, Appl. Microbiol Biotechnol. 77(1):13-22).

[0161] In some embodiments, one or more of the binding domains contained in the antibody may be a natural or engineered ligand of the target receptor, or a functional fragment of such a ligand, i.e., a fragment capable of specifically binding to the target receptor.

[0162] The antigen-binding domain can be in the form of a single scFv, a combination of Fab, sdAb, etc. For example, when the binding domain is in the form of scFv, formats such as tandem scFv ((scFv)2 or taFv) or trispecific antibody (3 scFvs) can be constructed, where the scFvs are linked together by a flexible linker. ScFv can also be used to construct diabody, trispecific antibody, and tetrabody (tandem diabody or TandAb) formats, which contain 2, 3, and 4 scFvs linked together by short linkers, respectively. The restricted linker length (usually about 5 amino acids in length) results in the dimerization of scFv in a head-to-tail manner. In any of the foregoing formats, the scFv can be further stabilized by including interdomain disulfide bonds. For example, disulfide bonds can be introduced between VL and VH by introducing additional cysteine residues in each chain (e.g., at position 44 in VH and position 100 in VL) (see, e.g., Fitzgerald et al., 1997, Protein Engineering, 10:1221-1225), or disulfide bonds can be introduced between two VHs to provide an antigen-binding domain with a DART format (see, e.g., Johnson et al., 2010, J Mol Biol., 399:436-449).

[0163] Similarly, formats containing two or more sdAbs (such as VH or VHH) linked together by a suitable linker can be used for biofunctional proteins. Other examples of antibody formats lacking a scaffold include those based on Fab fragments, such as Fab2, F(ab’)2, and F(ab’)3 formats, where the Fab fragments are linked by a linker or IgG hinge region.

[0164] Combinations of different forms of antigen-binding domains can also be employed to generate alternative formats. For example, scFv or sdAb can be fused to the C-terminus of one or both of the light and heavy chains of a Fab fragment, thereby generating bivalent (Fab-scFv) or (Fab-sdAb) or trivalent (Fab-(scFv)2 or Fab-(sdAb)2). Similarly, one or two scFvs or sdAbs can be fused in the hinge region of an F(ab’) fragment to generate trivalent or tetravalent F(ab’)2-scFv / sdAb. The binding domain can be one or a combination of the above forms (e.g., scFv, Fab, and / or sdAb, or ligand-based binding domains).

[0165] In certain specific embodiments, the biofunctional protein comprises a bispecific antibody that binds to immune cell antigens such as CD3 and tumor-associated antigens (TAAs) such as HER2. In certain more specific embodiments, the biofunctional protein comprises a bispecific antibody having a Fab-scFv format, wherein the Fab binds to an immune cell antigen and the scFv binds to a TAA. In certain more specific embodiments, the biofunctional protein comprises a bispecific antibody having a Fab-scFv format, wherein the Fab binds to CD3 and the scFv binds to HER2. In some embodiments, the biofunctional protein comprises a bispecific antibody having a Fab-Fab format, wherein one Fab binds to CD3 and the other Fab binds to HER2.

[0166] In certain embodiments, the biofunctional protein comprises two or more antigen-binding domains operably linked to a heterodimeric Fc. In this context, the biofunctional protein can be bivalent, trivalent, or tetravalent. Non-limiting examples of formats are described below. Other constructs are known in the art (see, e.g., Spiess et al., 2015, Mol Immunol., 67:95-106).

[0167] Exemplary constructs of biofunctional proteins comprising two binding domains operably linked to a heterodimeric Fc (i.e., bivalent antibodies) include, but are not limited to: a) the mAb format, wherein the first binding domain is a Fab operably linked to the N-terminus of the first Fc polypeptide of the heterodimeric Fc, and the second binding domain is a Fab operably linked to the N-terminus of the second Fc polypeptide; b) the hybrid format, wherein the first binding domain is a scFv operably linked to the N-terminus of one Fc polypeptide of the heterodimeric Fc, and the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide; and c) the biscFv format, wherein the first binding domain is a scFv operably linked to the N-terminus of the first Fc polypeptide of the heterodimeric Fc, and the second binding domain is a scFv operably linked to the N-terminus of the second Fc polypeptide.

[0168] Other examples include antibodies that comprise one binding domain (first or second) as a Fab or scFv operably linked to the N-terminus of the first Fc polypeptide and the other binding domain as a Fab or scFv operably linked to the C-terminus of the second Fc polypeptide.

[0169] Exemplary constructs of multispecific antibodies comprising three binding domains operably linked to a heterodimeric Fc (i.e., trivalent antibodies) include, but are not limited to:

[0170] A) mAb-Fv format, wherein the first binding domain is a Fab operably linked to the N-terminus of a first Fc polypeptide of a heterodimeric Fc, and the second binding domain is a Fab operably linked to the N-terminus of a second Fc polypeptide, wherein the third binding domain consists of a VH domain attached to the C-terminus of one Fc polypeptide and a VL domain attached to the C-terminus of the other Fc polypeptide;

[0171] B) mAb-scFv format, wherein the first binding domain is a Fab operably linked to the N-terminus of a first Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of a second Fc polypeptide, and the third binding domain is an scFv operably linked to the C-terminus of the first or second Fc polypeptide;

[0172] C) scFv-mAb format, wherein the first binding domain is a Fab operably linked to the N-terminus of a first Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of a second Fc polypeptide, and the third binding domain is an scFv operably linked to the N-terminus of the first or second Fc polypeptide;

[0173] D) Central scFv format, wherein the first binding domain is an scFv operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide, and the third binding domain is a Fab operably linked to the first binding domain (scFv);

[0174] E) Fab-hybrid format, wherein the first binding domain is an scFv operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide, and the third binding domain is a Fab operably linked to the N-terminus of the first or second binding domain;

[0175] F) scFv-hybrid format, wherein the first binding domain is an scFv operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide, and the third binding domain is an scFv operably linked to the N-terminus of the first or second binding domain;

[0176] G) Hybrid-scFv format, wherein the first binding domain is a scFv operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide, and the third binding domain is a scFv operably linked to the C-terminus of the first or second Fc polypeptide;

[0177] H) Hybrid-Fab format, wherein the first binding domain is a scFv operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide, and the third binding domain is a Fab operably linked to the C-terminus of the first Fc polypeptide or the second Fc polypeptide; and

[0178] I) Fab-mAb format, wherein the first binding domain is a Fab operably linked to the N-terminus of the first Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the second Fc polypeptide, and the third binding domain is a Fab operably linked to the N-terminus of the first or second binding domain.

[0179] Exemplary constructs of multispecific antibodies (i.e., tetravalent antibodies) comprising four binding domains operably linked to a heterodimeric Fc include, but are not limited to: i) central scFv2 format, wherein the first binding domain is a scFv operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a scFv operably linked to the N-terminus of the other Fc polypeptide, the third binding domain is a Fab operably linked to one of the scFvs, and the fourth binding domain is a Fab operably linked to the other scFv; and ii) dual variable domain format, wherein the first binding domain is a Fab operably linked to the N-terminus of one Fc polypeptide of a heterodimeric Fc, the second binding domain is a Fab operably linked to the N-terminus of the other Fc polypeptide, the third binding domain is a scFv operably linked to one of the Fabs, and the fourth binding domain is a scFv operably linked to the other Fab.

[0180] Antibodies to the biofunctional proteins described herein may comprise a label, a drug, or a combination thereof. Any label known in the art suitable for detecting the fusion proteins described herein may be used. Antibody-drug conjugates are described in more detail below.

[0181] In certain embodiments, the antigen-binding domain of an antibody of a biofunctional protein described herein binds to the same antigen on the same cell. In certain embodiments, the antigen-binding domain binds to more than one antigen on the same cell. In certain embodiments, the antigen-binding domain binds to more than one antigen, wherein at least one antigen is on a different cell from another antigen. In certain embodiments, one or more antigen-binding domains of the antibody bind to a tumor cell or an immune cell. In certain embodiments, the antigen-binding domain of the antibody binds to a tumor cell or an immune cell.

[0182] Chimeric, humanized, and variant antibodies

[0183] In some embodiments, the antibody can be derived from immunoglobulins from different species. For example, the antibody can be a chimeric antibody or a humanized antibody. A "chimeric antibody" refers to an antibody that typically comprises at least one variable domain from a rodent antibody (usually a murine antibody) and at least one constant domain from a human antibody. A "humanized antibody" is a class of chimeric antibody that contains the minimal sequence derived from a non-human antibody.

[0184] The human constant domain of a chimeric antibody need not have the same isotype as the non-human constant domain it replaces. Chimeric antibodies are discussed, for example, in Morrison et al., 1984, Proc. Natl. Acad. Sci. USA, 81:6851-55 and U.S. Patent No. 4,816,567. Generally, a humanized antibody is a human immunoglobulin (recipient antibody) in which residues from the recipient hypervariable regions are replaced by residues from the hypervariable regions of a non-human species (donor antibody) such as mouse, rat, rabbit, or non-human primate having the desired specificity and affinity for the target antigen. This technique for creating humanized antibodies is often referred to as "CDR grafting". Both "chimeric antibodies" and "humanized antibodies" generally refer to antibodies that combine immunoglobulin regions or domains from more than one species.

[0185] In some cases, additional modifications are made to further improve antibody performance. For example, framework region (FR) residues of human immunoglobulins are replaced with corresponding non-human residues, or a humanized antibody may contain residues not found in either the recipient or donor antibody. In general, the variable domains in a humanized antibody will contain all or substantially all of the hypervariable regions from a non-human immunoglobulin and all or substantially all of the FRs from a human immunoglobulin sequence. Humanized antibodies are described in more detail, for example, in Jones, et al., 1986, Nature, 321:522-525; Riechmann, et al., 1988, Nature, 332:323-329, and Presta, 1992, Curr.Op.Struct.Biol., 2:593-596.

[0186] Numerous methods are known in the art for selecting the most appropriate human framework into which to graft non-human CDRs. Early methods used a limited subset of well-characterized human antibodies without regard to sequence identity with the non-human antibody providing the CDRs (“fixed framework” method). More recent methods have employed variable regions having high amino acid sequence identity with the variable regions of the non-human antibody providing the CDRs (“homology matching” or “best fit” method). An alternative method is to select fragments of framework sequences from within each light or heavy chain variable region from several different human antibodies. CDR grafting can in some cases result in partial or complete loss of affinity of the grafted molecule for its target antigen. In such cases, affinity can be restored by backmutating some human-derived residues to the corresponding non-human residues. Methods for preparing humanized antibodies by these methods are well known in the art (see, e.g., Tsurushita & Vasquez, 2004, Humanization of MonoclonalAntibodies, Molecular Biology of B Cells, 533-545, Elsevier Science (USA); Jones et al., 1986, Nature, 321:522-525; Riechmann et al., 1988, Nature, 332:323.329; Presta et al., 1997, Cancer Res, 57(20):4593-4599).

[0187] Alternatively, or in addition to these traditional methods, more recent techniques can be employed to further reduce the immunogenicity of humanized antibodies transplanted via CDR. For example, frameworks based on human germline sequences or consensus sequences can be used as the recipient human framework instead of a human framework with one or more somatic mutations. Another technique aimed at reducing the potential immunogenicity of non-human CDRs is to transplant only the specific determining residues (SDRs). In this approach, only the minimal number of CDR residues required for antigen-binding activity ("SDRs") are transplanted into the human germline framework. This approach increases the "humanity" of the humanized antibody (i.e., similarity to the human germline sequence), thus helping to reduce the risk of variable region immunogenicity. These techniques have been described in various publications (see, e.g., Almagro & Fransson, 2008, Front Biosci, 13:1619-1633; Tan, et al., 2002, J Immunol, 169:1119-1125; Hwang, et al., 2005, Methods, 36:35-42; Pelat, et al., 2008, J Mol Biol, 384:1400-1407; Tamura, et al., 2000, J Immunol, 164:1432-1441; Gonzales, et al., 2004, Mol Immunol, 1:863-872, and Kashmiri, et al., 2005, Methods, 36:25-34).

[0188] In certain embodiments, the antibody comprises a humanized antibody sequence, e.g., one or more humanized variable domains. In some embodiments, the antibody is a humanized antibody.

[0189] In certain embodiments, the antigen-binding domain comprised by the fusion protein is a substitutional variant of a known antibody that contains one or more amino acid substitutions in the CDRs of the parental antibody. In certain embodiments, the substitutional variant has a modification (e.g., improvement) in certain biological properties relative to the parental antibody. For example, the substitutional variant may have increased affinity for the target protein, or it may have reduced immunogenicity. In some embodiments, the substitutional variant substantially retains certain biological properties of the parental antibody.

[0190] CDR hotspots are residues encoded by codons that mutate at high frequency during somatic maturation (see, e.g., Chowdhury, 2008, Methods Mol. Biol., 207:179-196). Affinity maturation by construction of secondary libraries and re-selection therefrom has been described (see, e.g., Hoogenboom et al. in Methods in Molecular Biology, 178:1-37, O’Brien et al., eds., Human Press, Totowa, N.J. (2001)).

[0191] Methods of affinity maturation are well known in the art. For example, diversity can be introduced into the variable genes selected for maturation by various techniques including, for example, error-prone PCR, chain shuffling or oligonucleotide-directed mutagenesis. A secondary library is then created and screened to identify any antibody variants having the desired affinity. Another method of introducing diversity involves CDR-directed methods, where several CDR residues (e.g., 2, 3, 4 or more residues at a time) are randomized. CDR3 of either or both the heavy or light chain is typically the target of CDR-directed methods. CDR residues involved in antigen binding can be identified, for example, using alanine-scanning mutagenesis (see, e.g., Cunningham and Wells, 1989, Science, 244:1081-1085) or by computer modeling of the crystal structure of the antigen-antibody complex to identify the contact points between the antibody and the antigen.

[0192] In certain embodiments, the substituted variants contain one or more substitutions located within one or more CDRs, provided that the substitutions do not substantially reduce the ability of the binding domain to bind its target antigen. For example, the substituted variants can contain one or more conservative substitutions as described herein within one or more CDRs that do not substantially reduce binding affinity. In some embodiments, the substituted variants contain one or more amino acid substitutions within CDRs that do not involve amino acids that contact the antigen. In some embodiments, the substituted variants contain variant VH or VL sequences, where each CDR is either unaltered or contains no more than one, two or three amino acid substitutions.

[0193] Glycosylation variant

[0194] In certain embodiments, the fusion proteins described herein comprise an IgG Fc-based biofunctional protein, wherein the initial glycosylation has been modified. As is known in the art, glycosylation of the Fc can be modified to increase or decrease effector function.

[0195] For example, mutation of the conserved asparagine residue at position 297 to alanine, glutamine, lysine, or histidine (i.e., N297A, Q, K, or H) results in a non-glycosylated Fc lacking all effector functions (Bolt et al., 1993, Eur. J. Immunol., 23: 403-411; Tao & Morrison, 1989, J. Immunol., 143: 2595-2601).

[0196] In contrast, removal of fucose from the N297-linked oligosaccharides of the heavy chain has been shown to enhance ADCC based on improved binding to FcγRIIIa (see, e.g., Shields et al., 2002, J Biol Chem., 277: 26733-26740, and Niwa et al., 2005, J. Immunol. Methods, 306: 151-160). Such afucosylated antibodies can be produced, for example, in knockout Chinese hamster ovary (cHO) cells lacking fucosyltransferase (FUT8) (Yamane-Ohnuki et al., 2004, Biotechnol. Bioeng., 87: 614-622), in the variant CHO cell line Lec 13 with reduced ability to attach fucose to N297-linked carbohydrates (International Publication No. WO 03 / 035835), or in other cells that produce non-fucosylated antibodies (see, e.g., Li et al., 2006, Nat Biotechnol, 24: 210-215; Shields et al., 2002, ibid, and Shinkawa et al., 2003, J. Biol. Chem., 278: 3466-3473). In addition, International Publication No. WO 2009 / 135181 describes adding fucose analogs to the culture medium during antibody production to inhibit incorporation of fucose into the carbohydrates on the antibody.

[0197] Other methods for generating antibodies with little or no fucose at the Fc glycosylation site (N297) are well known in the art. For example, techniques (ProBioGen AG) (see von Horsten et al., 2010, Glycobiology, 20(12): 1607-1618 and U.S. Patent No. 8,409,572).

[0198] Other glycosylation variants include those with bisecting oligosaccharides, e.g., variants in which the biantennary oligosaccharide attached to the Fc region of the antibody is bisected by N-acetylglucosamine (GlcNAc). Such glycosylation variants can have reduced fucosylation and / or improved ADCC function. See, e.g., International Publication No. WO 2003 / 011878, U.S. Patent No. 6,602,684, and U.S. Patent Application Publication No. US 2005 / 0123546. Useful glycosylation variants also include those having at least one galactose residue in the oligosaccharide attached to the Fc region, which can have improved CDC function (see, e.g., International Publications WO 1997 / 030087, WO 1998 / 58964, and WO 1999 / 22764).

[0199] Polypeptide scaffold

[0200] In certain embodiments, the biofunctional protein of the fusion proteins described herein is a polypeptide scaffold, which can serve, for example, to stabilize or extend the in vivo half-life of a ligand-receptor pair.

[0201] In certain embodiments, the biofunctional protein consists of a dimeric Fc region. In certain embodiments, the first and second polypeptides of the biofunctional protein consist of dimeric Fc, wherein the first polypeptide consists of a first Fc polypeptide and the second polypeptide consists of a second Fc polypeptide, and the first and second Fc polypeptides form a dimeric Fc region. In certain embodiments, the dimeric Fc region is a heterodimeric Fc. Heterodimeric Fc regions are described in more detail herein.

[0202] In certain embodiments, the polypeptide scaffold consists of a first and a second polypeptide. In certain embodiments, the ligand of the ligand-receptor pair is fused to the first polypeptide via a peptide linker, and the receptor is fused to the same corresponding end of the second polypeptide via a peptide linker. Thus, in certain embodiments, the ligand is fused to the N-terminus of the first polypeptide via a peptide linker, and the receptor is fused to the N-terminus of the second polypeptide via a second peptide linker. Conversely, in certain embodiments, the ligand is fused to the C-terminus of the first polypeptide via a peptide linker, and the receptor is fused to the second polypeptide via a second peptide linker.

[0203] In certain more specific embodiments, the biofunctional protein comprises a polypeptide scaffold consisting of a dimeric Fc region and a ligand-receptor pair (i.e., PDL-1 and PD-1). In certain embodiments, the fusion protein comprises a biofunctional protein consisting of a dimeric Fc region and a ligand-receptor pair (i.e., CD80 and CTLA4). In certain embodiments, the Fc domain of the polypeptide scaffold comprises amino acid sequences corresponding to SEQ ID NOs: 4 and 5 and optionally SEQ ID NO: 6. In certain embodiments, the polypeptide scaffold consists of a heterodimeric Fc comprising SEQ ID NO: 4 and SEQ ID NO: 5; wherein the first Fc polypeptide comprises SEQ ID NO: 4 and the second Fc polypeptide comprises SEQ ID NO: 5. In some embodiments, the polypeptide scaffold consisting of a heterodimeric Fc comprises the modified CH3 and / or CH2 domains of Tables 2 and 3, respectively.

[0204] Fc domain

[0205] In certain embodiments, the fusion proteins described herein include biofunctional proteins comprising a dimeric immunoglobulin Fc region, such as an antibody or a polypeptide scaffold. The term "Fc region" includes native sequence Fc regions and variant Fc regions. Unless otherwise specified herein, the numbering of amino acid residues in the Fc region or constant region is according to the EU numbering system, also known as the EU index, as described in Kabat et al., Sequences of Proteins of Immunological Interest, 5th Ed., Public Health Service, National Institutes of Health, Bethesda, MD (1991). An "Fc polypeptide" of a dimeric Fc refers to one of the two polypeptides that form the dimeric Fc region, i.e., a polypeptide comprising the C-terminal constant region of an immunoglobulin heavy chain capable of stable self-association.

[0206] The Fc region may comprise a CH3 domain or CH3 and CH2 domains. The CH3 domain comprises two CH3 sequences, each CH3 sequence comprising one of the two Fc polypeptides of the dimeric Fc. Similarly, the CH2 domain comprises two CH2 sequences, each CH2 sequence comprising one of the two Fc polypeptides of the dimeric Fc.

[0207] In certain embodiments, the fusion protein comprises an Fc based on human IgG Fc. In some embodiments, the fusion protein comprises an Fc based on human IgG1 Fc. In some embodiments, the fusion protein comprises an Fc based on a heterodimeric Fc comprising two different Fc polypeptides.

[0208] In certain embodiments, the fusion protein comprises an Fc based on a modified IgG Fc, wherein the CH3 domain comprises one or more amino acid modifications. In some embodiments, the fusion protein comprises an Fc based on a modified IgG Fc, wherein the CH2 domain comprises one or more amino acid modifications. In some embodiments, the fusion protein comprises an Fc based on a modified IgG Fc, wherein the CH3 domain comprises one or more amino acid modifications and the CH2 domain comprises one or more amino acid modifications.

[0209] Modified Fc CH3 domain

[0210] In certain embodiments, the fusion protein comprises a heterodimeric immunoglobulin Fc comprising a modified CH3 domain, wherein the modified CH3 domain comprises one or more asymmetric amino acid modifications. As used herein, "asymmetric amino acid modification" refers to a modification in which the amino acid at a particular position on a first Fc polypeptide is different from the amino acid at the corresponding position on a second Fc polypeptide. These asymmetric amino acid modifications can include modification of only one of two amino acids at the corresponding positions on each Fc polypeptide, or they can include modification of both amino acids at the corresponding positions on each of the first and second Fc polypeptides.

[0211] In certain embodiments, the fusion protein comprises a heterodimeric Fc comprising a modified CH3 domain, wherein the modified CH3 domain comprises one or more asymmetric amino acid modifications that promote the formation of the heterodimeric Fc rather than the homodimeric Fc. Amino acid modifications that can be made to the CH3 domain of an Fc to promote the formation of a heterodimeric Fc are known in the art and include, for example, those described in International Publication No. WO 96 / 027011 ("knobs-into-holes"); Gunasekaran et al., 2010, J Biol Chem, 285, 19637-46 ("electrostatic steering"); Davis et al., 2010, ProtEng Des Sel, 23(4):195-202 (strand-exchange engineered domain (SEED) technology) and Labrijn et al., 2013, Proc Natl Acad Sci USA, 110(13):5145-50 (Fab-arm exchange). Other examples include methods that combine plus and minus design strategies to produce a stable asymmetrically modified Fc region, as described in International Publications No. WO2012 / 058768 and No. WO 2013 / 063702.

[0212] In certain embodiments, the fusion protein comprises a heterodimeric Fc with a modified CH3 domain, as described in International Publication No. WO 2012 / 058768 or International Patent Publication No. WO 2013 / 063702.

[0213] In some embodiments, the fusion protein comprises a heterodimeric IgG1 Fc with a modified CH3 domain. Table 2 below provides the amino acid sequence of the human IgG1 Fc corresponding to amino acids 231 to 447 of the full-length human IgG1 heavy chain. The CH2 domain is generally defined as comprising amino acids 231 - 340 of the full-length human IgG1 heavy chain, and the CH3 domain is generally defined as comprising amino acids 341 - 447 of the full-length human IgG1 heavy chain.

[0214] In certain embodiments, the fusion protein comprises a heterodimeric Fc with a modified CH3 domain, the modified CH3 domain comprising one or more asymmetric amino acid modifications that promote heterodimeric Fc formation rather than homodimeric Fc formation, wherein the modified CH3 domain comprises a first Fc polypeptide having amino acid modifications at positions F405 and Y407 and a second Fc polypeptide having amino acid modifications at positions T366 and T394. In some embodiments, the amino acid modification at position F405 of the first Fc polypeptide of the modified CH3 domain is F405A, F405I, F405M, F405S, F405T, or F405V. In some embodiments, the amino acid modification at position Y407 of the first Fc polypeptide of the modified CH3 domain is Y407I or Y407V. In some embodiments, the amino acid modification at position T366 of the second Fc polypeptide of the modified CH3 domain is T366I, T366L, or T366M. In some embodiments, the amino acid modification at position T394 of the second Fc polypeptide of the modified CH3 domain is T394W. In some embodiments, the first Fc polypeptide of the modified CH3 domain further comprises an amino acid modification at position L351. In some embodiments, the amino acid modification at position L351 in the first Fc polypeptide of the modified CH3 domain is L351Y. In some embodiments, the second Fc polypeptide of the modified CH3 domain further comprises an amino acid modification at position K392. In some embodiments, the amino acid modification at position K392 of the second Fc polypeptide of the modified CH3 domain is K392F, K392L, or K392M. In some embodiments, one or both of the first and second Fc polypeptides of the modified CH3 domain further comprise the amino acid modification T350V.

[0215] In certain embodiments, the fusion protein comprises a heterodimeric Fc having a modified CH3 domain, the modified CH3 domain comprising one or more asymmetric amino acid modifications that promote heterodimeric Fc formation rather than homodimeric Fc formation, wherein the modified CH3 domain comprises a first Fc polypeptide comprising the amino acid modifications F405A, F405I, F405M, F405S, F405T, or F405V and the amino acid modification Y407I or Y407V, and a second Fc polypeptide comprising the amino acid modifications T366I, T366L, or T366M and the amino acid modification T394W. In some embodiments, the first Fc polypeptide of the modified CH3 domain further comprises the amino acid modification L351Y. In some embodiments, the second Fc polypeptide of the modified CH3 domain further comprises the amino acid modifications K392F, K392L, or K392M. In some embodiments, one or both of the first and second Fc polypeptides of the modified CH3 domain further comprises the amino acid modification T350V.

[0216] In certain embodiments, the fusion protein comprises a heterodimeric Fc comprising a modified CH3 domain, the CH3 domain having a first Fc polypeptide comprising amino acid modifications at positions F405 and Y407 and optionally further comprising an amino acid modification at position L351, and a second Fc polypeptide comprising amino acid modifications at positions T366 and T394 and optionally further comprising an amino acid modification at position K392, as described above, and the first Fc polypeptide further comprising an amino acid modification at one or both of positions S400 or Q347, and / or the second Fc polypeptide further comprising an amino acid modification at one or both of positions K360 or N390, wherein the amino acid modification at position S400 is S400E, S400D, S400R, or S400K; the amino acid modification at position Q347 is Q347R, Q347E, or Q347K; the amino acid modification at position K360 is K360D or K360E, and the amino acid modification at position N390 is N390R, N390K, or N390D.

[0217] In certain embodiments, the fusion protein comprises a heterodimeric Fc, the heterodimeric Fc comprising a modified CH3 domain, the modified CH3 domain comprising a modification of any one of variants 1, 2, 3, 4, or 5 as shown in Table 2. In certain embodiments, the CH3 domain has an amino acid sequence corresponding to SEQ ID NO: 4 or SEQ ID NO: 5. In certain embodiments, the CH3 has an amino acid sequence substantially identical to SEQ ID NO: 4 or SEQ ID NO: 5. In certain embodiments, the CH3 domain has an amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 4 or SEQ ID NO: 5.

[0218] Table 2: Human IgG1 Fc Sequences and Variants

[0219]

[0220]

[0221] Modified Fc CH2 domain

[0222] In certain embodiments, the fusion protein comprises an Fc based on an IgG Fc having a modified CH2 domain. In some embodiments, the fusion protein comprises an Fc based on an IgG Fc having a modified CH2 domain, wherein the modification of the CH2 domain results in an altered binding to one or more Fc receptors (FcRs) (such as receptors of the FcγRI, FcγRII, and FcγRIII subclasses).

[0223] Many amino acid modifications that selectively alter the affinity of the Fc for different Fcγ receptors for the CH2 domain are known in the art. Both amino acid modifications that result in increased binding and those that result in decreased binding can be used for certain indications. For example, increasing the binding affinity of the Fc for FcγRIIIa (the activated receptor) results in increased antibody-dependent cell-mediated cytotoxicity (ADCC), which in turn results in increased target cell lysis. In some cases, a decrease in binding to FcγRIIb (the inhibitory receptor) may also be beneficial in some cases. In certain indications, a reduction or elimination of ADCC and complement-mediated cytotoxicity (cDC) may be desired. In such cases, a modified CH2 domain comprising an amino acid modification that results in an increase in binding to FcγRIIb or an amino acid modification that reduces or eliminates the binding of the Fc region to all Fcγ receptors ("knockout" variants) may be useful.

[0224] Examples of amino acid modifications to the CH2 domain that alter the binding of Fcγ receptors to Fc include, but are not limited to, the following: S298A / E333A / K334A and S298A / E333A / K334A / K326A (increased affinity for FcγRIIIa) (Lu, et al., 2011, J Immunol Methods, 365(1-2):132-41); F243L / R292P / Y300L / V305I / P396L (increased affinity for FcγRIIIa) (Stavenhagen, et al., 2007, Cancer Res, 67(18):8882-90); F243L / R292P / Y300L / L235V / P396L (increased affinity for FcγRIIIa) (Nordstrom JL, et al., 2011, Breast Cancer Res, 13(6):R123); F243L (increased affinity for FcγRIIIa) (Stewart, et al., 2011, Protein Eng Des Sel., 24(9):671-8); S298A / E333A / K334A (increased affinity for FcγRIIIa) (Shields, et al., 2001, J Biol Chem, 276(9):6591-604); S239D / I332E / A330L and S239D / I332E (increased affinity for FcγRIIIa) (Lazar, et al., 2006, Proc Natl Acad Sci USA, 103(11):4005-10), and S239D / S267E and S267E / L328F (increased affinity for FcγRIIb) (Chu, et al., 2008, Mol Immunol, 45(15):3926-33).

[0225] Additional modifications that affect the binding of Fc to Fcγ receptors are described in Therapeutic Antibody Engineering (Strohl & Strohl, Woodhead Publishing series in Biomedicine No 11, ISBN 1 907568 37 9, October 2012, page 283).

[0226] In certain embodiments, the fusion protein comprises an Fc based on an IgG Fc having a modified CH2 domain, wherein the modified CH2 domain comprises one or more amino acid modifications that result in reduced or eliminated binding of the Fc region to all Fcγ receptors (i.e., a "knockout" variant).

[0227] A variety of publications describe strategies that have been used to engineer antibodies to produce “knockout” variants (see, e.g., Strohl, 2009, Curr Opin Biotech 20:685 - 691, and Strohl & Strohl, “Antibody Fc engineering for optimal antibody performance” In Therapeutic Antibody Engineering, Cambridge: Woodhead Publishing, 2012, pp. 225 - 249). These strategies include reducing effector function by modification of glycosylation (described in more detail below), use of IgG2 / IgG4 scaffolds, or introduction of mutations in the hinge or CH2 domains of the Fc (see also, U.S. Patent Publication No. 2011 / 0212087, International Publication No. WO 2006 / 105338, U.S. Patent Publication No. 2012 / 0225058, U.S. Patent Publication No. 2012 / 0251531, and Strop et al., 2012, J. Mol. Biol., 420:204 - 219).

[0228] Specific non - limiting examples of known amino acid modifications for reducing the binding of FcγR and / or complement to the Fc include those identified in Table 3.

[0229] Table 3: Modifications for reducing the binding of Fcγ receptor or complement to the Fc

[0230] Company Mutation GSK N297A Ortho Biotech L234A / L235A Protein Design labs IgG2 V234A / G237A Wellcome Labs IgG4 L235A / G237A / E318A GSK IgG4 S228P / L236E Merck IgG2 H268Q / V309L / A330S / A331S Bristol-Myers C220S / C226S / C229S / P238S Seattle Genetics C226S / C229S / E3233P / L235V / L235A Medimmune L234F / L235E / P331S

[0231] Additional examples include an Fc region engineered to include the amino acid modifications L235A / L236A / D265S. In addition, asymmetric amino acid modifications in the CH2 domain that reduce the binding of Fc to all Fcγ receptors are described in International Publication No. WO 2014 / 190441.

[0232] In certain embodiments, the CH2 domain has the amino acid sequence corresponding to SEQ ID NO:6. In certain embodiments, the CH2 has an amino acid sequence substantially identical to SEQ ID NO:6. In certain embodiments, the CH2 domain has an amino acid sequence about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:6.

[0233] Antibody-drug conjugate

[0234] Certain embodiments of the fusion proteins described herein include a biofunctional protein that is an antibody conjugated to a drug, i.e., an antibody-drug conjugate (ADC). The drug of the ADC can be any therapeutic molecule, such as a toxin, a chemotherapeutic agent, a small molecule inhibitor. The ADC can be conjugated to the drug via a linker, which can be a cleavable linker or a non-cleavable linker. A cleavable linker may be capable of being readily cleaved under intracellular conditions, e.g., by lysosomal processes. Examples of cleavable linkers include protease-sensitive, acid-sensitive, reduction-sensitive, or photo-labile linkers. Conjugation of the drug can be carried out by any method known in the art, which includes but is not limited to lysine or cysteine conjugation, dithiol linkers, conjugation using antibody glycosylation sites, ultraviolet light conjugation, and conjugation using unnatural amino acids.

[0235] Peptide linkers, proteases, and protease cleavage sites

[0236] The fusion proteins described herein include at least a first and a second peptide linker. A peptide linker is a peptide that joins or links other peptides or polypeptides. In certain embodiments, the peptide linker fuses a polypeptide of a biofunctional protein, such as an antibody or a dimeric Fc scaffold, to a ligand and / or a receptor of a ligand-receptor pair.

[0237] In certain embodiments, where the biofunctional protein includes an Fc region, the Fc polypeptide is fused to a ligand or a receptor of a ligand-receptor pair, or the linker can join the Fc polypeptide to a ligand or a receptor of a ligand-receptor pair. In certain embodiments, the ligand is fused to the end of a first polypeptide via a first peptide linker; the receptor is fused to the same corresponding end of a second polypeptide via a second peptide linker. In certain embodiments of the fusion proteins described herein, both the receptor and the ligand are fused to the corresponding N-termini of the first and second polypeptides via peptide linkers. In certain embodiments of the fusion proteins described herein, both the receptor and the ligand are fused to the corresponding C-termini of the first and second polypeptides via peptide linkers.

[0238] The peptide linker has a sufficient length to allow ligand and receptor pairing. In addition to providing a spacer function, the peptide linker can provide flexibility or rigidity suitable for correctly orienting one or more domains of the fusion proteins described herein within the fusion protein and between or among the fusion protein and its one or more targets. Further, the peptide linker can support the expression of the full-length fusion protein and the stability of the purified protein in vitro and in vivo after administration to a subject in need thereof, such as a human, and is preferably non-immunogenic or weakly immunogenic in such subjects. In certain embodiments, the peptide linker can include a portion or all of a human immunoglobulin hinge, the stalk region of a C-type lectin, a type II membrane protein family, or a combination thereof.

[0239] In certain embodiments, the peptide linker has sufficient length to allow ligand and receptor pairing and has from about 2 to about 150 amino acids. In certain embodiments, the peptide linker has a length ranging from about 3 to about 50 amino acids, or about 5 to about 20 amino acids, or about 10 to about 50 amino acids, or about 2 to about 40 amino acids, or about 8 to about 20 amino acids, about 10 to about 60 amino acids, about 10 to about 30 amino acids, or about 15 to about 25 amino acids. In some embodiments, the peptide linker is 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, or 60 amino acids.

[0240] At least one peptide linker in the peptide linker of the fusion protein described herein contains a protease cleavage site, also referred to as a cleavage sequence. In certain embodiments, the fusion protein contains at least one peptide linker containing a protease cleavage site and at least one peptide linker not containing a protease cleavage site. When used, the protease cleavage site is positioned within the peptide linker to maximize recognition and cleavage by a desired one or more proteases and to minimize recognition and non-specific cleavage by other proteases. The peptide linker may contain one or more cleavage sites. In these aspects, the fusion protein can be cleaved by 1, 2, 3, 4, 5, or more proteases. Additionally, the one or more protease cleavage sites may be positioned within the peptide linker (or in other words, may be surrounded by the linker) and as a whole be positioned within the fusion protein to achieve optimal desired cleavage and release of the fusion protein fragments (e.g., the ligand of a receptor-ligand pair, the receptor of a ligand-receptor pair, or both the ligand and the receptor) upon cleavage. The polypeptide portion that is fused to the fusion protein via the peptide linker and released from the fusion protein upon cleavage of the peptide linker is referred to herein as the cleavable moiety (CM). In certain embodiments where the fusion protein contains more than one CM, they may be fused to the fusion protein via the same or different peptide linkers (i.e., having the same or different cleavage sites).

[0241] The protease cleavage site or cleavage sequence can be selected based on proteases that are co-localized in tissues where the activity of the fusion protein or biofunctional protein is desired. The cleavage site can serve as a substrate for a variety of proteases, e.g., a substrate for serine proteases and a second different protease (e.g., matrix metalloproteinase (MMP)). In some embodiments, the cleavage site can serve as a substrate for more than one serine protease (e.g., plasmin and urokinase-type plasminogen activator (uPA)). In some embodiments, the peptide linker can serve as a substrate for more than one MMP (e.g., MMP9 and MMP14).

[0242] In certain embodiments, the peptide linker is specifically cleaved by a protease at a rate of about 0.001 - 1500×10 4 M -1 S -1 or at least 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1, 2.5, 5, 7.5, 10, 15, 20, 25, 50, 75, 100, 125, 150, 200, 250, 500, 750, 1000, 1250 or 1500×10 4 M -1 S -1 per second.

[0243] For specific cleavage with an enzyme, contact is made between the enzyme and the peptide linker. In certain embodiments, the peptide linker is cleaved when the fusion protein comprises at least a first peptide linker and there is sufficient enzyme activity present. Sufficient enzyme activity can refer to the ability of the enzyme to contact the peptide linker and effect cleavage. It can be readily envisioned that the enzyme may be in the vicinity of the peptide linker but unable to cleave due to protein modification by other cytokines or enzymes.

[0244] In certain embodiments, the peptide linker comprises a protease cleavage site that is 5-10 amino acids in length, or 7-10 amino acids in length, or 8-10 amino acids in length. In another embodiment, the peptide linker consists of a protease cleavage site that is 5-10 amino acids in length, or 7-10 amino acids in length, or 8-10 amino acids in length. In one embodiment, the protease cleavage site is preceded on the N-terminus by a linker sequence that is about 1-20 amino acids in length, 2-5 amino acids in length, 5-10 amino acids in length, 10-15 amino acids in length, 10-20 amino acids in length, 12-16 amino acids in length, or about 5 or about 10 amino acids in length. In another embodiment, the protease cleavage site is preceded on the C-terminus by a linker sequence that is about 1-20 amino acids in length, 2-5 amino acids in length, 5-10 amino acids in length, 10-15 amino acids in length, 10-20 amino acids in length, 12-16 amino acids in length, or in some instances about 5 or about 10 amino acids in length. In yet another embodiment, the protease cleavage site is preceded by a linker sequence located on the N-terminus and followed by a linker sequence located on the C-terminus. Thus, in certain embodiments, the protease cleavage site is located between two linkers. The linker on the N-terminus or C-terminus of the protease cleavage site can have different lengths, e.g., lengths between about 2-20, 6-20, 8-15, 8-10, 10-18, or 12-16 amino acids. In certain embodiments, the length of the N-terminus or C-terminus linker is about 3 or about 5 amino acids.

[0245] Exemplary peptide linkers of the present disclosure include one or more protease cleavage sites recognized by any of a variety of proteases, such as but not limited to serine proteases, MMPs (MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, MMP17, MMP18 (collagenase 4), MMP19, MMP20, MMP21, etc.), disintegrins, serralysins, astacins, caspases (e.g., caspase 1, caspase 2, caspase 3, caspase 4, caspase 5, caspase 6, caspase 7, caspase 8, caspase 9, caspase 10, caspase 11, caspase 12, caspase 13, caspase 14), cathepsins (e.g., cathepsin A, cathepsin B, cathepsin D, cathepsin E, cathepsin K, cathepsin S), granzyme B, guanidino benzoate enzyme (GB), heparase, elastase, podoplanin, proteinase 2, methyldopa, neuropilin, MT-SP1, neprilysin, plasmin, PSA, PSMA, TACE, TMPRSS3 / 4, uPA, and calpains, FAP, and KLK. In some embodiments, the protease is uPA or proteinase 2.

[0246] In certain embodiments, the peptide linker comprises cleavage sites that are cleaved by more than one protease. In these aspects, a single cleavage site can be cleaved by 1, 2, 3, 4, 5 or more proteases. In another embodiment, the peptide linker can comprise cleavage sites that are substantially cleaved by one enzyme and not by other enzymes. Thus, in some embodiments, the peptide linker comprises cleavage sites with high specificity. "High specificity" means that >90% cleavage by a specific protease is observed and <50% cleavage by other proteases is observed. In certain embodiments, the peptide linker comprises cleavage sites that exhibit >80% cleavage by one protease but <50% cleavage by other proteases. In certain embodiments, the peptide linker comprises cleavage sites that exhibit >70%, 75%, 76%, 77%, 78% or 79% cleavage by one protease but <65%, 60%, 55%, 54%, 53%, 52%, 51%, 50%, 49%, 48%, 47%, 46% or 45% cleavage by other proteases. For example, in one embodiment, the cleavage site can be cleaved >90% by a proteolytic enzyme and about 75% by uPa and plasmin. In another embodiment, the cleavage site can be cleaved by uPa and a proteolytic enzyme, but no specific cleavage by plasmin is observed. In yet another embodiment, the cleavage site can be cleaved by uPa and cannot be cleaved by a proteolytic enzyme or plasmin. In one embodiment, the cleavage site can exhibit a degree of resistance to non-specific protease cleavage (e.g., cleavage by plasmin or other non-specific proteases). In this regard, the protease cleavage site can have "high non-specific protease resistance" (<25% cleavage by plasmin or an equivalent non-specific protease), "medium non-specific protease resistance" (<75% cleavage by plasmin or an equivalent non-specific protease), or "low non-specific protease resistance" (cleavage by plasmin or an equivalent non-specific protease up to about 90%). Such cleavage activity can be measured using assays known in the art, such as by incubating with the appropriate protease followed by SDS-PAGE or other analysis. In certain embodiments, the protease cleavage site shows up to complete protease cleavage resistance for 24 hours of contact with the protease. In other embodiments, the protease cleavage sequence can show up to complete non-specific protease cleavage resistance after 0.5 hours to 36 hours of contact with the protease. In another embodiment, the protease cleavage sequence shows up to complete non-specific protease cleavage resistance after 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 20, 24, 36, 48 or 72 hours of contact with the appropriate protease.

[0247] Thus, in certain embodiments, the cleavage sites are selected based on the preferences of various desired proteases. In this way, a desired cleavage profile of a particular peptide linker containing a cleavage site can be selected for a desired purpose (e.g., high-specificity cleavage in a particular tumor microenvironment or a particular organ), where a particular protease or set of proteases may exhibit high, specific, elevated, efficient, moderate, low, or no cleavage at a particular cleavage site within the peptide linker. Methods for determining cleavage are known in the art.

[0248] In certain embodiments, a peptide linker may comprise one or more cleavage sites arranged in tandem, with or without additional linker between each cleavage site. In certain embodiments, the peptide linker comprises a first cleavage site and a second cleavage site, where the first cleavage site is cleaved by a first protease and the second cleavage site is cleaved by a second protease. As a non-limiting example, the peptide linker may comprise a first cleavage site cleaved by a proteolytic enzyme and uPa and a second cleavage site cleaved by MMP. In certain embodiments, the peptide linker comprises a first cleavage site, a second cleavage site, and a third cleavage site, where the first cleavage site is cleaved by a first protease, the second cleavage site is cleaved by a second protease, and the third cleavage site is cleaved by a third protease.

[0249] Exemplary proteolytic enzymes and their recognition sequences that can be used for the fusion proteins herein can be identified by one of ordinary skill in the art and are known in the art, such as those described in the MEROPS database (see, e.g., Rawlings, et al. Nucleic Acids Research, Vol. 46, D1, January 4, 2018, pp. D624–D632), and elsewhere (Hoadley et al., Cell, 2018; GTEx Consortium, Nature, 2017; Robinson et al., Nature, 2017).

[0250] Other methods can also be used to identify cleavage sites for use herein, such as those described in U.S. Patent Nos. 9,453,078, 10,138,272, 9,562,073, and Published International Applications Nos. WO 2015 / 048329, WO2015116933, WO2016118629.

[0251] Accordingly, one embodiment of the present disclosure provides a fusion protein comprising at least two peptide linkers, wherein at least one of the peptide linkers comprises one or more cleavage sites listed herein. In one embodiment, the present disclosure provides a fusion protein comprising a peptide linker, wherein the peptide linker comprises a protease cleavage site and is cleavable by uPA. In one embodiment, the present disclosure provides a fusion protein comprising a peptide linker, wherein the peptide linker comprises the amino acid sequence MSGRSANA (SEQ ID NO: 28). In certain embodiments, the peptide linker sequence comprises at least one protease cleavage site selected from TSGRSANP, LSGRSDNH, GSGRSAQV, GSSRNADV, GTARSDNV, GTARSDNV, GGGRVNNV, MSARILQV, or GKGRSANA (SEQ ID NOs: 30 - 37, respectively).

[0252] In certain embodiments, the fusion protein comprising the peptide linker described herein comprises two heterologous polypeptides: a first polypeptide at the amino (N)-terminal of the peptide linker and a second polypeptide at the carboxyl (C)-terminal of the peptide linker, and the two heterologous polypeptides are thus separated by the peptide linker.

[0253] In certain embodiments, the fusion protein comprises at least one peptide linker that does not comprise a protease cleavage site. In certain embodiments, the peptide linker comprises the amino acid sequence (EAAAK)n, where n is an integer from 1 to 5. In some embodiments, the peptide linker is EAAAK (SEQ ID NO: 39). In some embodiments, the peptide linker is EAAAKEAAAK (SEQ ID NO: 38). In some embodiments, the peptide linker comprises a polyproline linker, optionally having the amino acid sequence of PPP (SEQ ID NO: 41) or PPPP (SEQ ID NO: 40). In certain embodiments, the linker is a glycine (G)-proline (P) polypeptide linker, optionally GPPPG, GGPPPGG, GPPPPG, or GGPPPGG. In certain embodiments, the peptide linker is a Gly n Ser linker. In certain embodiments, the peptide linker comprises (Gly3Ser) n (Gly4Ser)1, (Gly3Ser)1(Gly4Ser) n , (Gly3Ser) n (Gly4Ser) n or (Gly4Ser) n amino acid sequence, where n is an integer from 1 to 5. In certain embodiments, peptide linkers suitable for linking different domains include sequences comprising glycine-serine linkers, such as, but not limited to, (Gm S) n -GG, (SGn)m, (SEGn)m, where m and n range from 0 - 20.

[0254] In certain embodiments, the peptide linker is an amino acid sequence obtained, derived, or designed from: an antibody hinge region sequence, a sequence that links a binding domain to a receptor, or a sequence that links a binding domain to a cell surface transmembrane region or membrane anchor. In some embodiments, the peptide linker has at least one cysteine capable of participating in at least one disulfide bond under physiological conditions or other standard peptide conditions (e.g., peptide purification conditions, conditions for peptide storage). In certain embodiments, a peptide linker corresponding to or similar to an immunoglobulin hinge peptide retains a cysteine corresponding to the hinge cysteine arranged at the amino - terminal end towards the hinge. In further embodiments, the peptide linker is from an IgG1 hinge and has been modified to remove any cysteine residues or is an IgG1 hinge having one or two cysteines corresponding to the hinge cysteines.

[0255] In certain embodiments, the peptide linker for use herein may comprise an "altered wild - type immunoglobulin hinge region" or an "altered immunoglobulin hinge region". Such altered hinge regions refer to (a) a wild - type immunoglobulin hinge region having up to 30% amino acid variation (e.g., up to 25%, 20%, 15%, 10%, or 5% amino acid substitutions or deletions), (b) a portion of a wild - type immunoglobulin hinge region having up to 30% amino acid variation (e.g., up to 25%, 20%, 15%, 10%, or 5% amino acid substitutions or deletions) and having a length of at least 10 amino acids (e.g., at least 12, 13, 14, or 15 amino acids), or (c) a portion of a wild - type immunoglobulin hinge region that includes the core hinge region (the length of this portion can be 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15, or at least 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, or 15 amino acids). In certain embodiments, one or more cysteine residues in a wild - type immunoglobulin hinge region (such as an IgG1 hinge that includes an upper region and a core region) may be replaced by one or more other amino acid residues (e.g., one or more serine residues). The altered immunoglobulin hinge region may alternatively or additionally have a proline residue in the wild - type immunoglobulin hinge region (such as an IgG1 hinge that includes an upper and a core region) replaced by another amino acid residue (e.g., a serine residue).

[0256] Alternative hinge and linker sequences that can be used as a linking region can be made from the portion of a cell surface receptor that connects IgV-like or IgC-like domains. The regions between IgV-like domains of a cell surface receptor containing multiple tandem IgV-like domains and the regions between IgC-like domains of a cell surface receptor containing multiple tandem IgC-like regions can also be used as a linking region or a linker peptide. In certain embodiments, the hinge and linker sequences have a length of 5 to 60 amino acids and can be predominantly flexible, but can also provide more rigid characteristics and can predominantly comprise helical structures with minimal beta-sheet structure.

[0257] In certain embodiments, the protease described herein is expressed at a higher amount in vivo near a particular target cell of interest (e.g., the tumor microenvironment of a target tumor cell). A variety of different medical conditions or diseases are known in which a target of interest (such as a particular tumor type, a particular tumor expressing a specific tumor-associated antigen) is co-localized with a protease, and the substrate of the protease is known in the art. In an example of cancer, the target tissue can be cancerous tissue, particularly cancerous tissue of a solid tumor. Elevated protease levels in many cancers (e.g., liquid tumors or solid tumors) have been reported in the literature. See, e.g., LaRocca et al., (2004) British J. of Cancer 90(7):1414-1421.

[0258] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker-VL, receptor-linker-VL, ligand-linker-VH, or receptor-linker-VH.

[0259] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-cleavable linker-VL, receptor-cleavable linker-VL, ligand-cleavable linker-VH, or receptor-cleavable linker-VH.

[0260] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:114)-VL, receptor-linker (SEQ ID NO:114)-VL, ligand-linker (SEQ ID NO:14)-VH, or receptor-linker (SEQ ID NO:14)-VH.

[0261] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:145)-VL, receptor-linker (SEQ ID NO:145)-VL, ligand-linker (SEQ ID NO:145)-VH, or receptor-linker (SEQ ID NO:145)-VH.

[0262] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:147)-VL, receptor-linker (SEQ ID NO:147)-VL, ligand-linker (SEQ ID NO:147)-VH, or receptor-linker (SEQ ID NO:147)-VH.

[0263] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:154)-VL, receptor-linker (SEQ ID NO:154)-VL, ligand-linker (SEQ ID NO:154)-VH, or receptor-linker (SEQ ID NO:154)-VH.

[0264] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker (SEQ ID NO:203)-VL, receptor-linker (SEQ ID NO:203)-VL, ligand-linker (SEQ ID NO:203)-VH, or receptor-linker (SEQ ID NO:203)-VH.

[0265] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker-Fc or receptor-linker-Fc.

[0266] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-cleavable linker-Fc or receptor-cleavable linker-Fc.

[0267] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-cleavable linker (SEQID NO:28)-Fc or receptor-cleavable linker (SEQ ID NO:28)-Fc.

[0268] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-linker-Fc1 or receptor-linker-Fc1.

[0269] In certain embodiments, the fusion protein comprises, from the N-terminus to the C-terminus, ligand-cleavable linker-Fc2 or receptor-cleavable linker-Fc2.

[0270] In certain embodiments, Fc1 and Fc2 can form a heterodimer. In certain embodiments, Fc1 is linked to a ligand and Fc2 is linked to a receptor. In certain embodiments, the linker that links the ligand to Fc1 is cleavable and the linker that links the receptor to Fc2 is non-cleavable. In certain embodiments, the linker that links the ligand to Fc1 is non-cleavable and the linker that links the receptor to Fc2 is cleavable. In certain embodiments, the linker that links the ligand to Fc1 is cleavable and the linker that links the receptor to Fc2 is cleavable. In certain embodiments, the linker that links the ligand to Fc1 is non-cleavable and the linker that links the receptor to Fc2 is non-cleavable.

[0271] Target

[0272] In some embodiments, the antigen-binding domain of the fusion proteins described herein binds specifically to a cell surface molecule. In certain embodiments, the antigen-binding domain of the fusion protein binds specifically to a tumor-associated antigen (TAA). The TAA is any antigenic substance expressed on the surface of tumor cells. In some embodiments, the antigen-binding domain binds specifically to a TAA selected from the group consisting of fibroblast activation protein alpha (FAPα), trophoblast glycoprotein (5T4), tumor-associated calcium signal transducer 2 (Trop2), fibronectin EDB (EDB-FN), fibronectin F.IIIB domain, CGS-2, EpCAM, EGER, HER-2, HER-3, cMet, CEA, and FOLR1, EpCAM, EGFR, HER-2, HER-3, cMet, CEA, and FOLR1, EpCAM, EGFR, HER-2, HER-3, c-Met, FOLR1, PSMA, CD38, BCMA, and CEA. 5T4, AFP, B7-H3, cadherin-6, CAIX, CD117, CD123, CD138, CD166, CD19, CD20, CD205, CD22, CD30, CD33, CD40, CD352, CD37, CD44, CD52, CD56, CD70, CD71, CD74, CD79b, DLL3, DR5, EphA2, FAP, FGFR2, FGFR3, GPC3, gpA33, FLT-3, gpNMB, HPV-16E6, HPV-16E7, ITGA2, ITGA3, SLC39A6, MAGE, mesothelin (MSLN), Muc1, Muc16, NaPi2b, nectin-4, P-cadherin, NY-ESO-1, PRLR, PSCA, PTK7, ROR1, SLC44A4, SLTRK5, SLTRK6, STEAP1, TIM1, tissue factor (TF), Trop2, WT1.

[0273] In some embodiments, the antigen-binding domain binds specifically to an immune checkpoint protein. Examples of immune checkpoint proteins include, but are not limited to, CD27, CD137, 2B4, TIGIT, CD155, ICOS, HVEM, CD40L, LIGHT, TIM-1, OX40, DNAM-1, PD-L1, PD1, PD-L2, CTLA-4, CD80, CD40, CEACAM1, CD48, CD70, A2AR, CD39, CD73, B7-H3, B7-H4, BTLA, IDO1, IDO2, TDO, KIR, LAG-3, TIM-3, VISTA, CD47, or SIRPα.

[0274] In some embodiments, the antigen-binding domain specifically binds to an antigen expressed on a virus-infected cell, a bacterium-infected cell, a damaged red blood cell, an arterial plaque cell, an inflamed or fibrotic tissue cell.

[0275] In certain embodiments, the antigen-binding domain specifically binds to a cytokine receptor. Examples of cytokine receptors include, but are not limited to, type I cytokine receptors such as GM-CSF receptor, G-CSF receptor, type I IL receptor, Epo receptor, LIF receptor, CNTF receptor, TPO receptor; type II cytokine receptors such as IFN-α receptor (IFNAR1, IFNAR2), IFN-β receptor, IFN-γ receptor (IFNGR1, IFNGR2), type II IFN receptor; chemokine receptors such as CC chemokine receptor, CXC chemokine receptor, CX3C chemokine receptor, XC chemokine receptor; tumor necrosis factor receptor superfamily receptors such as TNFRSF5 / CD40, TNFRSF8 / CD30, TNFRSF7 / CD27, TNFRSF1A / TNFR1 / CD120a, TNFRSF1B / TNFR2 / CD120b; TGF-β receptors such as TGF-β receptor 1, TGF-β receptor 2; Ig superfamily receptors such as IF-1 receptor, CSF-1R, PDGFR (PDGFRA, PDGFRB), SCFR.

[0276] In certain embodiments, the antigen-binding domain of the fusion protein described herein specifically binds in vivo to at least one molecule or target of interest. In certain embodiments, the target of interest is cluster of differentiation 3 (CD3), human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), mesothelin (MSLN), tissue factor (TF), cluster of differentiation 19 (CD19), tyrosine-protein kinase Met (c-Met), cluster of differentiation 40 (CD40), cadherin 3 (CDH3), or a combination thereof. In certain embodiments, the fusion protein comprises an antibody and at least one antigen-binding domain of the antibody specifically binds to an epitope on CD3, HER2, EGFR, MSLN, TF, CD19, c-Met, CD40, CDH3, or a combination thereof.

[0277] In some embodiments, the target of interest is HER2, and the anti-HER2 complementarity-determining region of the fusion protein has: a VH having an amino acid sequence corresponding to SEQ ID NO: 120 and a VL having an amino acid sequence corresponding to SEQ ID NO: 124. In certain embodiments, the anti-HER2 complementarity-determining region has a VH amino acid sequence substantially identical to SEQ ID NO: 120 and a VL amino acid sequence substantially identical to SEQ ID NO: 124. In certain embodiments, the anti-HER2 complementarity-determining region has a VH amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 120 and a VL amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 124. In certain embodiments, the anti-HER2 complementarity-determining region has a VH amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 120 and a VL amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 124. In some embodiments, the anti-HER2 complementarity-determining region comprises an scFv having an amino acid sequence corresponding to SEQ ID NO: 3. In some embodiments, the anti-HER2 has: a VH having three CDRs (i.e., HCDR1, HDR2, and HCDR3 having amino acid sequences corresponding to SEQ ID NOs: 121, 122, and 123, respectively) and a VL having three CDRs (i.e., LCDR1, LCDR2, and LCDR3 having amino acid sequences corresponding to SEQ ID NOs: 125, 126, and 127, respectively).

[0278] In some embodiments, the target of interest is EGFR, and the anti-EGFR complementarity determining regions (CDRs) of the fusion protein have: a VH having an amino acid sequence corresponding to SEQ ID NO:14 and a VL having an amino acid sequence corresponding to SEQ ID NO:13. In certain embodiments, the anti-EGFR CDRs have a VH amino acid sequence substantially identical to SEQ ID NO:14 and a VL amino acid sequence substantially identical to SEQ ID NO:13. In certain embodiments, the anti-EGFR CDRs have a VH amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:14 and a VL amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO:13. In certain embodiments, the anti-EGFR CDRs have a VH amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:14 and a VL amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO:13. In some embodiments, the anti-EGFR has: a VH having 3 CDRs (i.e., HCDR1, HCDR2, and HCDR3 having amino acid sequences corresponding to SEQ ID NOs:84, 85, and 86, respectively) and a VL having 3 CDRs (i.e., LCDR1, LCDR2, and LCDR3 having amino acid sequences corresponding to SEQ ID NOs:59, 60, and 61, respectively).

[0279] In some embodiments, the target of interest is MSLN, and the MSLN - specific binding moiety of the fusion protein has: a VH having an amino acid sequence corresponding to SEQ ID NO:16 and a VL having an amino acid sequence corresponding to SEQ ID NO:15. In certain embodiments, the MSLN - specific binding moiety has a VH amino acid sequence that is substantially identical to SEQ ID NO:16 and a VL amino acid sequence that is substantially identical to SEQ ID NO:15. In certain embodiments, the MSLN - specific binding moiety has a VH amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO:16 and a VL amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO:15. In certain embodiments, the MSLN - specific binding moiety has a VH amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO:16 and a VL amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO:15. In some embodiments, the anti - MSLN has: a VH having 3 CDRs (i.e., HCDR1, HCDR2, and HCDR3 having amino acid sequences corresponding to SEQ ID NOs:69, 70, and 71, respectively) and a VL having 3 CDRs (i.e., LCDR1, LCDR2, and LCDR3 having amino acid sequences corresponding to SEQ ID NOs:74, 75, and 76, respectively).

[0280] In some embodiments, the target of interest is TF (tissue factor), and the anti-TF complementarity determining region of the fusion protein has: a VH having an amino acid sequence corresponding to SEQ ID NO: 18 and a VL having an amino acid sequence corresponding to SEQ ID NO: 17. In certain embodiments, the anti-TF complementarity determining region has a VH amino acid sequence substantially identical to SEQ ID NO: 18 and a VL amino acid sequence substantially identical to SEQ ID NO: 17. In certain embodiments, the anti-TF complementarity determining region has a VH amino acid sequence about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO: 18 and a VL amino acid sequence about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO: 17. In certain embodiments, the anti-TF complementarity determining region has a VH amino acid sequence about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO: 18 and a VL amino acid sequence about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO: 17. In some embodiments, the anti-TF has: a VH having 3 CDRs (i.e., HCDR1, HCDR2 and HCDR3 having amino acid sequences corresponding to SEQ ID NOs: 54, 55 and 56 respectively) and a VL having 3 CDRs (i.e., LCDR1, LCDR2 and LCDR3 having amino acid sequences corresponding to SEQ ID NOs: 48, 49 and 50 respectively).

[0281] In some embodiments, the target of interest is CD19, and the anti-CD19 complementarity-determining region of the fusion protein has: a VH having an amino acid sequence corresponding to SEQ ID NO: 20 and a VL having an amino acid sequence corresponding to SEQ ID NO: 19. In certain embodiments, the anti-CD19 complementarity-determining region has a VH amino acid sequence substantially identical to SEQ ID NO: 20 and a VL amino acid sequence substantially identical to SEQ ID NO: 19. In certain embodiments, the anti-CD19 complementarity-determining region has a VH amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO: 20 and a VL amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO: 19. In certain embodiments, the anti-CD19 complementarity-determining region has a VH amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO: 20 and a VL amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO: 19. In some embodiments, the anti-CD19 has: a VH having 3 CDRs (i.e., HCDR1, HCDR2 and HCDR3 having amino acid sequences corresponding to SEQ ID NOs: 64, 65 and 66, respectively) and a VL having 3 CDRs (i.e., LCDR1, LCDR2 and LCDR3 having amino acid sequences corresponding to SEQ ID NOs: 74, 75 and 165, respectively).

[0282] In some embodiments, the target of interest is c-Met, and the anti-c-Met complementarity determining region (CDR) of the fusion protein has: a VH having an amino acid sequence corresponding to SEQ ID NO: 22 and a VL having an amino acid sequence corresponding to SEQ ID NO: 21. In certain embodiments, the anti-c-Met CDR has a VH amino acid sequence substantially identical to SEQ ID NO: 22 and a VL amino acid sequence substantially identical to SEQ ID NO: 21. In certain embodiments, the anti-c-Met CDR has a VH amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 22 and a VL amino acid sequence that is about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 21. In certain embodiments, the anti-c-Met CDR has a VH amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 22 and a VL amino acid sequence that is about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 21. In some embodiments, the anti-c-Met has: a VH having three CDRs (i.e., HCDR1, HCDR2, and HCDR3 having amino acid sequences corresponding to SEQ ID NOs: 99, 100, and 101, respectively) and a VL having three CDRs (i.e., LCDR1, LCDR2, and LCDR3 having amino acid sequences corresponding to SEQ ID NOs: 94, 95, and 96, respectively).

[0283] In some embodiments, the target of interest is CDH3, and the anti-CDH3 complementarity determining region of the fusion protein has: a VH having an amino acid sequence corresponding to SEQ ID NO: 24 and a VL having an amino acid sequence corresponding to SEQ ID NO: 23. In certain embodiments, the anti-CDH3 complementarity determining region has a VH amino acid sequence substantially identical to SEQ ID NO: 24 and a VL amino acid sequence substantially identical to SEQ ID NO: 23. In certain embodiments, the anti-CDH3 complementarity determining region has a VH amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO: 24 and a VL amino acid sequence that is about 80%, about 85%, about 90% or about 95% identical to SEQ ID NO: 23. In certain embodiments, the anti-CDH3 complementarity determining region has a VH amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO: 24 and a VL amino acid sequence that is about 96%, about 97%, about 98% or about 99% identical to SEQ ID NO: 23. In some embodiments, the anti-CDH3 has: a VH having 3 CDRs (i.e., HCDR1, HCDR2 and HCDR3 having amino acid sequences corresponding to SEQ ID NOs: 89, 90 and 91, respectively) and a VL having 3 CDRs (i.e., LCDR1, LCDR2 and LCDR3 having amino acid sequences corresponding to SEQ ID NOs: 94, 95 and 96, respectively).

[0284] In some embodiments, the target of interest is CD40, and the anti-CD40 complementarity determining regions of the fusion protein have: a VH having an amino acid sequence corresponding to SEQ ID NO: 172 and a VL having an amino acid sequence corresponding to SEQ ID NO: 177. In certain embodiments, the anti-CD40 complementarity determining regions have a VH amino acid sequence substantially identical to SEQ ID NO: 172 and a VL amino acid sequence substantially identical to SEQ ID NO: 177. In certain embodiments, the anti-CD40 complementarity determining regions have a VH amino acid sequence about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 172 and a VL amino acid sequence about 80%, about 85%, about 90%, or about 95% identical to SEQ ID NO: 177. In certain embodiments, the anti-CD40 complementarity determining regions have a VH amino acid sequence about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 172 and a VL amino acid sequence about 96%, about 97%, about 98%, or about 99% identical to SEQ ID NO: 177. In some embodiments, the anti-CD40 has: a V having 3 CDRs (i.e., HCDR1, HCDR2, and HCDR3 having amino acid sequences corresponding to SEQ ID NOs: 173, 174, and 175, respectively) and a VL having 3 CDRs (i.e., LCDR1, LCDR2, and LCDR3 having amino acid sequences corresponding to SEQ ID NOs: 178, 179, and 180, respectively).

[0285] In certain embodiments, the antigen-binding domain of the fusion protein specifically binds to a molecule (e.g., a polypeptide) on an immune cell. In certain embodiments, the fusion protein comprises an antigen-binding domain that specifically binds to a TAA and an antigen-binding domain that specifically binds to a molecule (e.g., a polypeptide) on an immune cell. Thus, in certain embodiments, the fusion protein binds to both tumor cells and immune cells. In certain embodiments, the immune cell is a T cell. In certain embodiments, the immune cell is a macrophage, dendritic cell, neutrophil, B cell, or NK cell.

[0286] In certain embodiments, the fusion protein binds to the CD3 antigen on a T cell and one or more TAAs on a tumor cell.

[0287] Masked T cell engager

[0288] T cell engagers (TCEs) are polypeptide constructs, often bispecific antibodies, which can simultaneously bind a TAA on a tumor cell and the CD3 epitope on a T cell, thereby forming an artificial immune synapse independent of the TCR. This results in activation of the T cell and cytotoxic effects on the tumor cell. The efficacy of bispecific antibodies capable of targeting T cells to tumor cells has been identified and tested in cancer therapy. Blinatumomab is an example of a bispecific anti-CD3-CD19 antibody in the format called BiTE TM (bispecific T cell engager), which has been identified for the treatment of B cell diseases such as relapsed B cell non-Hodgkin lymphoma and chronic lymphocytic leukemia (Baeuerle et al. (2009) Cancer Research 12:4941-4944) and has received FDA approval. T cell engagers against other tumor-associated target antigens have also been made and several have entered clinical trials: AMG110 / MT110 EpCAM for lung, gastric, and colorectal cancer; AMG211 / MEDI565 CEA for gastrointestinal adenocarcinoma; and AMG 212 / BAY2010112 PSMA for prostate cancer (see Suruadevara, C.M. et al., Oncoimmunology. June 2015; 4(6):e1008339). Although these studies have shown promising clinical efficacy, they have also been hampered by severe dose-limiting toxicities mainly caused by cytokine release syndrome (CRS). This results in a narrow therapeutic window. Using a masked T cell-binding paratope that is mainly activated in the tumor microenvironment may reduce the toxicity of the TCE.

[0289] In certain embodiments, the fusion protein binds the CD3 antigen on a T cell and the TAA on a tumor cell. In certain embodiments, the fusion protein binds the CD3 antigen on a T cell, the TAA on a tumor cell, and the IgSF extracellular domain on a tumor cell. In certain embodiments, the fusion protein binds the CD3 antigen on a T cell, the TAA on a tumor cell, and the IgSF extracellular domain on a T cell.

[0290] In certain embodiments, the fusion protein is protease-unmasked in the tumor microenvironment and binds the TAA on the tumor cell and the CD3 antigen on the T cell, resulting in bridging of the T cell and the tumor cell, as demonstrated in Example 20. In certain embodiments, the unmasked fusion protein binds the CD3 antigen on the T cell and both the TAA and the IgSF ligand on the tumor cell, as illustrated in Figure 31. In certain embodiments, binding of the IgSF ligand (e.g., PD-L1) on the tumor cell prevents binding of its IgSF receptor (e.g., PD-1) on the T cell, thereby blocking checkpoint inhibition (Figure 31C )。

[0291] In certain embodiments, the fusion protein comprises anti-CD3 complementarity determining regions VH and VL that are substantially identical to the VH and VL of the complementarity determining regions shown in Table BB. In certain embodiments, the CD3 complementarity determining regions comprise the following VH and VL amino acid sequences:

[0292] (a) a VH comprising the amino acid sequence corresponding to SEQ ID NO:2 and a VL comprising the amino acid sequence according to SEQ ID NO:1;

[0293] (b) a VH comprising the amino acid sequence corresponding to SEQ ID NO:206 and a VL comprising the amino acid sequence according to SEQ ID NO:210;

[0294] (c) a VH comprising the amino acid sequence corresponding to SEQ ID NO:215 and a VL comprising the amino acid sequence according to SEQ ID NO:219;

[0295] (d) a VH comprising the amino acid sequence corresponding to SEQ ID NO:223 and a VL comprising the amino acid sequence according to SEQ ID NO:227;

[0296] (d) a VH comprising the amino acid sequence corresponding to SEQ ID NO:231 and a VL comprising the amino acid sequence according to SEQ ID NO:235; or

[0297] (e) a VH comprising the amino acid sequence corresponding to SEQ ID NO:239 and a VL comprising the amino acid sequence according to SEQ ID NO:243.

[0298] In certain embodiments, the CD3 complementarity determining regions comprise VH and VL that are about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 96%, about 97%, about 98% or about 99% identical to the following:

[0299] (a) a VH comprising the amino acid sequence corresponding to SEQ ID NO:2 and a VL comprising the amino acid sequence according to SEQ ID NO:1;

[0300] (b) a VH comprising the amino acid sequence corresponding to SEQ ID NO:206 and a VL comprising the amino acid sequence according to SEQ ID NO:210;

[0301] (c) a VH comprising the amino acid sequence corresponding to SEQ ID NO:215 and a VL comprising the amino acid sequence according to SEQ ID NO:219;

[0302] a VH comprising the amino acid sequence corresponding to SEQ ID NO: 223 and a VL comprising the amino acid sequence according to SEQ ID NO: 227;

[0303] a VH comprising the amino acid sequence corresponding to SEQ ID NO: 231 and a VL comprising the amino acid sequence according to SEQ ID NO: 235; or

[0304] a VH comprising the amino acid sequence corresponding to SEQ ID NO: 239 and a VL comprising the amino acid sequence according to SEQ ID NO: 243.

[0305] In certain embodiments, the anti-CD3 complementarity determining regions comprise: a VH comprising three heavy chain CDRs (i.e., HCDR1, HCDR2, and HCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 207, 208, and 209) and a VL comprising three light chain CDRs (i.e., LCDR1, LCDR2, and LCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 211, 212, and 214). In certain embodiments, the anti-CD3 complementarity determining regions comprise: a VH comprising three heavy chain CDRs (i.e., HCDR1, HCDR2, and HCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 224, 225, and 226) and a VL comprising three light chain CDRs (i.e., LCDR1, LCDR2, and LCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 228, 229, and 230). In certain embodiments, the anti-CD3 complementarity determining regions comprise: a VH comprising three heavy chain CDRs (i.e., HCDR1, HCDR2, and HCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 232, 233, and 234) and a VL comprising three light chain CDRs (i.e., LCDR1, LCDR2, and LCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 236, 237, and 238). In certain embodiments, the anti-CD3 complementarity determining regions comprise: a VH comprising three heavy chain CDRs (i.e., HCDR1, HCDR2, and HCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 240, 241, and 242) and a VL comprising three light chain CDRs (i.e., LCDR1, LCDR2, and LCDR3 comprising the amino acid sequences corresponding to SEQ ID NOs: 244, 245, and 246).

[0306] CAR construct

[0307] In certain embodiments, the fusion protein may be included in a chimeric antigen receptor (CAR) or a CAR fragment. The CAR may include one or more extracellular ligand-binding domains, optionally a hinge region, a transmembrane region, and an intracellular signaling region. The one or more extracellular ligand-binding domains may include one or more fusion proteins. The extracellular ligand-binding domain typically may include a single-chain immunoglobulin variable fragment (scFv) or other ligand-binding domains, such as Fab or a native protein ligand. The hinge region typically may include a polypeptide hinge of variable length (such as one or more amino acids), a CD8a hinge region, or an IgG4 region (or others), and combinations thereof. The transmembrane domain typically may include a transmembrane region derived from CD8α, CD28, or other transmembrane proteins (such as DAP10, DAP12, or NKG2D), and combinations thereof. The intracellular signaling region may include one or more intracellular signaling domains, such as CD28, 4-1BB, CD3ζ, OX40, 2B4, or other intracellular signaling domains and combinations thereof. For example, the one or more intracellular signaling domains may include CD28 and CD3ζ, 4-1BB and CD3ζ, or CD3ζ. Lymphocytes such as T cells and NK cells can be modified to produce chimeric antigen receptor cells (e.g., CAR-T). CAR-T cells can recognize specific soluble antigens or antigens on the surface of target cells (such as the surface of tumor cells) or cells in the tumor microenvironment. When the extracellular ligand-binding domain binds to a cognate ligand, the intracellular signaling domain of the CAR can activate lymphocytes. See, e.g., Brudno et al., Nature Rev. Clin. Oncol. (2018) 15:31-46; Maude et al., N. Engl. J. Med. (2014) 371:1507-1517; Sadelain et al., Cancer Disc. (2013) 3:388-398 (2018); U.S. Patent Nos. 7,446,190 and 8,399,645.

[0308] In certain embodiments, a CAR construct is provided that comprises a ligand-receptor pair construct as described herein. In certain embodiments, the CAR construct comprises a scFv that can be fused to the ligand-receptor pair construct. In certain embodiments, the ligand-receptor pair construct is a single-chain ligand-receptor pair construct that can be fused to the N-terminus of the scFv, with or without a linker. In certain embodiments, the single-chain ligand-receptor pair construct comprises a protease-cleavable linker. In certain embodiments, the receptor is fused to the N-terminus of the scFv, with or without a first linker, and the ligand is fused internally to a second linker that connects the heavy and light chains of the scFv. In certain embodiments, the linker comprises a protease cleavage site that can be cleaved by a protease. In certain embodiments, the ligand is fused to the N-terminus of the scFv, with or without a first linker, and the receptor is fused internally to a second linker that connects the heavy and light chains of the scFv. In certain embodiments, the first linker is cleavable and the second linker is not protease-cleavable. In certain embodiments, T cells can be modified to express a ligand-receptor pair CAR.

[0309] Sequence homology

[0310] Certain embodiments of the present disclosure relate to isolated polynucleotides or polynucleotide collections that encode fusion proteins as described herein. Polynucleotides in this context can encode all or part of the fusion protein.

[0311] The terms "nucleic acid", "nucleic acid molecule", and "polynucleotide" are used interchangeably herein and refer to nucleotides in polymeric form of any length, which are deoxyribonucleotides or ribonucleotides or analogs thereof. Non-limiting examples of polynucleotides include genes, gene fragments, messenger RNA (mRNA), cDNA, recombinant polynucleotides, plasmids, vectors, isolated DNA of any sequence, isolated RNA of any sequence, nucleic acid probes, and primers.

[0312] A polynucleotide that "encodes" a given polypeptide is a polynucleotide that, when placed under the control of appropriate regulatory sequences, is transcribed (in the case of DNA) or translated (in the case of mRNA) in vivo into the polypeptide. The boundaries of the coding sequence are determined by the start codon at the 5' (amino) terminus and the translation stop codon at the 3' (carboxyl) terminus. The transcription termination sequence can be located 3' of the coding sequence.

[0313] In certain embodiments, the present disclosure relates to polynucleotide and polypeptide sequences that are identical or substantially identical to at least a portion of a polypeptide encoding a fusion protein described herein, such as a first or second polypeptide of a biologically functional protein. In the context of two or more polynucleotide or polypeptide sequences, the term "identical" refers to two or more identical sequences or subsequences. When sequences are compared and aligned to obtain maximum correspondence as measured over a comparison window or over a specified region using one of the commonly used sequence comparison algorithms known to those of ordinary skill in the art or by manual alignment and visual inspection, the sequences are "substantially identical" if they have a certain percentage of identical amino acid residues or nucleotides (e.g., about 80%, about 85%, about 90%, or about 95% identity within the specified region). This definition also refers to the components of the test polynucleotide sequence. Identity can exist over a region of at least about 50 amino acids or nucleotides in length, or over a region of 75 - 100 amino acids or nucleotides in length, or, when not specified, over the entire sequence of the polynucleotide or polypeptide. For sequence comparison, a test sequence is generally compared to a designated reference sequence. When using a sequence comparison algorithm, the test sequence and the reference sequence are entered into a computer, and subsequence coordinates are specified if necessary, and sequence algorithm program parameters are designated. Default program parameters can be used, or alternative parameters can be specified. The sequence comparison algorithm then calculates the percentage of sequence identity of the test sequence relative to the reference sequence based on the program parameters.

[0314] As used herein, a "comparison window" refers to a sequence segment that encompasses contiguous amino acid or nucleotide positions, which can range from 20 to 1000 contiguous amino acid or nucleotide positions, such as from about 50 to about 600 or from about 100 to about 300 or from about 150 to about 200 contiguous amino acid or nucleotide positions, and in which the two sequences can be compared after the test sequence is optimally aligned with a reference sequence having the same number of contiguous positions. In certain embodiments, longer segments up to and including the full-length sequence may also be used as the comparison window. Methods of sequence alignment for comparison are known to those of ordinary skill in the art. Optimal sequence alignments for comparison can be performed, for example, by the local homology algorithm of Smith & Waterman, 1970, Adv. Appl. Math., 2:482; the homology alignment algorithm of Needleman & Wunsch, 1970, J. Mol. Biol., 48:443; the similarity search method of Pearson & Lipman, 1988, Proc. Natl. Acad. Sci. USA, 85:2444; or computerized implementations of these algorithms (e.g., GAP, BESTFIT, FASTA, or TFASTA in the Wisconsin Genetics Software Package, Genetics Computer Group, Madison, WI), or by manual alignment and visual inspection (see, e.g., Ausubel et al., Current Protocols in Molecular Biology, (Supplement 1995), Cold Spring Harbor Laboratory Press). Examples of available algorithms suitable for determining the percent sequence identity are the BLAST and BLAST 2.0 algorithms, which are described in Altschul et al., 1997, Nuc. Acids Res., 25:3389-3402 and Altschul et al., 1990, J. Mol. Biol., 215:403-410, respectively. Software for performing BLAST analyses is publicly available through the website of the National Center for Biotechnology Information (NCBI).

[0315] Certain embodiments described herein relate to variant sequences comprising one or more amino acid substitutions. In some embodiments, the amino acid substitutions are conservative amino acid substitutions. Generally, a "conservative substitution" is considered to be the replacement of one amino acid with another amino acid having similar physical, chemical, and / or structural properties. Common conservative substitutions are listed in column 1 of Table 4. Those skilled in the art will understand that the primary factors determining what constitutes a conservative substitution are typically the size of the amino acid side chain and its physical / chemical properties, but certain circumstances allow for the replacement of a given amino acid with a broader range of amino acids than those listed in column 1. These additional amino acids tend to have similar properties to the amino acid being replaced, but vary more in size, or have similar size but greater differences in physical / chemical properties. A broader range of such conservative substitutions is listed in column 2 of Table 4. The skilled person can readily determine the most appropriate set of substituents to choose, given the specific protein context in which the amino acid substitution is being made.

[0316] Table 4: Conservative Amino Acid Substitutions

[0317]

[0318]

[0319] Preparation of fusion protein

[0320] The fusion proteins described herein can be produced using standard recombinant methods known in the art (see, e.g., U.S. Patent No. 4,816,567 and "Antibodies: A Laboratory Manual," 2nd ed., edited by Greenfield, Cold Spring Harbor Laboratory Press, New York, 2014).

[0321] Vector encoding fusion protein

[0322] To recombinantly produce the fusion proteins described herein, polynucleotides or sets of polynucleotides encoding the fusion proteins are generated and inserted into one or more vectors for further cloning and / or expression in a host cell. One or more polynucleotides encoding the fusion protein can be produced by standard methods known in the art (see, e.g., Ausubel et al., Current Protocols in Molecular Biology, John Wiley & Sons, New York, 1994 & updates, and “Antibodies: A Laboratory Manual,” 2nd ed., Greenfield ed., Cold Spring Harbor Laboratory Press, New York, 2014). As will be understood by those skilled in the art, the number of polynucleotides required to express the fusion protein will depend on the format of the fusion protein, including whether the fusion protein contains an antibody and the number of polypeptides within the fusion protein. For example, when the fusion protein contains two polypeptide chains, two polynucleotides each encoding one polypeptide chain will be required. Similarly, in certain embodiments, when the fusion protein contains a biologically functional protein in an mAb format, two polynucleotides each encoding one polypeptide chain are required. When multiple polynucleotides are needed, they can be incorporated into one vector or into more than one vector.

[0323] Generally, for expression, the polynucleotide or set of polynucleotides is incorporated into an expression vector along with one or more regulatory elements, such as transcriptional elements required for efficient transcription of the polynucleotide. Examples of such regulatory elements include, but are not limited to, promoters, enhancers, terminators, and polyadenylation signals. Those skilled in the art will understand that the choice of regulatory elements depends on the host cell selected for expressing the polypeptide of the fusion protein, and such regulatory elements can be derived from a variety of sources, including bacterial, fungal, viral, mammalian, or insect genes. The expression vector can optionally further contain heterologous nucleic acid sequences that facilitate the expression or purification of the expressed protein. Examples include, but are not limited to, signal peptides and affinity tags, such as metal affinity tags, histidine tags, avidin / streptavidin encoding sequences, glutathione-S-transferase (GST) encoding sequences, and biotin encoding sequences. The expression vector can be an extrachromosomal vector or an integrating vector.

[0324] Certain embodiments of the present disclosure relate to vectors (such as expression vectors) comprising one or more polynucleotides encoding at least a portion of the fusion proteins described herein. The one or more polynucleotides can be contained in a single vector or in more than one vector. In some embodiments, the polynucleotides are contained in a polycistronic vector.

[0325] Expression vectors for use in expressing polynucleotides include, but are not limited to, pTT5 and pUC15, cells containing vectors encoding fusion proteins.

[0326] Suitable host cells for cloning or expressing fusion protein polypeptides include a variety of prokaryotic or eukaryotic cells known in the art. Eukaryotic host cells include, for example, mammalian cells, plant cells, insect cells, and yeast cells (such as Saccharomyces or Pichia cells). Prokaryotic host cells include, for example, Escherichia coli, Aeromonas salmonicida, or Bacillus subtilis cells.

[0327] In certain embodiments, the fusion protein is produced in bacteria, particularly when glycosylation and Fc effector functions are not required, as described, for example, in U.S. Pat. Nos. 5,648,237, 5,789,199, and 5,840,523, and Charlton, Methods in Molecular Biology, Vol. 248, pp. 245 - 254, edited by B.K.C. Lo, Humana Press, Totowa, N.J., 2003.

[0328] In certain embodiments, eukaryotic microorganisms such as filamentous fungi or yeast are suitable expression host cells, particularly fungal and yeast strains in which the glycosylation pathway has been "humanized" to result in the production of antibodies with a partially or fully human glycosylation pattern (see, for example, Gerngross, 2004, Nat. Biotech. 22:1409 - 1414, and Li et al., 2006, Nat. Biotech. 24:210 - 215).

[0329] Suitable host cells for expressing glycosylated fusion proteins are generally eukaryotic cells. For example, U.S. Pat. Nos. 5,959,177, 6,040,498, 6,420,548, 7,125,978, and 6,417,429 describe PLANTIBODIES for producing antibodies in transgenic plants. TMTechniques. Mammalian cell lines suitable for suspension growth are particularly useful for the expression of fusion proteins. Examples include, but are not limited to, monkey kidney CV1 lines transformed by SV40 (COS-7), human embryonic kidney (HEK) line 293 or 293 cells (see, e.g., Graham et al., 1977, J. Gen Virol., 36:59), baby hamster kidney cells (BHK), mouse Sertoli TM4 cells (see, e.g., Mather, 1980, Biol Reprod, 23:243-251); monkey kidney cells (CV1), African green monkey kidney cells (VERO-76), human cervical carcinoma (HeLa) cells, canine kidney cells (MDCK), buffalo rat hepatocytes (BRL 3A), human lung cells (W138), human hepatocytes (Hep G2), mouse mammary tumor (MMT060562), TRI cells (see, e.g., Mather et al., 1982, Annals N.Y. Acad Sci, 383:44-68), MRC 5 cells, FS4 cells, Chinese hamster ovary (CHO) cells (including DHFR-CHO cells, see Urlaub et al., 1980, Proc Natl Acad Sci USA, 77:4216), and myeloma cell lines (such as Y0, NS0, and Sp2 / 0). Exemplary mammalian host cell lines suitable for antibody production are reviewed in Yazaki & Wu, Methods in Molecular Biology, Vol. 248, pp. 255-268 (B.K.c.Lo, ed., Humana Press, Totowa, N.J., 2003).

[0330] In certain embodiments, the host cell is a transient or stable higher eukaryotic cell line, such as a mammalian cell line. In some embodiments, the host cell is a mammalian HEK293T, CHO, HeLa, NS0, or COS cell. In some embodiments, the host cell is a stable cell line that permits the mature glycosylation of the fusion protein.

[0331] Host cells containing one or more expression vectors encoding a fusion protein can be cultured using conventional methods to produce the fusion protein. Alternatively, in some embodiments, host cells containing one or more expression vectors encoding a fusion protein can be used therapeutically or prophylactically to deliver the fusion protein to a subject, or a polynucleotide or expression vector can be administered ex vivo to cells from a subject and the cells then returned to the subject's body.

[0332] In some embodiments, the host cell comprises a vector (e.g., has been transformed with a vector) that comprises polynucleotides encoding the VL of the binding domain described herein and the VH of the binding domain. In some embodiments, the host cell comprises a first vector that comprises a polynucleotide encoding the VL of the binding domain described herein; and a second vector that comprises a polynucleotide encoding the corresponding VH of the binding domain. In some embodiments, the host cell is eukaryotic, such as Chinese hamster ovary (CHO) cells, human embryonic kidney (HEK) cells, or lymphoid cells (e.g., Y0, NS0, Sp20 cells).

[0333] In certain embodiments, the host cell is Expi293 TM (Thermo Fisher, Waltham, MA). In certain embodiments, the host cell is CHO-S cells (National Research Council Canada) or HEK293 cells.

[0334] Certain embodiments of the present disclosure relate to methods of preparing a fusion protein, the method comprising culturing a host cell that has been introduced with one or more polynucleotides encoding the fusion protein or one or more expression vectors encoding the fusion protein under conditions suitable for expression of the fusion protein, and optionally recovering the fusion protein from the host cell (or from the host cell culture medium).

[0335] Cell culture media that can be used include, but are not limited to, DMEM (Thermo Fisher, Waltham, MA), Opti-MEM TM (Thermo Fisher, Waltham, MA), Opti-MEM TM I Reduced Serum Medium (Thermo Fisher, Waltham, MA), RPMI-1640 medium, Expi293 TM Expression Medium (Thermo Fisher, Waltham, MA), and FreeStyle CHO Expression Medium (Thermo Fisher Scientific, Waltham, MA).

[0336] The cell culture medium can be supplemented with serum (e.g., fetal bovine serum (FBS)), amino acids (e.g., L-glutamine), antibiotics (e.g., penicillin and streptomycin), and / or antifungals (e.g., amphotericin) or any other supplements conventionally used to support cell culture.

[0337] Purification of fusion protein

[0338] Typically, the fusion protein is purified after expression. Proteins can be isolated or purified in a variety of ways known to those skilled in the art (see, e.g., Protein Purification: Principles and Practice, 3rd ed., Scopes, Springer-Verlag, NY, 1994). Standard purification methods include chromatographic techniques using systems such as FPLC and HPLC at atmospheric or elevated pressure, including ion exchange, hydrophobic interaction, affinity, molecular exclusion or gel filtration, and reverse-phase chromatography. Additional purification methods include electrophoresis, immunoprecipitation, precipitation, dialysis, and chromatofocusing techniques. The combination of ultrafiltration and diafiltration techniques with protein concentration is also useful. As is well known in the art, a variety of native proteins bind to Fc and antibodies, and these proteins are used for the purification of certain antibodies. For example, the bacterial proteins A and G bind to the Fc region. Similarly, the bacterial protein L binds to the Fab region of some antibodies. Purification can also be achieved by specific fusion partners. For example, if GST fusion is employed, glutathione resin can be used to purify the antibody, if His-tag is used, Ni +2 affinity chromatography purification, or if Flab-tag is used, immobilized anti-flag antibody purification. The degree of purification required will vary depending on the use of the antibody. In some cases, purification may not be required.

[0339] In certain embodiments, the fusion protein is substantially pure. The term "substantially pure" (or "substantially purified") when used in reference to the fusion proteins described herein means that the fusion protein is substantially or essentially free of components (such as the starting cells, or in the case of a recombinantly produced fusion protein, the host cells) that are normally associated with or interact with the protein as found in its natural environment. In certain embodiments, a substantially pure fusion protein is a protein preparation having less than about 30%, less than about 25%, less than about 20%, less than about 15%, less than about 10%, or less than about 5% (by dry weight) contaminating protein.

[0340] Assessment of protein purification and / or homogeneity can be performed by any method known in the art, including but not limited to, non-reducing / reducing CE-SDS, non-reducing / reducing SDS-PAGE, ultra-performance liquid chromatography-size exclusion chromatography (UPLC-SEC), high-performance liquid chromatography (HPLC), mass spectrometry, multi-angle light scattering (MALS), dynamic light scattering (DLS).

[0341] Post-translational modification

[0342] In certain embodiments, the fusion proteins described herein comprise one or more post-translational modifications. Such post-translational modifications can occur in vivo, or they can be performed in vitro after the fusion protein has been isolated from the host cell.

[0343] Post-translational modifications include various modifications known in the art (see, e.g., Proteins - Structure and Molecular Properties, 2nd ed., T.E. Creighton, W.H. Freeman and Company, New York, 1993; Post-Translational Covalent Modification of Proteins, ed. B.C. Johnson, Academic Press, New York, pp. 1-12, 1983; Seifter et al., 1990, Meth. Enzymol., 182:626-646, and Rattan et al. 1992, Ann. N.Y. Acad. Sci., 663:48-62). In those embodiments where the fusion protein comprises one or more post-translational modifications, the fusion protein can comprise the same type of modification at one or more sites, or it can comprise different modifications at different sites.

[0344] Examples of post-translational modifications include glycosylation, acetylation, phosphorylation, amidation, derivatization by known protecting / blocking groups, formylation, oxidation, reduction, proteolytic cleavage, or specific chemical cleavage by cyanogen bromide, trypsin, chymotrypsin, papain, V8 protease, or NaBH4.

[0345] Other examples of post-translational modifications include, for example, the addition or removal of N-linked or O-linked carbohydrate chains, chemical modification of N-linked or O-linked carbohydrate chains, processing of the N-terminus or C-terminus, attachment of chemical moieties to the amino acid backbone, and addition or deletion of N-terminal methionine residues produced by expression in a prokaryotic host cell. Post-translational modifications can also include modifications with detectable labels such as enzymatic, fluorescent, isotopic, or affinity labels to permit detection and isolation of the protein. Examples of suitable enzyme labels include, but are not limited to, horseradish peroxidase, alkaline phosphatase, β-galactosidase, and acetylcholinesterase. Examples of suitable cofactor complexes include, but are not limited to, streptavidin / biotin and avidin / biotin. Examples of suitable fluorescent materials include, but are not limited to, umbelliferone, fluorescein, fluorescein isothiocyanate, rhodamine, dichlorotriazinylamine fluorescein, dansyl chloride, and phycoerythrin. An example of a luminescent material is luminol, examples of bioluminescent materials include luciferase, luciferin, and aequorin, and examples of suitable radioactive materials include iodine, carbon, sulfur, tritium, indium, technetium, thallium, gallium, palladium, molybdenum, xenon, and fluorine.

[0346] Additional examples of post-translational modifications include acetylation, ADP-ribosylation, amidation, covalent attachment of flavin, covalent attachment of heme moieties, covalent attachment of nucleotides or nucleotide derivatives, covalent attachment of lipids or lipid derivatives, covalent attachment of phosphatidylinositol, cross-linking, cyclization, disulfide bond formation, demethylation, formation of covalent cross-links, formation of cysteine, formation of pyroglutamate, γ-carboxylation, GPI anchor formation, hydroxylation, iodination, methylation, myristoylation, pegylation, prenylation, racemization, selenoylation, sulfation, transfer-RNA-mediated addition of amino acids such as arginylation and ubiquitination.

[0347] Masking and programmed activation of fusion proteins

[0348] According to the present disclosure, the fusion protein is masked and does not engage with one or more of its intended targets. The degree to which the binding of the fusion protein to one or more of its targets is reduced can be measured by standard techniques such as enzyme-linked immunosorbent assay (ELISA), biolayer interferometry (BLI), surface plasmon resonance (SPR), fluorescence-activated cell sorting (FACS), flow cytometry, kinetic exclusion assay (KinExA), Meso Scale Discovery (MSD), microfluidics, or isothermal titration calorimetry (ITC). In certain embodiments, the fusion protein comprises an antigen-binding domain masked by a ligand-receptor pair, and the binding of the antigen-binding domain to its cognate antigen is reduced by at least 3-fold, for example, at least 5-fold, at least 10-fold, at least 20-fold, at least 25-fold, or at least 30-fold, or at least 40-fold, or at least 50-fold, or at least 70-fold, or at least 80-fold, or at least 90-fold, or at least 100-fold, or at least 200-fold, or at least 400-fold, or at least 600-fold, or at least 800-fold, or at least 1000-fold, or at least 2000-fold, or at least 5000-fold, or at least 10,000-fold as compared to the corresponding unmasked antigen-binding domain.

[0349] According to the present disclosure, proteolytic cleavage of at least one peptide linker in the peptide linker between the ligand or receptor of the ligand-receptor pair and the biofunctional protein unmasks (activates) the fusion protein such that it can bind to one or more of its intended targets. The sensitivity of the peptide linker to cleavage can be tested in vitro by standard techniques including those described in the examples herein. The degree to which the binding of the fusion protein to one or more of its targets is restored after proteolytic cleavage can also be tested by standard techniques such as enzyme-linked immunosorbent assay (ELISA), biolayer interferometry (BLI), surface plasmon resonance (SPR), fluorescence-activated cell sorting (FACS), flow cytometry, kinetic exclusion assay (KinExA), Meso Scale Discovery (MSD), microfluidics, or isothermal titration calorimetry (ITC). The restoration of binding of the fusion protein to one or more of its targets can be partial or complete. Partial restoration of binding is defined as measurable binding of a relevant domain of the fusion protein (e.g., ligand, receptor, or antigen-binding domain) to its intended target and can be, for example, between 2-fold and 100-fold lower than the binding of the parental domain. Partial restoration can be about 100-fold, 75-fold, 50-fold, 25-fold, 10-fold, 5-fold, or 2-fold lower than the binding of the parental domain.

[0350] Therapeutic method

[0351] In some aspects, the present disclosure includes methods for treating a disease or disorder, which include administering a fusion protein described herein to a subject in need thereof. In certain embodiments, the subject is a mammal. In certain embodiments, the subject is a human.

[0352] In certain embodiments, the methods disclosed herein are for treating cancer. The cancer can include, but is not limited to, blood neoplasms (including leukemia, myeloma, and lymphoma), carcinomas (including adenocarcinoma and squamous cell carcinoma), melanoma, and sarcoma. Carcinomas and sarcomas are also often referred to as "solid tumors". In certain embodiments, the cancer is a solid tumor. In certain embodiments, the cancer is leukemia. In certain embodiments, the cancer is lymphoma.

[0353] The fusion protein can exert a cytotoxic or cytostatic effect and can result in one or more of the following: a decrease in tumor size, slowing or preventing an increase in tumor size, disappearance or removal of the tumor and an increase in the disease-free survival time between its recurrence, preventing the initial or subsequent occurrence of the tumor (e.g., metastasis), an increase in the time to progression, a reduction in one or more adverse symptoms associated with the tumor, or an increase in the overall survival time of a subject having the tumor.

[0354] In certain embodiments, the methods disclosed herein are for treating immunodeficiency disorders or diseases.

[0355] In certain embodiments, the methods disclosed herein are for treating autoimmune diseases or disorders.

[0356] The methods described herein include administering a fusion protein described herein to a subject in need thereof. The fusion protein can be administered to the subject by a suitable route of administration. As will be understood by those skilled in the art, the route of administration and / or mode will vary depending on the desired result. Generally, immunotherapeutic antibodies are administered by systemic or local administration. Local administration can be administration at the tumor site or into the tumor-draining lymph nodes. Generally, the fusion protein will be administered parenterally, e.g., by intravenous, intramuscular, intradermal, intraperitoneal, subcutaneous, or spinal administration, such as by injection or infusion.

[0357] Treatment is achieved by administering a "therapeutically effective amount" of the fusion protein. A "therapeutically effective amount" means an amount effective to achieve the desired therapeutic result at a necessary dosage and for a necessary period of time. The therapeutically effective amount can vary depending on factors such as the disease state, age, sex, and weight of the subject. A therapeutically effective amount is also an amount where the beneficial effects of treatment outweigh any toxic or detrimental effects of the fusion protein. "A sufficient amount" means an amount sufficient to produce the desired effect, e.g., an amount sufficient to modulate the immune response to a target cell or tissue, for example, by binding of an immunomodulatory ligand-receptor to an immune cell.

[0358] A suitable dose of the fusion protein can be determined by a skilled practitioner. The selected dose level will depend on various pharmacokinetic factors, including the activity of the particular fusion protein employed, the route of administration, the time of administration, the excretion rate of the polypeptide, the duration of treatment, other drugs, compounds, and / or materials used in combination with the fusion protein, such as anti-cancer drugs, and the age, sex, weight, medical condition, general health, and prior medical history of the subject being treated, as well as similar factors well known in the medical arts.

[0359] Methods of modulating immune cells or immune responses

[0360] In certain embodiments, the fusion proteins described herein are administered to a subject in need thereof, such as a subject suffering from cancer, to modulate the immune system of the subject. Thus, in certain embodiments, the fusion proteins described herein down-regulate or up-regulate an immune response.

[0361] According to this embodiment, administering a sufficient amount of the fusion protein to a subject can achieve one or more of the following to activate or up-regulate an immune response: modulating immune checkpoints, modulating T cell receptor signal transduction, modulating T cell activation, modulating pro-inflammatory cytokines, modulating interferon-γ production by T cells, modulating T cell suppression, modulating the survival and / or differentiation of M2-type tumor-associated macrophages (TAMs) or myeloid-derived suppressor cells (MDSCs), and / or modulating cytotoxicity or cytostatic effects on cells.

[0362] In certain embodiments, the present disclosure provides methods of modulating an immune response, which include inhibiting immune checkpoints, stimulating immune checkpoints, immune cell activation, stimulating T cell receptor signal transduction, and stimulating antibody-dependent cell cytotoxicity (ADCC), T cell-dependent cell cytotoxicity (TDCC), cell-dependent cell cytotoxicity (CDC), or antibody-dependent cell phagocytosis (ADCP).

[0363] In certain embodiments, the fusion protein is capable of agonizing a target leukocyte co-stimulatory receptor when activated by a protease. Functional effects of leukocyte co-stimulatory receptor agonism include activation of T effector cells, differentiation and activation of inflammatory myeloid cells, and / or recruitment of B cells and / or NKT cells. Activation of T effector cells can lead to increased production of one or more cytokines by T cells, such cytokines being, for example, interferon gamma (IFN-γ), interleukin 2 (IL-2), interleukin 12 (IL-12), interleukin 17 (IL-17), interleukin 21 (IL-21), granulocyte-macrophage colony-stimulating factor (GM-CSF), tumor necrosis factor-α (TNF-α), macrophage inflammatory protein 1β (MIP-1β), and / or C-X-C motif ligand 13 (CXCL13). Increased production of IL-21 and CXCL13 by T effector cells may, for example, support the differentiation and activation of inflammatory myeloid cells in the TME, recruit anti-tumor lymphoid cells, such as B cells and NKT cells, and / or support the formation of tertiary lymphoid structures.

[0364] In certain embodiments, the fusion protein activates T effector cells. In some embodiments, the fusion protein increases production of GM-CSF, TNF-α, MIP-1β, IL-17, IL-12, IL-21, and / or C-X-C motif ligand 13 (CXCL13) by T effector cells.

[0365] In certain embodiments, the fusion protein reduces CSF1-dependent viability of monocytes and activates T effector cells.

[0366] Certain embodiments of the present disclosure relate to methods of using a fusion protein to regulate leukocyte co-stimulatory receptor agonism in vivo, for example, to treat cancer.

[0367] In certain embodiments, the method involves, for example, inhibition or downregulation of immune cells or immune responses for treating autoimmune diseases or disorders. Thus, in certain embodiments, the fusion protein is administered in an amount sufficient to regulate immune cells. In certain embodiments, downregulation of the immune response is achieved by regulating immune checkpoints, regulating T cell receptor signaling, regulating T cell activation, regulating pro-inflammatory cytokines, regulating interferon-γ production by T cells, regulating T cell inhibition, regulating the survival and / or differentiation of M2-type tumor-associated macrophages (TAMs) or myeloid-derived suppressor cells (MDSCs), and / or regulating cytotoxicity or cytostatic effects on cells.

[0368] Method for modifying ADCC of target cells

[0369] In certain embodiments, the fusion proteins described herein induce antibody-dependent cell-mediated cytotoxicity (ADCC), which in turn results in increased lysis of target cells. In certain embodiments, the fusion protein comprises an Fc region having an increased binding affinity of the Fc to FcγRIIIa (activating receptor), and the increased binding affinity results in increased antibody-dependent cell-mediated cytotoxicity (ADCC) and increased lysis of target cells. In certain embodiments, the Fc region has a modified CH2 domain, and the modified CH2 domain comprises an amino acid modification that results in an increased binding affinity of the Fc to FcγRIIIa (activating receptor), and the increased binding affinity results in increased antibody-dependent cell-mediated cytotoxicity (ADCC).

[0370] In certain embodiments, the fusion proteins described herein mitigate less antibody-dependent cell-mediated cytotoxicity (ADCC). In certain indications, mitigation or elimination of ADCC and complement-mediated cytotoxicity (CDC) is desired. In certain embodiments, it may be useful for the fusion protein to comprise an Fc region having a modified CH2 domain that contains an amino acid modification that results in increased binding to FcγRIIb or an amino acid modification that reduces or eliminates binding of the Fc region to all Fcγ receptors ("knockout" variants). In certain embodiments, the fusion protein comprises a pharmaceutical composition of an Fc region having reduced binding to FcγRIIb (inhibitory receptor)

[0371] The fusion proteins according to the present disclosure can be formulated into pharmaceutical compositions. These compositions may further comprise pharmaceutically acceptable excipients, carriers, buffers, stabilizers or other materials well known to those skilled in the art in addition to one or more of the fusion proteins in the fusion protein. Such materials should be non-toxic and should not interfere with the efficacy of the active ingredient. The exact nature of the carrier or other material may depend on the route of administration, for example, oral, intravenous, cutaneous or subcutaneous, nasal, intramuscular, intraperitoneal routes.

[0372] Pharmaceutical compositions for oral administration may be in the form of tablets, capsules, powders or liquids. Tablets may contain solid carriers such as gelatin or adjuvants. Liquid pharmaceutical compositions generally contain liquid carriers such as water, petroleum, animal or vegetable oils, mineral oils or synthetic oils. Physiological saline solutions, dextrose or other sugar solutions or glycols such as ethylene glycol, propylene glycol or polyethylene glycol may be included.

[0373] For intravenous, cutaneous, or subcutaneous injection, or injection at the site of pain, the active ingredient will be in the form of a parenterally acceptable aqueous solution that is pyrogen-free and has a suitable pH, isotonicity, and stability. A person of ordinary skill in the art can readily prepare a suitable solution using, for example, isotonic vehicles such as sodium chloride injection, Ringer's injection, and lactated Ringer's injection. Preservatives, stabilizers, buffers, antioxidants, and / or other additives can be included as needed.

[0374] For the fusion proteins according to the present disclosure to be administered to an individual, administration is preferably carried out in a "therapeutically effective amount" sufficient to show a benefit to the individual. A "prophylactically effective amount" can also be administered when sufficient to show a benefit to the individual. The actual amount administered, as well as the rate and time course of administration, will depend on the nature and severity of the protein aggregation disease being treated. The treatment regimen (e.g., determination of dosage, etc.) is the responsibility of general practitioners and other physicians and will generally take into account the disorder to be treated, the condition of the individual patient, the site of delivery, the method of administration, and other factors known to the practitioner. Examples of the techniques and protocols mentioned above can be found in Remington's Pharmaceutical Sciences, 16th Edition, Osol, A. (ed.), 1980.

[0375] The composition can be administered alone or in combination with other therapies, either simultaneously or sequentially, depending on the disorder to be treated.

[0376] Kit

[0377] The present disclosure also provides a kit comprising one or more of the compositions described herein and instructions for use. Thus, in certain embodiments, a kit is described herein that includes a vector for expressing the fusion protein described herein and instructions for use. In certain embodiments, a kit is described herein that includes a host cell comprising a vector for expressing the fusion protein and instructions for use. In certain embodiments, it is a kit comprising a purified fusion protein and instructions for use. The purified fusion protein can be provided lyophilized or in a dry form, such as a powder or granule, and the kit can additionally contain a suitable solvent for reconstituting one or more of the lyophilized or dry components.

[0378] The medicament kit will generally include a container and a label / or packaging insert on or associated with the container. The label or packaging insert contains instructions typically included in the commercial packaging of a therapeutic product, which provide information or instructions regarding indications, usage, dosage, administration, contraindications, and / or warnings for the use of such therapeutic product. The label or packaging insert may also include a notice in a form required by a government agency regulating the manufacture, use, or sale of a drug or biologic, which reflects the agency's approval for manufacture, use, or sale for administration to humans or animals. The container contains a composition comprising a fusion protein. In some embodiments, the container may have a sterile access port. For example, the container can be an intravenous solution bag or a vial having a stopper that can be pierced by a hypodermic needle.

[0379] In addition to the container containing the composition comprising a fusion protein, the medicament kit may further include one or more additional containers containing other components of the medicament kit. For example, a pharmaceutically acceptable buffer (such as bacteriostatic water for injection (BWFI), phosphate buffered saline, Ringer's solution, or dextrose solution), other buffers or diluents.

[0380] Suitable containers include, for example, bottles, vials, syringes, and intravenous solution bags, etc. The container can be formed of various materials such as glass or plastic. As appropriate, one or more components of the medicament kit may be lyophilized or provided in a dry form (such as a powder or granule), and the medicament kit may additionally contain a suitable solvent for reconstituting one or more lyophilized or dry components.

[0381] The medicament kit may also include other materials that are desirable from a commercial and user perspective, such as filters, needles, and syringes.

[0382] Examples

[0383] The following examples are provided for illustrative purposes only and are not intended to limit the scope of the present disclosure in any way. Efforts have been made to ensure the accuracy of the numerical values (e.g., amounts, temperatures, etc.) used, but of course some experimental error and deviation should be allowed.

[0384] Practice of the present disclosure will employ, unless otherwise indicated, conventional methods of protein chemistry, biochemistry, recombinant DNA techniques and pharmacology within the skill of the art. Such techniques are sufficiently explained in the literature. See, for example, T.E. Creighton, Proteins: Structures and Molecular Properties (W.H. Freeman and Company, 1993); A.L. Lehninger, Biochemistry (Worth Publishers, Inc., current edition); Sambrook, et al., Molecular Cloning: A Laboratory Manual (2nd Edition, 1989); Methods In Enzymology (S. Colowick and N. Kaplan eds., Academic Press, Inc.); Remindton’s Pharmaceutical Sciences, 18th Edition (Easton, Pennsylvania: Mack Publishing Company, 1990); Carey and Sundberg Advanced Organic Chemistry 3rd Edition (Plenum Press) Volumes A and B (1992).

[0385] Example 1 Design of a Masked Anti-CD3 x Anti-Her2 T Cell Engager Fusion Protein

[0386] An anti-CD3 Fab x anti-Her2 scFv Fc was attached to a masking moiety on the anti-CD3 Fab by linking one of the ligand-receptor pair PD-1-PDL-1 to the N-terminus of the Fab light chain and the other to the N-terminus of the heavy chain. The fusion protein construct was designed as follows.

[0387] Methods

[0388] The fusion protein is in a modified bispecific Fab x scFv Fc format, which has a half - antibody comprising an anti - CD3 heavy chain and light chain, and the anti - CD3 heavy chain and light chain form a heterodimer with an anti - Her2 scFv fused to Fc. The anti - CD3 complementarity determining regions are described in US20150232557A1 (VL SEQ ID NO: 1, VH SEQ ID NO: 2). The anti - Her2 complementarity determining regions are in scFv format, which is based on the VL and VH of trastuzumab (Carter, P. et al. Humanization of an anti - p185HER2 antibody for human cancer therapy. Proc Natl Acad Sci U S A 89, 4285 - 4289, doi: 10.1073 / pnas.89.10.4285 (1992)), and the VL and VH of trastuzumab are linked by a glycine - serine linker (SEQ ID NO: 3) as described in US10000576B1. To allow for selective heterodimer pairing, mutations are introduced in the anti - CD3 CH3 and the anti - Her2 scFv - Fc cH3 chains as previously described (Von Kreudenstein, T.S. et al. Improving biophysical properties of a bispecific antibody scaffold to aid developability: quality by molecular design. MAbs 5, 646 - 654, doi: 10.4161 / mabs.25632 (2013); (A - chain CH3 domain SEQ ID NO: 4, B - chain CH3 domain SEQ ID NO: 5)). Mutations are also introduced in the two CH2 domains (L234A_L235A_D265S, as compared to wild - type human IgG1 CH2) to reduce binding to Fcγ receptors (SEQ ID NO: 6). In addition, a linker composed of a variable number of repeats of sequences predicted to form a helix - turn is used ((EAAAK) n, Chen, X., Zaro, J.L. & Shen, W.C. Fusion protein linkers: property, design and functionality. Adv Drug Deliv Rev 65, 1357 - 1369, doi:10.1016 / j.addr.2012.09.039(2013)) Polypeptides of modified protein sequences based on the IgV domains of human PD-1 (SEQ ID NO: 7) and / or PD-L1 (SEQ ID NO: 8) (West, S.M. & Deng, X.A. Considering B7-CD28 as a family through sequence and structure. Exp Biol Med (Maywood), 1535370219855970, doi:10.1177 / 1535370219855970(2019)) were respectively fused to the N-terminus of the heavy chain (VH-CH1-hinge-CH2-CH3) and κ light chain (VL-CL) of the anti-CD3 variable domain. These PD-1 and PD-L1 moieties were expected to dimerize and sterically block epitope binding. In all variants, the PD-1 or PD-L1 sequence used as one half of the masker contained mutations to increase the affinity of the PD-1:PD-L1 complex as described above (Maute, R.L. et al. Engineering high-affinity PD-1 variants for optimized immunotherapy and immuno-PET imaging. Proc Natl Acad Sci U S A 112, E6506 - 6514, doi:10.1073 / pnas.1519623112(2015); SEQ ID NO: 9; Liang, z. et al. High-affinity human PD-L1 variants attenuate the suppression of T cell activation. Oncotarget 8, 88360 - 88375, doi:10.18632 / oncotarget.21729(2017); SEQ ID NO: 10). Additionally, in all WT PD-1 moieties, unpaired cysteines were mutated to serine to eliminate the liability of the exposed reducing groups (SEQ ID NO: 11).Some variants also contain a cleavage sequence for the tumor microenvironment (TME)-associated protease uPa (MSGRSANA SEQ ID NO: 28) to allow removal of some or all of the mask by exposing the fusion protein to the protease. A schematic of the masked Fab construct design and the proposed mechanism of action is shown in Figure 1. The final design is a bispecific Fab x scFv Fc molecule that contains a masked anti-CD3 Fab and an anti-Her2 scFv. The schematic is shown in. Figure 2 and the sequences used are listed in Table A.

[0389] Table A: Sequence composition of the variants tested *

[0390]

[0391]

[0392]

[0393] * The PD-1 IgV domain attached to the heavy chain is represented by a striped pattern in the cartoon, and the PD-L1 IgV domain attached to the light chain is shown as a checkerboard pattern.

[0394] Example 2 Generation of Masked Anti-CD3 Variants

[0395] The sequence of the modified CD3 x Her2 Fab x scFv variant designed in Example 1 was introduced into an expression vector, and the sequence was expressed and purified as follows.

[0396] Method

[0397] The heavy chain vector insert containing the heavy chain clone with a signal peptide (Barash et al., 2002, Biochem and Biophys Res.Comm., 294: 835-842, SEQ ID 27) and terminating at G446 (EU number) of CH3 was ligated into the pTT5 vector to generate a heavy chain expression vector. The light chain vector insert containing the same signal peptide and the light chain clone was ligated into the pTT5 vector to generate a light chain expression vector. The resulting light and heavy chain expression vectors were sequenced to confirm the correct reading frame and sequence of the coding DNA.

[0398] The heavy and light chains of the modified CD3 x Her2 Fab x scFv Fc variant were co-expressed in 25 mL of Expi293F TM cells (Thermo Fisher, Waltham, MA) cultures. The Expi293 TM cells were grown at 37 °C in Expi293TM Cultured in expression medium (Thermo Fisher, Waltham, MA) on an orbital shaker rotating at 125 rpm in a humidified atmosphere of 8% CO2. The total cell count of 25 mL volume of 7.5×10 7 cells was transfected with a total of 25 μg DNA at a transfection ratio of H1:L1:H2 of 40:40:20. Before transfection, the DNA was diluted in 1.5 mL Opti-MEM TM I Reduced-Serum Medium (Thermo Fisher, Waltham, MA). 80 μL of ExpiFectamine TM 293 reagent (Thermo Fisher, Waltham, MA) was diluted in 1.42 mL volume of Opti-MEM TM I Reduced-Serum Medium and, after incubation for 5 minutes, was mixed with the DNA transfection mixture to a total volume of 3 mL. After 10 to 20 minutes, the DNA-ExpiFectamine TM 293 reagent mixture was added to the cell culture. After incubation at 37 °C for 18 - 22 hours, 150 μL of ExpiFectamine TM 293 Enhancer 1 and 1.5 mL of ExpiFectamine TM 293 Enhancer 2 (Thermo Fisher, Waltham, MA) were added to each culture. The cells were incubated for 5 to 7 days and the supernatant was harvested for protein purification.

[0399] The clarified supernatant samples were applied to 1 mL of a slurry containing 50% mAb Select SuRe resin (GE Healthcare, Chicago, IL) in batch mode. The column was balanced in PBS. After loading, the column was washed with PBS and the protein was eluted with 100 mM sodium citrate buffer pH 3.5. The eluted sample was pH adjusted by adding 10% (v / v) 1M Tris pH 9 to produce a final pH of 6-7. After concentration, all materials were injected into an AKTA Pure FPLC system (GE LifeSciences) and run on a Superdex 200 Increase 10 / 300GL (GE LifeSciences) column pre-equilibrated with PBS pH 7.4. The protein was eluted from the column at a rate of 0.75 mL / min and collected in 0.5 mL fractions. Peak fractions were pooled and concentrated using a Vivaspin 20, 30 kDa MWCO polyethersulfone concentrator (MilliporeSigma Burlington MA, USA). TM Membrane 0.2μm PALL Acrodisc TM After sterile filtration through a syringe filter, the protein was quantified based on A280nm (Nanodrop) and frozen and stored at -80°C until further use.

[0400] result

[0401] After protein A purification, the sample contained a large amount of higher molecular weight species as determined by UPLC-SEC (not shown), and preparative SEC was used to obtain a high purity sample. The yield after preparative SEC ranged from 1.5 mg to 5 mg per variant. Sample purity and stability were evaluated in Examples 3 and 4.

[0402] Example 3 Purity and Homogeneity Assessment of Masked Anti-CD3 Variants

[0403] Purity and sample homogeneity of the purified variants were assessed by non-reducing / reducing CE-SDS UPLC-SEC as described below.

[0404] method

[0405] After purification, CE-SDS Sample purity was assessed by non-reduced and reduced high-throughput protein expression assays using the GXII (Perkin Elmer, Waltham, MA). The User Guide, 2nd Edition, implementation procedure, with the following modifications. Add 2 μL or 5 μL (concentration range 5 - 2000 ng / μL) of the mAb sample and 7 μL of HT Protein Express sample buffer (PerkinElmer #760328) to individual wells in a 96-well plate (BioRad, Hercules, CA). Prepare the reducing buffer by adding 3.5 μL of DTT (1 M) to 100 μL of HT Protein Express sample buffer. Then denature the mAb sample at 90 °C for 5 minutes and add 35 μL of water to each sample well. Run the instrument using the HT Protein Express Chip (Perkin Elmer #760499) and the HT Protein Express 200 detection settings (14 kDa - 200 kDa).

[0406] Perform UPLC-SEC on an Agilent Technologies 1260 Infinity LC system using an Agilent Technologies AdvanceBio SEC 300A column at 25 °C. Before injection, centrifuge the sample at 10000 g for 5 minutes and then inject 5 μL into the column. Run the sample at a flow rate of 1 mL / min in PBS, pH 7.4 for 7 minutes and monitor the elution by UV absorbance at 190 - 400 nm. Extract the chromatogram at 280 nm. Perform peak integration using OpenLAB CDS ChemStation software.

[0407] Results

[0408] Figure 3A , 3C , representative UPLC-SEC traces of the samples after preparative SEC purification of the variants in 3E and 3G show that the samples are highly homogeneous, containing 89% - 94% of the correct species. The presence of small peaks at low retention times compared to the major species indicates the presence of minor amounts of high molecular weight species, such as oligomers and aggregates, in all samples.

[0409] Analysis of non-reducing CE-SDS ( Figure 3B , 3D, 3E, and 3F) showed a single dominant species, and only bands corresponding to the full-length chains of all variants were found in the reducing CE-SDS runs. Notably, the masked heavy and light chains showed significantly higher apparent molecular weights than expected (110 kDa vs. 63 kDa for HC and 54 kDa vs. 37 kDa for LC). This was also reflected in the high apparent molecular weights of the non-reduced disulfide-bonded species (215 kDa and 152 kDa). Glycosylation of the PD1 and PD-L1 moieties in the design may contribute to the increased apparent molecular weight (Tan, S. et al. An unexpected N-terminal loop in PD-1 dominates binding by nivolumab. Nat Commun 8, 14369, doi:10.1038 / ncomms14369 (2017), Li, C.W. et al. Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity. Nat Commun 7, 12632, doi:10.1038 / ncomms12632 (2016)).

[0410] Example 4 Stability Assessment of Masked Anti-CD3 Variants

[0411] The thermal stability of the purified variants was evaluated by differential scanning calorimetry (DSC) as described below.

[0412] Method

[0413] Samples of a representative set of modified CD3 x Her2 Fab x scFv Fc variants were diluted to 0.5 - 1 mg / ml in PBS. For DSC analysis using NanoDSC (TA Instruments, New Castle, DE, USA), 950 μl of sample and matching buffer (PBS) were added to the sample and reference 96-well plates, respectively. At the start of the DSC run, a buffer (PBS) blank injection was performed to stabilize the baseline. Then, each sample was injected and scanned from 25 °C to 95 °C at 1 °C / min using a nitrogen pressure of 60 psi. The thermograms were analyzed using NanoAnalyze software. The matching buffer thermogram was subtracted from the sample thermogram, and a baseline fit was performed using a sigmoidal curve. The data were then fit with a two-state scaled DSC model.

[0414] Results

[0415] DSC thermal analysis plot of the unmodified CD3 x Her2 Fab x scFv Fc variant (30421, Figure 4 ) shows transitions at 68 °C and 83 °C. T m The transition at 68 °C may correspond to an unresolved single transition of the unfolding of the anti-CD3 Fab, anti-Her2 scFv, and CH2 domain, while the T m transition at = 83 °C may correspond to the unfolding of the CH3 domain in the heavy chain. The thermal analysis plot of the variant carrying the PD-1:PD-L1 mask (30430, 30436; Figure 4 ) also shows two transitions at similar temperatures and has a thermal analysis trace similar to that of the de-masked variant. This indicates that the fused masking domain does not affect the T m of the anti-CD3 Fab and unfolds either cooperatively or non-cooperatively with the Fab but has a T m similar to that of the Fab, scFv, and CH2.

[0416] Example 5 uPa cleavage of anti-CD3 variants

[0417] To evaluate the release of some or all of the mask from the anti-CD3 Fab of the fusion protein due to the protease cleavage site introduced in the cleavage linker, samples were treated in vitro with uPa. The reaction was monitored by reducing Caliper as follows.

[0418] Method

[0419] For preparative cleavage of the variant, 25 - 100 μg of the purified sample was diluted in PBS + 0.05% Tween 20 to a final variant concentration of 0.2 mg / mL, and recombinant human u-plasminogen activator (uPa) / urokinase (R&D Systems #P00749) was added at a protease:substrate molar ratio of 1:50. After incubation at 37 °C for 24 hours, the sample fragments were analyzed in reducing CE-SDS as described in Example 2 and then frozen at -80 °C and stored until further use.

[0420] Results

[0421] Analysis of the reducing CE-SDS profiles of the masked variants with and without uPa treatment revealed that under the conditions studied, cleavage at the introduced cleavage site effectively removed some or all of the mask from the Fab ( Figure 5)。For successfully cleaved variants (30430, 30436, 31934), the bands representing the masked heavy and / or light chain fragments completely disappeared after cleavage, and the unmasked heavy and / or light chain fragments appeared. Although a low-intensity broad band corresponding to the free PD-1 fragment could be observed for variant 30430, this was not the case for the released PD-L1 in variant 30436. The size heterogeneity caused by small size and glycosylation (Tan, S. et al. An unexpected N-terminal loop in PD-1 dominates binding by nivolumab. Nat Commun 8, 14369, doi:10.1038 / ncomms14369 (2017), Li, C.W. et al. Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity. Nat Commun 7, 12632, doi:10.1038 / ncomms12632 (2016)) may render the free PD-1 and PD-L1 fragments almost undetectable and undetectable, respectively. No cleavage was observed in the variants without the cleavage sequence (30421, 30423).

[0422] Example 6 Masking / Unmasking of CD3 Binding

[0423] The binding of the uncleaved and cleaved samples of the anti-CD3 variants from Example 5 to Jurkat cells expressing CD3 was tested by ELISA as follows and the binding of the samples to pan T cells was tested by flow cytometry.

[0424] Method

[0425] ELISA

[0426] Human Jurkat cells (Fujisaki Cell Center, Japan) were maintained in RPMI-1640 medium supplemented with 2 mM L-glutamine and 10% heat-inactivated fetal bovine serum (FBS) (containing 1X penicillin / streptomycin) in a humidified +5% CO2 incubator at 37°C.

[0427] Samples of the modified CD3 x Her2 variants from Example 5 were diluted 2-fold in a blocking buffer containing a saturating amount of irrelevant human Ig and then serially diluted three-fold seven times in the blocking buffer to obtain a total of eight concentration points. Separate blocking buffer was added to the control wells to measure the background signal on the cells (negative / blank control).

[0428] All incubations were carried out at 4 °C. On the day of the assay, exponentially growing cells were centrifuged and seeded in a 96-well filter plate (MilliporeSigma, Burlington, MA, USA) in a 1:1 mixture of complete medium and blocking buffer. Equal volumes of 2X variant or control were added to the cells and incubated for 1 h. The plates were then washed 4 times using vacuum filtration. An HRP-conjugated anti-human IgG Fcγ-specific secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA) was added to the wells and incubated for a further 1 h. The plates were washed 7 times by vacuum filtration and then the TMB substrate (Thermo Scientific, Waltham MA, USA) was added at room temperature. The reaction was terminated by adding 0.5 volume of 1 M sulfuric acid and the supernatant was transferred to a clear 96-well plate (Coming, Coming, NY, USA) by filtration. Absorbance at 450 nm was read on a Spectramax 340PC plate reader with pathcheck correction.

[0429] Binding curves of OD450 minus blank versus linear or log antibody concentration were fitted using GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA). A one-site specific four-parameter non-linear regression curve fitting model with a Hill slope was employed to determine the Bmax and apparent Kd values for each test article.

[0430] Flow cytometry

[0431] Antibodies were titrated from 300 nM to 1.7 pM at a 1:3 dilution in a total volume of 20 μL / well in a V-bottom 96-well plate (Sarstedt AG, Nümbrecht, Germany) in FACS buffer (PBS containing 2% FBS (Thermo Fisher Scientific, Waltham, MA)). Peripheral blood pan T cells from healthy donors (BioIVT, Westbury, NY) were thawed and washed in medium consisting of RPMI 1640 medium (A1049101, ATCC modified) (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA). Cells were counted, resuspended in FACS buffer, and added to the 96-well plate at 50,000 cells per well. Cells were incubated with the variants at 4 °C for 1 hour and then washed twice with FACS buffer and 1 mg / mL secondary antibody AF647 goat anti-human IgG Fc (Jackson ImmunoResearch, West Grove, PA). A 1000-fold diluted viability dye (Biolegend, San Diego, CA) was also added to the wells. The plate was incubated at room temperature for 30 minutes with shaking (200 rpm). Cells were then washed twice in FACS buffer and resuspended in 100 μL of FACS buffer. For assay readings, the geometric mean of APC fluorescence was measured by flow cytometry on a BD LSR Fortessa (BD Biosciences, San Jose, CA). Raw data were analyzed on FlowJo, LLC software (Becton, Dickinson & Company, Ashland, OR). Graphs were generated using GraphPad Prism version 8.1.2 for Mac OS X (GraphPad Software, La Jolla, CA).

[0432] Results

[0433] ELISA

[0434] As in Figure 6As can be seen, variants (30423, 30430, 30436) containing a full-length PD1:PD-L1-based mask attached to CD3 Fab showed 40 - 180-fold reduced binding compared to the unmasked control (30421). After treatment with uPa, the CD3-binding portion of cleavable variants 30430 and 30436 was restored (within 6 - 7-fold of the unmasked control). This partial restoration may be due to steric hindrance of epitope binding by the mask portion remaining on the mask after cleavage. Meanwhile, controls (31929, 31931) having only PD-1 or PD-L1 attached to the heavy or light chain respectively showed a binding reduction (4 - 5-fold) compared to the unmasked control similar to that of the uPa-cleaved samples of the fully masked variants.

[0435] Flow cytometry

[0436] As in Figure 22 As can be seen, variants (30423, 30430) containing a full-length PD1:PD-L1-based mask attached to CD3 Fab showed >43-fold reduced binding compared to the unmasked control (30421). After treatment with uPa, the CD3-binding portion of cleavable variant 30430 was restored (within 29-fold of the unmasked control). This partial restoration may be due to steric hindrance of epitope binding by the mask portion remaining on the mask after cleavage. Meanwhile, the binding reduction (14-fold) of control (31929) having only PD-1 attached to the heavy chain compared to the unmasked control was similar to that of the uPa-cleaved samples of the fully masked variants. In a separate experiment, the binding reduction (6-fold) of variant (32497) having a non-functional PD-1 domain attached to the heavy chain compared to the unmasked control was similar to the binding reduction (31929, 5-fold) observed for an equivalent variant having a functional PD-1 ( Figure 32 ).

[0437] Example 7 T cell-dependent cytotoxicity of masked and unmasked variants

[0438] The functional effect of PD-1:PD-L1-based masks on the ability of CD3 x Her2 Fab x scFv Fc variants to engage and activate T cells to kill Her2-bearing cells was evaluated in a T cell-dependent cytotoxicity (TDCC) assay as follows.

[0439] Method

[0440] Co-culture assay

[0441] JIMT-1 (Leibniz Institute, Braunschweig, Germany), which was cultured in a growth medium consisting of DMEM medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA), HCC1954 (ATCC, Manassas, VA) and HCC827 (ATCC, Manassas, VA), which were cultured in a growth medium consisting of RPMI-1640 ATCC modified (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum, and MCF-7 (ATCC, Manassas, VA), which was cultured in a growth medium consisting of MEM medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum and 0.01 mg / mL human recombinant insulin (Thermo Fisher Scientific, Waltham, MA), were maintained horizontally in T-175 flasks (Corning, Corning, NY) in an incubator containing 5% carbon dioxide at 37 °C. On the day of setting up the assay, the variants were titrated directly in triplicate from 5 nM to 0.08 pM at a 1:3 dilution in a 384-well cell culture-treated optical bottom plate (Thermo Fisher Scientific, Waltham, MA). The tumor cells were rinsed with PBS (Thermo Fisher Scientific, Waltham, MA), harvested with TrypLE Express (Thermo Fisher Scientific, Waltham, MA), diluted in the medium, and counted using Vi-Cell (Beckman Coulter, Indianapolis, IN). A vial of primary human pan T cells (BioIVT, Westbury, NY) was thawed in a 37 °C water bath, washed in the medium, and counted using Vi-Cell. The pan T cell suspension was mixed with the tumor cells at an effector:target ratio of 5:1, washed, and resuspended at 0.55E6 cells / mL. 20 μL of the mixed cell suspension was added to the plate containing the titrated variants. The plate was incubated in an incubator containing 5% carbon dioxide at 37 °C for 48 hours. Then, high-content cytotoxicity assessment of the samples was performed, and the supernatant was collected for IFNγ analysis.

[0442] High-content cytotoxicity analysis

[0443] To visualize the cell nuclei and assess viability, cells were stained with Hoechst 33342. 10 μL of Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA) was diluted 1:1000 in the medium, added to the cells after a 48-hour period, and incubated for an additional 1 hour at 37 °C. Then, the plate was subjected to high-content image analysis on a CellInsight CX-5 (Thermo Fisher Scientific, Waltham, MA) to distinguish and quantify viable tumor cells, dead tumor cells, and effector cells. The plate was scanned on a CellInsight CX5 high-content instrument using SpotAnalysis.V4 Bioapplication with the following settings: objective: 10x, channel 1 - 386 nm: Hoechst (fixed exposure time 0.008 ms, gain of 2).

[0444] IFNγ Quantification

[0445] To perform IFNγ quantification in an MSD U-PLEX 384-well single-point assay, a streptavidin-coated multi-array plate (MA6000384SA plate, Meso Scale Diagnostics, Rockville, MD) was blocked with 50 μL of diluent 100, sealed, and incubated at room temperature with shaking (800 rpm) for 30 minutes. At the end of the incubation, all wells were aspirated. Biotinylated capture IFNγ antibody was added to diluent 100 at a 1:16.5 ratio, and 10 μL of the capture antibody solution was added to each well of the blocked plate. The plate was sealed and incubated overnight at 4 °C. The next day, the frozen supernatants from the co-culture assays were thawed on wet ice. The plate was washed and 5 μL of diluent 43 was added to each well, followed by 5 μL of the thawed supernatant sample or standard. The plate was sealed and incubated at room temperature with shaking (800 rpm) for 1 hour. After incubation, the plate was washed and 10 μL of SULFO-TAG detection antibody diluted 1:1000 in diluent 3 was added to each well. The plate was sealed and incubated at room temperature with shaking (800 rpm) for 1 hour. After incubation, the plate was washed and 40 μL of MSD GOLD read buffer was added to each well. The plate was read on a MESO SECTOR R600 instrument (Meso Scale Diagnostics, Rockville, MD).

[0446] PD-L1 and Her2 Receptor Quantification

[0447] Her2 and PD-L1 receptor quantification was performed via flow cytometry using the Quantum Simply Cellular anti-human and anti-mouse IgG kits (Bangs Laboratories, Fishers, Indiana) respectively. Tumor cells were rinsed with PBS (Thermo Fisher Scientific, Waltham, MA) and harvested with TrypLE Express (Thermo Fisher Scientific, Waltham, MA). Cells were counted using Vi-Cell (Beckman Coulter, Indianapolis, IN), washed and resuspended at 4×10^6 cells / mL in FACS buffer - PBS containing 2% FBS (Thermo Fisher Scientific, Waltham, MA). 25 μL of the tumor cell suspension was added in triplicate to 96-well V-bottom plates (Sarstedt AG, Nümbrecht, Germany). To wells and Eppendorf tubes (Thermo Fisher Scientific, Waltham, MA) containing Quantum Simply Cellular IgG beads (anti-human or anti-mouse) and blank beads, 15 μg / mL of anti-Her2-AF647 (trastuzumab, monovalent antibody, Zymeworks, Vancouver, BC), anti-PDL1-APC (clone MIH1, BD Biosciences, San Jose, CA) or irrelevant negative control IgG-AF657 (Zymeworks, Vancouver, BC) antibody was added. Cells and beads were incubated with the antibody in the dark at 4°C for 1 hour. Cells and beads were washed, resuspended and analyzed by flow cytometry. For analysis, a standard curve was generated using a spreadsheet provided by Bangs Laboratories (Fishers, Indiana) for a specific batch of beads, and the surface antigen binding capacity (ABC) was generated using the same spreadsheet by inputting the geometric mean of the cell population. The ABC value represents the number of receptor molecules expressed on the cell surface assuming a monovalent binding model. The standard curve for the reliable determination range of the receptor number was from 3500 receptors / cell to 330000 receptors / cell for Her2 and from 4400 receptors / cell to 630000 receptors / cell for PD-L1.

[0448] Results

[0449] When probing the function of the same sample in a TDCC assay using JIMT-1 cells expressing Her2, the masking effect in CD3 binding observed in Example 6 was reproduced ( Figure 7 . The de-masked variant (30421) showed robust tumor cell killing at low variant concentrations, while the potency of the masked, non-cleavable variant (30423) was reduced by approximately 49,000-fold. The fully masked variant (30430) with a cleavable PD-L1 moiety on the light chain also had a potency reduction of approximately 5,800-fold without uPa treatment. This masking difference between the non-cleavable and cleavable variants was also observed in CD3 binding (Example 6). When the mask was cleaved with uPa, the potency of 30430 was restored to that of the de-masked (30421) variant. A control variant (31929) with only the PD-1 moiety of the mask attached showed a potency similar to 30421 and 30430 treated with uPa. An irrelevant anti-respiratory syncytial virus (RSV) antibody (22277) showed no activation of T cells for tumor cell killing.

[0450] The TDCC using JIMT-1 as a Her2 and PD-L1 positive cell line was repeated and extended to 3 other cell lines using different levels of those receptors and different T cell donors than in the previous experiment. The cytotoxicity data for two repeats are shown in Figure 23. For repeat n = 1, the level of the cytokine IFNγ was also monitored as an indicator of T cell immune activation ( Figure 24 ). The number of receptors for all cell lines used was determined and is shown in Figure 25Among the cytotoxicities for different cell lines, the potency of the de-masking control (30421) was determined to be between 0.03 pM (HCC1954: high Her2, high PD-L1) and 3 pM (MCF-7: medium Her2, low PD-L1). The potency of this de-masking control as determined by IFNγ release was between 8.4 pM (HCC1954: high Her2, high PD-L1) and 50 pM (HCC829: low Her2, medium PD-L1). Masking as measured by the increase in EC50 of the non-cleavable (30423) and cleavable (30430) masked variants was confirmed in all cell lines and was in the range of 72 to >450-fold for the cytotoxicity readout and 8.2 to >350-fold for the IFNγ readout. Variants having only the PD-1 portion attached to the heavy chain (31929), the cleavable masked variant after uPa treatment (30430 + uPa), and the combination of the de-masking control and saturating amounts of anti-PD-L1 antibody (30421 + 120 nM atezolizumab) showed higher potency (cytotoxicity EC50 reduced 0.019 to 0.84-fold) than the de-masking control (30421) in cell lines with significant PD-L1 expression (HCC1954, JIMT-1, HCC827) due to their ability to engage PD-L1. Cell lines with very low PD-L1 expression (MCF-7) did not show significant differentiation in cytotoxicity readouts between the de-masking control (30421) and those variants capable of engaging PD-L1 (31929, 30421 + 120 nM atezolizumab, 30430 + uPa). However, for all cell lines tested, these variants with anti-PD-L1 portions did show higher potency in terms of IFNγ release compared to the de-masking control (30421). For any of these cell lines, an irrelevant anti-RSV antibody (22277) did not show activity in TDCC.

[0451] Example 8 PD1 and PD-L1 Binding Analysis of Masked Anti-CD3 Variants

[0452] As an indicator of the biological activity of the PD-1 and PD-L1 portions used as masking domains, the binding of the modified variants to CHO cells expressing PD-L1 and PD-1 was determined as follows.

[0453] Method

[0454] CHO Cell Transfection

[0455] CHO-S cells (National Research Council of Canada) were cultured in FreeStyle CHO Expression Medium (Thermo Fisher Scientific, Waltham, MA) containing 1% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA). Transfection was performed using the Neon Transfection System (Thermo Fisher Scientific, Waltham, MA). CHO-S cells were counted and the cells were washed 2 times with PBS and once in resuspension buffer R (Thermo Fisher Scientific, Waltham, MA) and then resuspended at 100E6 cells / mL. PD-1, PDL-1 or GFP plasmid DNA (GenScript, Piscataway, NJ) was added at 1 μg / 1E6 cells. The Neon tip was loaded with 3 mL of electroporation buffer E2 (Thermo Fisher Scientific, Waltham, MA). Using a 100 μL Neon tip (Thermo Fisher Scientific, Waltham, MA), each plasmid was transfected under the following settings: voltage - 1620, width - 10, pulse - 3. The transfected cells were transferred to pre-warmed flasks at a concentration of 1E6 cells / mL for each condition.

[0456] PD1 / PDL1 Binding according to Flow Cytometry

[0457] The variant purified in Example 2 and subjected to uPa treatment in Example 5 was directly titrated from 200 nM at a 1:3 dilution in a V-bottom 96-well plate (VWR, Radnor, PA, USA). CHO-PD1, CHO-PDL-1, and CHO-GFP cells were thawed and washed in RPMI 1640 medium (A1049101, ATCC-modified) (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA, USA) and resuspended in FACS buffer (PBS + 2% FBS). Each of the CHO-PD1 and CHO-PDL-1 cells was combined with CHO-GFP cells at a 2:1 ratio, and 20 μL of the cell suspension was added to the plate containing the titrated variant. The cells were incubated with the variant at 4 °C for 1 hour. After incubation, the cells were washed twice with FACS buffer, and 1 μg / mL of secondary antibody AF647 goat anti-human IgG Fc (Jackson ImmunoResearch, West Grove, PA, USA) and a 1000-fold dilution of the viability dye (Biolegend, San Diego, CA, USA) were added to the wells. The plate was incubated at room temperature for 30 minutes. The cells were washed twice in FACS buffer and resuspended in 50 μL of FACS buffer.

[0458] For assay readings, the geometric mean of APC fluorescence was measured by flow cytometry on a BD LSRFortessa (BD Life Sciences, Gurugram, India). Nonspecific binding was determined by measuring the geometric mean of APC fluorescence of GFP-positive cells. Graphs were generated using GraphPad Prism version 8.1.2 for Mac OS X (GraphPad Software, La Jolla, CA, USA).

[0459] Results

[0460] As shown in Figure 8, for the masked variants (30423, 30426, 30430, 30436) without uPa treatment (-uPa), no binding to PD-L1 (A) or PD-1 (B) was observed. Variants with only affinity-matured PD-1 or PD-L1 moieties attached to the heavy or light chain showed binding with IC50 values of 0.3 nM and 6 nM, respectively. The uncleavable variants (30423, 30426) did not bind to PD-L1 or PD-1 when treated with protease (+uPa), while the uPa-treated samples containing a uPa cleavage sequence between the Fab and the PD-1:PD-L1 mask restored partial binding. Specifically, the binding of 30430 to PD-L1 was partially restored, within 53-fold of the relevant unidirectional masking control 31929 (A). The binding of 30436 to PD-1 was partially restored, within 12-fold of the unidirectional masking control 31931 (B). This is consistent with the properties of the immunomodulators (PD-1 on 30430, PD-L1 on 30436) designed to remain on these variants when cleaved by protease. As expected, variants without a PD-1:PD-L1-based mask (30421) and unrelated controls (22277) did not show binding to PD-L1 or PD-1. In a separate experiment using JIMT-1 as the target cell, the variant (32497) with a non-functional PD-1 domain attached to the heavy chain showed a reduced TDCC potency (55-fold EC 50 , Figure 33 ) compared to the unmasked control (v30421), while the equivalent variant (31929) with a functional PD-1 showed an elevated TDCC potency (0.2-fold EC 50 ) compared to the previously observed unmasked control (v30421).

[0461] Study of the functionality of the addition of a PD-1 mask in the hybrid PD-I / PD-L1 reporter assay of Example 9

[0462] To investigate the blockade of the PD-1:PD-L1 checkpoint engagement by the PD-1 moiety of the mask plus the blockade of the T cell engagement function of the variant, a custom hybrid PD-1 / PD-L1 reporter assay (RGA) was performed as follows.

[0463] Methods

[0464] Before the assay setup, JIMT-1 (Leibniz Institute, Braunschweig, Germany) cultured in growth medium consisting of DMEM medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA), HCC1954 (ATCC, Manassas, VA) and HCC827 (ATCC, Manassas, VA) cultured in growth medium consisting of RPMI-1640 ATCC modified (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum, MCF-7 (ATCC, Manassas, VA) cultured in growth medium consisting of MEM medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum and 0.01 mg / mL human recombinant insulin (Thermo Fisher Scientific, Waltham, MA), and Jurkat T cells stably expressing human PD-1 and NFAT-induced luciferase cultured in RPMI-1640 medium ATCC modified supplemented with 10% fetal bovine serum (PD-1 / PD-L1 Blockade Bioassay Promega catalog number J1250, Madison, WI) were maintained in T-75 or T-175 flasks (Corning, Coming, NY) in an incubator containing 5% carbon dioxide at 37 °C. On the day of the experiment, the variants were directly titrated in triplicate at a 1:3 dilution from 150 nM to 0.85 pM in a total volume of 20 μL per well into a 384-well low-profile white flat-bottom polystyrene TC-treated microplate (Corning Cat#3570, Corning, NY). The tumor cells were dissociated using cell dissociation buffer and the cells were mixed with Jurkat cells at a 1:1 ratio in RPMI 1640 supplemented with 1% fetal bovine serum. 20 μL of the mixed cell suspension was added to the plate containing the titrated variants. The plate was incubated at 37 °C in 5% carbon dioxide for 16 hours. After incubation, 40 μL of Bio-Glo™ luciferase assay reagent (Promega catalog number G7940, Madison, WI) was added to all wells, ensuring no bubbles were formed, and the plate was read in luminescence mode on a microplate reader (Biotek Synergy H1, Winooski, VT) at a gain of 150 after 10 minutes. A schematic diagram of the assay setup is shown in Figure 9A in.

[0465] Results

[0466] The customized RGA analysis for detecting the added functionality of the mask is shown in Figure 9B . High RGA responses were observed when cells were treated with a combination of a de-masking variant (30421) capable of crosslinking T cells and tumor cells and a saturating amount (150 nM) of anti-PD-L1 antibody. Although the de-masked bispecific CD3 x Her2 antibody can effectively crosslink T cells and tumor cells, high concentrations of anti-PD-L1 antibody strongly block the PD-1:PD-L1 checkpoint engagement, resulting in high signals at all tested variant concentrations. In contrast, when treated with only the de-masking variant (30421), the signal was significantly reduced due to the engagement of PD-1 and PD-L1 between the modified T cells and JIMT-1 cells. When compared with the de-masked 30421, the uncleavable (30423) and cleavable (30430) masked variants without uPa treatment (-uPa) showed significantly reduced activity at variant concentrations below 10 nM, indicating that steric blockade of the CD3 epitope can effectively inhibit T cell adapter functionality. The 30430 sample without uPa treatment was more potent than 30423 in eliciting an RGA response. When treated with uPa (+uPa), at variant concentrations above 100 pM, the cleavable masked variant (30430) showed higher activity in the RGA than the de-masked control (30421), indicating the de-masking of the CD3 epitope and the blockade of the PD-1:PD-L1 checkpoint engagement by the functional PD-1 portion of the mask remaining on the variant after cleavage. Based on this finding, a control with only the PD-1 domain attached to the heavy chain of the CD3 Fab showed a similar profile and increased activity in the RGA compared to the de-masked control (30421) at variant concentrations above 100 pM. The irrelevant anti-RSV antibody (22277) showed no activity in the RGA.

[0467] Repeated use of JIMT-1 as an RGA for Her2- and PD-L1-positive cell lines and extension of this RGA to three other cell lines at different levels of those receptors, as determined in Example 7. Data for the RGA conducted here are shown in Figure 26. Masking was confirmed in all cell lines as measured by the increase in EC50 of the uncleavable (30423) and cleavable (30430) masked variants and was in the range of 4- to 530-fold compared to the unmasked control (30421). The potency of this unmasked control (30421) was comparable between cell lines (EC50 = 20-50 pM), while variants tested on cell lines with lower Her2 and / or PD-L1 receptor numbers (HCC827 and MCF-7) showed stronger masking than variants with higher receptor expression (HCC1954 and JIMT-1). After treatment with uPa (+uPa), the potency of the cleavable masked variant (30430) recovered to within 1.7- to 3.6-fold of the unmasked control. Variants having only the PD-1 portion attached to the heavy chain (31929), the cleavable masked variant after uPa treatment (30430+uPa), and the combination of the unmasked control and saturating amounts of anti-PD-L1 antibody (30421+120 nM atezolizumab) showed higher efficacy (1.6- to 3.3-fold above maximum RLU) in cell lines with significant PD-L1 expression (HCC1954, JIMT-1, HCC827) due to their ability to engage PD-L1. Higher potency (EC50 of 0.2- to 0.4-fold) was also observed for these variants in cell lines with high TAA and PD-L1 expression (HCC1954, JIMT-1). Cell lines with very low PD-L1 expression (MCF-7) did not show a differentiation between the unmasked control (30421) and those variants capable of engaging PD-L1 (31929, 30421+120 nM atezolizumab, 30430+uPa). For any of these cell lines, the irrelevant anti-RSV antibody (22277) did not show activity in terms of RGA.

[0468] Example 10 Preparation of Masked Anti-EGFR, Anti-Mesothelin, Anti-TF, Anti-CD19, Anti-cMet, and Anti-CDH3 Variants

[0469] To investigate the applicability of the masking technique to antibodies targeting different antigens, the variable domains of mAbs targeting several different epitopes were attached to a masking domain containing the PD-1:PD-L1 complex. The fusion protein constructs were designed as follows.

[0470] Method

[0471] As described in Example 1, the protein sequences of the WT and modified IgV domains of human PD-1 and PD-L1 were separately conjugated to the N-termini of the IgG1 heavy chain and κ light chain (VL-CL) of antibodies targeting several different epitopes (EGFR, mesothelin, TF, CD19, cMet, CDH3) via non-cleavable and uPa-cleavable linkers. The sequences of VL and VH and their sources are described in Table 2. The significant difference from the constructs in Example 1 was the use of wild-type (WT) CH3 (SEQ ID 12), allowing the assembly of homodimeric full-size antibodies. A schematic of the masked Fab construct design and the expected mechanism of action (MoA) is shown in Figure 1. A schematic of the final design (a bivalent fully masked mAb with two identical heavy and light chains) is shown in Figure 10 . The sequences used for the final variants are listed in Table B.

[0472] Table B: Sequences of the complementarity-determining regions for compatibility studies with the mask

[0473]

[0474]

[0475] Table C: Sequence composition of the tested variants

[0476]

[0477]

[0478]

[0479]

[0480] Example 11 Production of Masked Anti-EGFR, Anti-Mesothelin, Anti-TF, Anti-CD19, Anti-cMet, and Anti-CDH3 Variants

[0481] The sequences of the modified variants designed in Example 10 were cloned into an expression vector, and the sequences were expressed and purified as follows.

[0482] Methods

[0483] The heavy and light chain sequences of the modified variants targeting several different epitopes (EGFR, MSLN, TF, CD19, cMet, CDH3) from Example 10 were transfected into Expi293F TM cells at an equimolar ratio, and expression and purification were performed as described in Example 2.

[0484] Results

[0485] Use preparative SEC as described in Example 2 to obtain high-purity samples. The yield after preparative SEC ranges from 1.5 mg to 6 mg per variant. Evaluate the sample purity as described in Example 12.

[0486] Example 12 Quality assessment of masked anti-EGFR, anti-mesothelin, anti-TF, anti-CD19, anti-cMet, and anti-CDH3 variants

[0487] Evaluate the purity and sample homogeneity of the purified variants from Example 10 by UPLC-SEC and non-reducing / reducing SDS-PAGE.

[0488] Methods

[0489] For non-reducing SDS-PAGE, dilute 2 μL of the sample with 10 μL of PBS and then mix it with 4 μL of 4X Laemmli buffer (BioRad, Hercules, CA). Then heat the sample at 95 °C for 5 minutes and run it on a Mini-PROTEAN 4-20% precast gel (BioRad, Hercules, CA) in the provided Tris / glycine / SDS buffer, and then stain, decolorize, and image with Coomassie G-250. UPLC-SEC and non-reducing and reducing CE-SDS are performed as described in Example 3.

[0490] Results

[0491] Figures 11A to 11J The UPLC-SEC traces of the samples after preparative SEC purification show that the samples are homogeneous and contain 85%-98% of the correct substance. The presence of small peaks at low retention times compared to the main substance indicates the presence of a small amount of high-molecular-weight substances, such as oligomers and aggregates, in all samples. These high-molecular-weight substances are more prevalent in the CD19- and EGFR-targeted samples compared to the samples targeting MSLN, TF, c-Met, and CDH3.

[0492] Analysis of non-reducing SDS-PAGE and CE-SDS ( Figure 11K , Figure 11L)All variants were shown to have a single dominant species. Notably, the apparent molecular weight of this species was significantly higher than expected (>250 kDa vs 200 kDa). Reducing CE-SDS of representative variants targeting c-Met and CDH3 showed only bands corresponding to the intact heavy and light chains. These bands showed the same high apparent molecular weight as described in Example 3. Glycosylation of both the PD1 and PD-L1 moieties in the design may contribute to the increased apparent molecular weight (Tan, S. et al. An unexpected N-terminal loop in PD-1 dominates binding by nivolumab. Nat Commun 8, 14369, doi:10.1038 / ncommsl4369 (2017), Li, C.W. et al. Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity. Nat Commun 7, 12632, doi:10.1038 / ncomms12632 (2016)).

[0493] uPa cleavage of masked anti-EGFR, anti-mesothelin, anti-TF, anti-CD19, anti-cMet, and anti-CDH3 variants in Example 13

[0494] To evaluate the release of partial or all of the maskers from Fab's at several different complementarity positions caused by the cleavage sites of the expected proteases in the cleavage linker, selected samples generated in Example 11 were treated in vitro with uPa. The reaction was monitored by reducing SDS-PAGE as follows.

[0495] Method

[0496] A preparative cleavage assay of modified variants targeting different epitopes was set up as described in Example 5 and analyzed by non-reducing SDS-PAGE. SDS-PAGE was set up as described in Example 12, except that the samples were denatured using reducing Laemmli buffer. The reducing buffer was obtained by supplementing 4X Laemmli buffer with 10% β-ME.

[0497] Results

[0498] Variants without the uPa cleavage sequence did not show any processing under the tested conditions, while all variants that did contain the uPa-specific sequence between the PD-L1 portion and the VL of the Fab showed complete cleavage (Figure 12) and release of the PD-L1 domain from the light chain. For the unprotected κ light chain, as expected, a decrease in the apparent MW of the LC to approximately 25 kDa was observed after uPa treatment. Due possibly to heterogeneous glycosylation (Li, C.W. et al. Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity. Nat Commun 7, 12632, doi:10.1038 / ncomms12632 (2016)) and small molecular weight (approximately 13 kDa), free PD-L1 portions were not detected in variants targeting EGFR and TF. For MSLN and CD19, faint bands indicating species with an apparent molecular weight of 15 - 20 kDa were detected.

[0499] Example 14 Masking / Unmasking of Anti-EGFR, Anti-Mesothelin, Anti-TF, Anti-CD19, Anti-cMet, and Anti-CDH3 Variants

[0500] The target binding of different complementary site / epitope pairs was evaluated by SPR and flow cytometry on samples generated in Example 11 and processed with uPa in Example 13 as follows.

[0501] Method

[0502] Initial Binding According to Flow Cytometry

[0503] Various cancer cell lines (MDA-MB231, OVCAR3, MDA-MB468, Raji) expressing surface proteins containing the analyte of interest were maintained in their recommended medium supplemented with L-glutamine and appropriate concentrations of serum (complete medium) in a humidified +5% CO2 incubator at 37°C.

[0504] Modified variants targeting different epitopes were diluted 2-fold in complete medium and then serially diluted three-fold in cold complete medium to obtain a total of eight to ten concentration points starting from 300 nM or 150 nM.

[0505] All media were maintained at 4 °C and all incubations were carried out on wet ice. On the day of the assay, exponentially growing cells were harvested using a warm non-enzymatic cell dissociation solution, centrifuged and resuspended in complete medium at a cell density of 2E+06 cells / mL. 50 μL / well of cells were distributed in polypropylene V-bottom 96-well plates (Coming, Coming, NY, USA). An equal volume of 2X test antibody or control was added to the cells and incubated for 2 h. The cells were then washed twice by centrifugation and the supernatant removed. Detection of bound variants was achieved by re-incubating for 1 h with a fluorescently labeled Fc-specific secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA). The cells were washed twice by centrifugation and the cell pellet resuspended in complete medium containing propidium iodide (Invitrogen, Carlsbad, CA, USA), filtered using a 96-well filter plate with a pore size of 0.60 μm (MilliporeSigma, Burlington, MA, USA) and analyzed by flow cytometry using an HTS autosampler (mounted on a BD-LSRII or BD-LSR Fortessa). 2000 live cell / single cell events were acquired per sample.

[0506] The specific MFI for each sample point was calculated by subtracting the MFI value of the negative control (background). A four-parameter non-linear regression curve fitting model with one-site specificity with Hill slope was used to fit the binding curve (curve of specific MFI vs. linear or log antibody concentration) using GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA) to determine the Bmax and apparent Kd values for each test article.

[0507] SPR

[0508] At a temperature of 25 °C using PBS-T (PBS + 0.05% (v / v) Tween 20, pH 7.4) running buffer on a Biacore TMSPR (surface plasmon resonance) binding assays were performed on a T200 instrument (GE Healthcare, Mississauga, ON, Canada) to determine the kinetics and affinity of different subsets of antigens (EGFR, TF, mesothelin) for the modified mAb variants. The CM5 series S sensor chip, Biacore amine coupling kit (NHS, EDC, and 1 M ethanolamine), and 10 mM sodium acetate buffer were all purchased from GE Healthcare. PBS running buffer containing 0.05% Tween 20 (PBS-T) (v / v) was purchased from Teknova Inc. (Hollister, CA). Goat polyclonal anti-human Fc antibody was purchased from Jackson Immuno Research Laboratories Inc. (West Grove, PA). Recombinant proteins of the extracellular domain of human EGFR (Genscript, catalog number Z03194-50) and mature human mesothelin (R&D systems, catalog number 3265-MS-050) were purchased and purified by SEC prior to SPR analysis to ensure the purity and homogeneity of the analytes. The recombinant protein of human TF was expressed in HEK293 cells and purified by anion exchange (Q Sepharose HP, GE Healthcare) followed by SEC purification, and then the recombinant protein was used for SPR.

[0509] Screening the binding of mAb variants to different antigens was performed in two steps: The mAb variants were indirectly captured onto the surface of anti-human Fc specific polyclonal antibody, and then five concentrations of SEC-purified antigens were injected. As described by the manufacturer (GE Healthcare), the anti-human Fc surface was prepared on a CM5 series S sensor chip by standard amine coupling. Briefly, a solution of 25 μg / mL anti-human Fc in 10 mM NaOAc pH 4.5 was injected at a flow rate of 10 μL / min for 7 minutes immediately after EDC / NHS activation until approximately 4500 response units (RU) were immobilized on all four flow cells. The remaining active groups were quenched by injecting 1 M ethanolamine at a flow rate of 10 μL / min for 7 minutes. The mAb to be analyzed was indirectly captured onto the anti-Fc surface (flow cells 2 - 4) by injecting a 2 - 20 μg / mL solution at a flow rate of 10 μL / min for 60 seconds, resulting in an mAb capture level in the range of 130 to 470 RU, depending on the mAb variant. Using single-cycle kinetics, five concentrations of a two-fold dilution series of the antigen were continuously injected at 40 μL / min onto all flow cells including the reference flow cell 1, and buffer blanks were injected onto all flow cells as controls. Details regarding the concentration range of the analyte and the contact and dissociation times are shown in Table 53. The anti-human Fc surface was regenerated for the next injection cycle by applying a 10 mM glycine / HCl pH 1.5 pulse at 30 μL / min for 120 seconds. Using Biacore TM The double-reference-subtracted sensograms were analyzed using the Biacore T200 evaluation software v3.0 and then fitted to a 1:1 Langmuir binding model.

[0510] Table D: SPR Analyte Parameters

[0511] Analyte Concentration range [nM] Contact time [s] Dissociation time [s] EGFR 2.5-40 180 300 TF 0.125-2 300 1800 Mesothelin 0.1252 / 1.25-20 300 / 180 1800

[0512] Results

[0513] Figure 13 shows that the antigen binding of all uncleavable variants (variants 31722, 31728, 31736, 31732, 28647, 28664 for EGFR, MSLN, TF, CD19, cMet, CDH3 respectively) is reduced 30 - 190-fold compared to the corresponding unmasked controls (variants 32474, 16417, 6323, 4372, 17606, 17214 for EGFR, MSLN, TF, CD19, cMet, CDH3 respectively), as determined by cell binding studies. In cases where cleavable variants were included, samples were tested in the absence (-uPa) and presence (+uPa) of uPa treatment. While the uncleavable variants (variants 31722, 31728, 31736, 31732 for EGFR, MSLN, TF, CD19 respectively) showed only minor differences between the uncleaved and uPa-treated samples, the cleavable samples (variants 31723, 31729, 31737, 31733 for EGFR, MSLN, TF, CD19 respectively) showed a significant recovery of binding after uPa treatment. Specifically, the binding levels were similar to those of the uncleavable variants prior to protease treatment, and after uPa cleavage, the binding recovered to within 1.3 - 85-fold of the unmasked controls. Where available (EGFR, TF, mesothelin), SPR binding results showed the same trend of masking and binding recovery after cleavage.

[0514] Example 15 Functional Analysis of Masked Anti-EGFR Variants

[0515] To study the effect of the mask on the function of EGFR-targeting variants generated in Example 11, treated with uPa in Example 13, and tested for target binding in Example 14 in a cell-based assay, the growth inhibition of NCI-H292 cells was studied as follows.

[0516] Method

[0517] For this assay, NCI-H292 cells were routinely grown in 75 cm 2(T75) Flasks were grown at 37 °C + 5% CO2 and passaged twice weekly in FBS medium without added antibiotics. One day before adding antibiotics, cells were seeded in 384-well plates (Corning 3570) at 300, 1000, and 125 cells / 25 μL / well in medium supplemented with 1000 units penicillin, 1000 μg streptomycin, and 2.5 μg amphotericin B per milliliter. On the day of the assay, antibodies and controls were serially diluted at 6-fold the desired final concentration in an 11-point dose-response curve and then added to the plated cells to obtain the final incubation concentrations described in Table 5: Variant Concentration Ranges. Their effect on cell proliferation was measured after incubation at 37 °C, 5% CO2 for 5 days. Incubation with an irrelevant antibody (22277) was used to assess off-target cytotoxicity. Cell viability was determined using CellTiterGlo TM (Promega, Madison) based on quantification of ATP, which indicates the presence of metabolically active cells, in each well. Signal output was measured on a luminescence plate reader (Envision, Perkin Elmer) set to an integration time of 0.1 seconds. The integration time was adjusted to minimize signal saturation at high ATP concentrations.

[0518] Data expressed as relative light units (RLU) were normalized against untreated control wells and expressed as % survival calculated according to the following formula:

[0519] % survival = RLU Ab / RLU untreated × 100.

[0520] Using GraphPad Prism software, dose-response curves were generated to measure efficacy (the maximum saturable growth inhibition response observed at high concentrations) and potency (relative IC50, the concentration required to achieve half-maximal efficacy).

[0521] Table E: Variant Concentration Ranges

[0522]

[0523] Results

[0524] As Figure 14As shown, the anti-EGFR antibody based on cetuximab (32474) inhibits the growth of NCI-H292 cells with an IC50 of 0.11 nM. Treatment with uPa has only a minimal effect on this function. The PD-1:PD-L1 masked variants (31722, 31723) are less potent (IC50 increased 40 - 80-fold) without uPa treatment. When treated with uPa, although the function of the non-cleavable variant 31722 is still significantly inhibited (100-fold), the cleavable 31723 shows a functional recovery to within 2.5-fold of the unmasked v32474. The irrelevant antibody (22277) shows no functional role in the growth inhibition assay.

[0525] Example 16 B7:CD28 family ligand-receptor pairs as masks - CTLA4:CD80

[0526] To determine whether other members of the B7:CD28 family could be used to effectively mask the Fab, the CTLA4:CD80 masked form of the CD3 x Her2 Fab x scFv Fc antibody from Example 1 was generated as follows and its CD3 binding was evaluated.

[0527] Methods

[0528] The masked CTLA4:CD80 CD3 Fab was designed to be equivalent to the PD1:PD-L1 masked variant in Example 1. Briefly, the IgV domain sequences of human CD80 and CTLA4 (West, S.M. & Deng, X.A. Considering B7-CD28 as a family through sequence and structure. Exp Biol Med (Maywood), 1535370219855970, doi: 10.1177 / 1535370219855970 (2019); SEQ ID 25, 26) were attached to the N-termini of the heavy and light chains of the CD3 Fab using one of the linker combinations described in Examples 1 and 10, respectively. Specifically, the CTLA4 IgV domain was fused to the LC with a uPa-cleavable sequence, while the CD80 moiety cannot be removed by proteases. A schematic diagram of the architecture of the variants studied is shown in Figure 15In addition, in order to reduce the homodimerization via CD80 described above (C.C. Stamper et al., Crystal structure of the B7-1 / CTLA-4 complex that inhibits human immune responses. Nature 410, 608-611 (2001)), mutations were introduced into the CD80 portion in some variants. The sequences of the individual chains of the variants are listed in Table F. The production of the antibodies, the evaluation of their sample purity and the cleavage of upa, and the evaluation of the binding to Jurkat cells carrying CD3 were carried out as in Examples 2, 3, 5, and 6, respectively.

[0529] Table F: Sequence composition of the variants tested *

[0530]

[0531] * The CD80 IgV domain attached to the heavy chain is represented by a striped pattern in the cartoon, and the CTLA-4 IgV domain attached to the light chain is shown as a checkerboard pattern.

[0532] Results

[0533] The production of the modified CD3 x Her2 Fab x scFv Fc variant (30444) carrying the CTLA4:CD80-based mask yielded 6.7 mg after preparative SEC, an amount similar to that of the equivalent PD-1:PD-L1 masked variant in Example 2. UPLC-SEC analysis after protein A purification ( Figure 16A ) showed that the dimer was the major species, consistent with the homodimerization interfaces on CD80 and CTLA4 far from the heterodimer interface (Trang, V.H. et al. A coiled-coil masking domain for selective activation of therapeutic antibodies. Nat Biotechnol 37, 761-765, doi:10.1038 / s41587-019-0135-x (2019)). A large amount of high molecular weight species, such as aggregates and oligomers, were also observed and preparative SEC was carried out to remove these unwanted particles. UPLC-SEC of the final SEC purified sample ( Figure 16B ) showed the presence of 84% dimer and 9% monomeric species. In addition, 7% of high molecular weight species remained. Non-reducing CE-SDS ( Figure 16C) shows a spectrum corresponding to a single dominant species, the molecular weight of which is significantly higher than the expected molecular weight of the intact molecule. Bands of the modified heavy and light chains show significantly higher apparent molecular weights in the CE-SDS reducing profile. Similar to the modification based on PD-1:PD-L1 in Example 3, this may be caused by extensive glycosylation of CD80 and CTLA4 (Stamper, C.C. et al Crystal structure of the B7-1 / CTLA-4 complex that inhibits human immune responses. Nature 410, 608-611, doi:10.1038 / 35069118 (2001)). When mutations are introduced in the homodimerization interface of the CD80 moiety, the amount of dimer species found in UPLC-SEC after Protein A purification is reduced to 19-59%, while the amount of monomer species increases to 28-66% ( Figures 16D to 16F ).

[0534] When treated with uPa, the CTLA4 moiety is effectively removed from the light chain, as seen in Figure 17 . Here, the band corresponding to the modified light chain disappears after cleavage, and a band corresponding to the molecular weight of the unmasked light chain appears. No released CTLA4 component was detected after cleavage, which may be due to the heterogeneity caused by small size and glycosylation.

[0535] Binding to CD3 on Jurkat cells was evaluated by ELISA as described in Example 6 ( Figure 18 ), and the results showed that the CD80:CTLA4-based modification (v30444) reduced target binding by approximately 80-fold. This is similar to what was seen in an equivalent variant with a PD-1:PD-L1-based masker (Example 6, v30430 is included here for reference). After uPa cleavage of the CTLA4 moiety, CD3 binding was partially restored (within approximately 4-fold of WT).

[0536] Example 17 Conditional active immunomodulators based on masked immunomodulator-Fc-fusions

[0537] Immune regulatory pairs (e.g., PD-1:PD-L1 (Table G), CD80:CTLA-4) are used as non-targeted conditionally activating molecules in this example. Here, the immune regulatory pair does not provide a masking function for a specific epitope, but is directly fused to Fc as follows.

[0538] Methods

[0539] The constructs studied herein are IgV domains of immune modulator pairs (such as PD-1:PD-L1) that are hinge-fused to a heterodimeric IgG Fc at the N-terminus. The Fc portion of these constructs contains mutations in the CH3 domain that drive heterodimeric pairing of the two chains as previously described (e.g., Kreudenstein, T.S. et al. Improving biophysical properties of a bispecific antibody scaffold to aid developability: quality by molecular design. MAbs 5, 646-654, doi:10.4161 / mabs.25632 (2013); SEQ ID 4,5; other mutations forming heterodimeric Fc are also available in the literature). In one embodiment, mutations are also introduced in the two CH2 domains to abolish binding to Fcγ receptors (SEQ ID 6,). When one immune modulator IgV domain (e.g., the high-affinity form of PD-1, Maute, R.L. et al. Engineering high-affinity PD-1 variants for optimized immunotherapy and immuno-PET imaging. Proc Natl Acad Sci U S A 112, E6506-6514, doi:10.1073 / pnas.1519623112 (2015), SEQ ID 9) is directly fused to the N-terminus of the IgG hinge, the amino acid sequence (MSGRSANA) recognized and cleaved by uPa is introduced between the hinge and another immune modulatory IgV domain (e.g., WT PD-L1, SEQ ID 8) on the other chain.

[0540] This design results in the generation of a conditionally active monovalent PD-L1 targeting molecule that is directly fused to IgG Fc via a protease-cleavable peptide linker ( Figure 19)。In the absence of uPa, high-affinity PD-1:PD-L1 dimers are formed intramolecularly and prevent undesired systemic binding to PD-L1. When exposed to uPa, e.g., in the tumor microenvironment (TME), PD-L1 is released and the PD-1 moiety can bind to PD-L1 expressed on tumor cells. In the TME, checkpoint activity is thus selectively blocked and the sensitivity of tumor cells to cytotoxic T cells is enhanced. Other immunomodulatory ligand-receptor pairs, such as CD80:CTLA-4 or SIRPa:CD47, are also used as masks. For CD80:CTLA-4, CTLA-4 is released only in the presence of the correct TME-related protease and the remaining CD80 can bind to CD28 or CTLA-4 on T cells and thereby exert its immunomodulatory function. For the SIRPα:CD47 mask, the CD47 moiety is released by proteolytic cleavage in the TME, freeing SIRPα to bind to CD47 on macrophages, thereby inhibiting checkpoint activity and increasing phagocytosis and tumor cell killing.

[0541] The binding of the variants treated with or without uPa to PD-L1 was tested by flow cytometry as described in Example 8. The same samples were tested in a reporter gene assay (RGA) (Promega, Madison, WI, USA) sensitive to PD-1:PD-L1 checkpoint inhibition. The RGA was performed similarly to the RGA in Example 9, except that CHO cells expressing PD-L1 and directed to the TCR were used together with modified Jurkat T cells according to the manufacturer's protocol.

[0542] Table G: Sequence composition of the variants tested *

[0543]

[0544] *PD-1IgV is represented by a striped pattern in the cartoon and the PD-L1 IgV domain is shown as a checkerboard pattern.

[0545] Results

[0546] In the absence of uPa treatment, ZW Fc1 does not bind to PD-L1 in flow cytometry assays. This is due to the tight intramolecular interaction between the high-affinity form of the PD-1 IgV domain and PD-L1 in the Fc assembly. When treated with uPa, ZW Fc1 binds tightly to PD-L1 in SPR and flow cytometry assays. This is expected because after cleavage of the uPa-specific sequence in the linker, the PD-L1 moiety is released and the PD-1 domain retained on the Fc binds freely to PD-L1 in the assay. Similarly, uncleaved ZW Fcl by uPa is inactive in the PD-1:PD-L1 RGA, while robust activity is exhibited when treated with uPa.

[0547] Example 18 Evaluation of Masking Techniques in the Anti-CD40 System

[0548] The PD-1:PD-L1-based masks described in Examples 1-15 were applied to the complementary site targeting CD40, and the sample quality, target binding, and effect of the mask on function of the resulting variants were evaluated as follows.

[0549] Methods

[0550] Variant Design and Production

[0551] The PD-1:PD-L1 masking form of the full-size antibody containing the anti-CD40 complementary site described above (R. H. Vonderheide et al., Clinical activity and immune modulation in cancer patients treated with CP-870,893, a novel CD40 agonist monoclonal antibody. J Clin Oncol 25, 876-883 (2007)) was constructed as described in Example 10. The resulting constructs and their sequences are summarized in Table H.

[0552] Table H: Sequences of Anti-CD40 Variants

[0553]

[0554]

[0555] The heavy and light chain sequences of the described variants were introduced into an expression vector to express the heavy and light chain sequences in Expi293F TMThe heavy and light chain sequences were expressed in cells and purified using the two-step purification process described in Example 11. The purity and sample homogeneity of the purified samples were then evaluated by UPLC-SEC and non-reducing gel electrophoresis as described in Example 3. After purification, the samples were treated with uPa as described in Example 5 and their processing was evaluated by non-reducing CE-SDS. Then, the target binding of both the samples without uPa treatment and the samples treated with uPa to Raji cells was evaluated by flow cytometry as described in Example 14.

[0556] CD40 RGA

[0557] To functionally evaluate the untreated (-uPA) and uPa-treated (+uPA) variants, a CD40 reporter gene assay (RGA) was performed. HEK Blue CD40L cells (Invivogen, San Diego, CA, USA, hkb-cd40 lot 38-01-hkbcd40) were stripped with PBS and then resuspended at 2.78 × 10 5 cells / mL in pre-warmed test medium (Gibco TM DMEM (Thermo Fisher Scientific, Waltham, MS, USA, 1195-040) supplemented with 10% heat-inactivated Gibco TM FBS (Thermo Fisher Scientific, Waltham, MS, USA, 12483-020 lot 1996160) (56 °C, 30 min) and 100 U / mL Gibco TM Pen-Strep (Thermo Fisher Scientific, Waltham, MS, USA, 15070-063 lot 1989510)). WT-CHOK1 (ATCC, Manassas, VA, USA, ATCC CCL-61, lot 70014310) and FcgR2B-CHOK1 cells (BPS Bioscience, San Diego, CA, USA, 79511, lot 191104-41) were stripped with trypsin and resuspended at 5.56 × 10 5Resuspend in test medium at the indicated cell density. Then, add 25,000 HEK Blue CD40 cells (90 μL) to 20 μL of the variants serially diluted in test medium (10 μg / mL - 0.000001 μg / mL), followed by 50,000 WT-CHOK1, FcYR2B-CHOK1 cells (90 μL) or 90 μL of test medium. After incubation at 37 °C, 5% CO2 for 20 - 24 h, mix 20 μL of the supernatant with 180 μL of Quanti-Blue TM solution (Invivogen, San Diego, CA, USA), incubate at 37 °C, 5% CO2 for 3 h, and measure the OD 620nm . Test articles included un-uPa-treated and uPa-treated CD40-targeted variants, as well as irrelevant control antibodies targeting RSV and CD40L (Invivogen, San Diego, CA, USA) as negative and positive controls, respectively.

[0558] Results

[0559] As Figures 20A to 20C shown, for the masked variants v32478 and v32479, after SEC purification, the anti-CD40 variants showed a dominant species with a purity of 92% - 100% in UPLC-SEC, but there was a small amount of higher molecular weight species (7 - 8%). For all variants, non-reducing CE-SDS analysis ( Figure 20D ) also showed a single dominant species. While the apparent molecular weight of the major species of the unmasked v32477 was approximately 150 kDa, as expected, the PD-1:PD-L1 masked variants v32478 and v32479 showed significantly higher apparent molecular weights (>250 kDa), which may be due to glycosylation, as seen in constructs using the same masking domain in Examples 3 and 12. For all variants, reducing CE-SDS ( Figure 20D ) showed two species with different molecular weights corresponding to the heavy and light chains. For the masked variants v32478 and v32479, the apparent molecular weights of both the heavy and light chains were also higher than expected (approximately 100 kDa vs 63 kDa for HC and approximately 50 kDa vs 37 kDa for LC), which may be due to glycosylation of both PD-1 and PD-L1, and as seen in Examples 3 and 12.

[0560] The three anti-CD40 variants studied here were treated with uPa after production, and cleavage was monitored by reducing CE-SDS ( Figure 20E)。Due to the lack of specific cleavage sites, v32477 and v32478 showed no change after incubation with uPa, while processing was seen in the light chain of v32479. Here, the PD-L1 portion was removed by cleavage of the upa-specific sequence in the linker between the C-terminus of PD-L1 and the N-terminus of the VL domain. This resulted in the detection of three fragments in reduced CE-SDS after cleavage: the unchanged PD-1 masking heavy chain lacking the uPa site, the chain corresponding to VL-CL of the κ light chain, and the chain corresponding to the released PD-L1 portion.

[0561] The binding of uPa-treated and untreated samples to CD40 on Raji cells was tested by flow cytometry. As Figure 20F shown, the unmasked v32477 showed a binding curve with an EC50 value of 1 nM, while the binding of the masked v32478 was reduced 40 - 70-fold. Both variants lack the uPa cleavage site, so the binding is not affected by uPa treatment. The binding of untreated v32479 was reduced 14-fold but recovered to within 5-fold when treated with uPa.

[0562] When the functionality of the same samples was probed in a CD40-specific RGA, these trends were reproduced ( Figure 20G ). While v32477 showed robust independent activity that could be further enhanced by FcγR2B-CHOK1, the function of v32478 was seen to be reduced 90 - 110-fold. Since both variants lack the uPa cleavage site, they showed the same activity in RGA experiments with or without uPa treatment. For v32479 untreated with uPa, a similar level of activity masking (55-fold) to v32478 was observed. The activity of v32487 treated with uPa could be detected within 2-fold of v32477. The positive control CD40L induced CD40 activity independent of the presence of FcγR2B, and the negative control (v22277) could not activate CD40 in this assay. The maximum activity level (B max ) observed among the tested variants in this assay was greater in the presence of the FcgR2B-positive cell line, which was the opposite of when there was no FcgR2B on the secondary cell line. Even in the absence of the FcgR2B-positive cell line, treatment with CD40L caused the same B max increase.

[0563] Example 19: SIRPα as a masking agent: CD47 immunomodulation pair

[0564] To determine whether immune checkpoint pairs outside the B7:CD28 family can be used for highland masking of Fab, the CD47:SIRPα masking format of the anti-EGFR antibody described in Production Example 10 below was produced and its EGFR binding was evaluated.

[0565] Method

[0566] The CD47:SIRPα masked anti-EGFR antibody was designed to be equivalent to the PD1:PD-L1 masked variant described in Example 10. Briefly, the IgV domain sequence of human CD47 and a modified, affinity-increased variant of human SIRPα (K. Weiskopf et al., Engineered SIRPα variants as immunotherapeutic adjuvants to anticancer antibodies. Science 341, 88 - 91 (2013)) were attached to the N-termini of the heavy and light chains of the anti-EGFR Fab, respectively, using the uPa-cleavable linker described in Examples 1 and 10. A schematic diagram of the architecture of the variants studied is shown in Figure 27 Figure. The sequences of the individual chains of the variants are listed in Table I. The production of the antibodies, their sample purity, and the evaluation of uPa cleavage were carried out as described in Examples 2, 3, and 5, respectively. Then, the binding to H292 cells carrying EGFR was evaluated by quantitative fluorescence microscopy.

[0567] Table I: Sequence composition of the test variants *

[0568]

[0569] * The SIRPα IgV domain attached to the heavy chain is represented by a striped pattern in the cartoon, and the CD47 IgV domain attached to the light chain is shown as a checkerboard pattern.

[0570] Initial binding to H292 cells according to fluorescence microscopy

[0571] The NCI-H292 cell line expressing EGFR was maintained at 37 °C in a humidified +5% CO2 incubator in RPMI-1640 (complete medium) supplemented with L-glutamine and 10% FBS. One day before the assay, exponentially growing cells were harvested using 0.05% trypsin and resuspended in complete medium at a cell density of 1.2×10 5 cells / ml. At TC-treated microplates (Code 3882, Coming, Coming, NY, USA) made of 96-well half-area transparent flat-bottom black polystyrene were each dispensed with 50 μL of cells to obtain a final concentration of 6000 cells / mL, and the cells were incubated overnight at 37 °C in a humidified +5% CO2 incubator. On the day of the experiment, before performing the assay, the plates with cells were cooled to 4 °C and held for 30 minutes. The modified variants were diluted to 2-fold their final concentration in cold DPBS (Wisent Bioproduct, St-Bruno, Quebec, Canada) containing Ca 2+ and Mg 2+ , and then serially diluted in triplicate to obtain a total of 11 concentration points starting from 100 nM. All solutions were kept at 4 °C, and all incubations were performed at 4 °C. An equal volume of 2X test variant or control was added to the cells and incubated for 2 hours. Then, in a BioTek EL405 select plate washer (BioTek, Winooski, VT, USA), with cold DPBS containing Ca 2+ and Mg 2+Wash the cells with cold DPBS for 3 wash cycles, 150 μL per well per cycle, with a final residual volume of 25 μL. Detection of the bound variants was achieved by re-incubating for one hour with a fluorescent labeling mixture containing AF488-labeled human Fc-specific secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA), erythrocyte masking agent (Molecular Probes, Eugene, Oregon, USA), and Hoechst 33342 (Molecular probes, Eugene, Oregon, USA) in the presence of FBS (Wisent Bioproduct, St-Bruno, Quebec, Canada). Wash the cells twice (3 cycles, 150 μL per wash per well) in a BioTek EL405 select (BioTek, Winooski, VT, USA) plate washer. Capture images in an ImageXpress Micro XLS (Molecular Devices, San Jose, CA, USA) using transmitted light, DAPI (blue channel), Cy5 (far red channel), and FITC (green channel). Perform image analysis using MetaXpress analysis software Custom Module Editor (CME) (Molecular Devices, San Jose, CA, USA). For each well, measure the total green fluorescence intensity in the well area covered by cells, and then normalize this total green fluorescence intensity against the cell area. This normalized value, "total intensity per cell area", was used for curve fitting analysis in GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA). The baseline value was calculated using the average normalized green background fluorescence signal of the control wells (these are the wells incubated only with the fluorescent labeling mixture of the secondary antibody). The baseline value for each plate was subtracted from all the data before applying the non-linear fitting model. For each test article, fit the curve of the specific total intensity (baseline-corrected) per cell area against the logarithm of the antibody concentration with a non-linear regression curve fitting model of "unit point-specific binding with Hill slope".

[0572] Results

[0573] Production of the modified anti-EGFR variant (34164) carrying a CD47:SIRPα-based mask yielded 0.33 mg after preparative SEC. UPLC-SEC analysis after Protein A purification showed the presence of a large amount of high molecular weight species, such as aggregates and oligomers, in addition to the main species, and preparative SEC was performed to remove these unwanted particles. UPLC-SEC of the final SEC-purified sample ( Figure 28A ) showed the presence of 91% of the desired species. Non-reducing CE-SDS ( Figure 28B ) showed a profile corresponding to a single dominant species with a molecular weight significantly higher than the expected molecular weight of the intact molecule. The band for the CD47-modified light chain showed a significantly higher apparent molecular weight than expected in the reducing CE-SDS profile, overlapping with the modified heavy chain. Similar to the PD-1:PD-L1-based modification in Example 3, this may be due to extensive glycosylation of CD47 (W.J. Mawby, C.H. Holmes, D.J. Anstee, F.A. Spring, M.J. Tanner, Isolation and characterization of CD47 glycoprotein: a multispanning membrane protein which is the same as integrin-associated protein (IAP) and the ovarian tumour marker OA3. Biochem J 304 (Pt 2), 525-530 (1994)).

[0574] When treated with uPa, both the CD47 and SIRPα moieties were effectively removed from the light chain, as seen in Figure 29 . Here, the bands corresponding to the modified heavy and light chains disappeared after cleavage, and bands corresponding to the molecular weights of the de-masked heavy and light chains appeared. The released CD47 and SIRPα components could not be clearly identified after cleavage, probably due to their small size and heterogeneity caused by glycosylation.

[0575] Binding to EGFR on H292 cells, as assessed by high content analysis (Figure 30), showed that the CD47:SIRPα-based mask in v34164 reduced target binding by 37-fold. This was similar to that observed in the equivalent variant with a PD-1:PD-L1 mask in Example 14. After cleavage of both mask components with uPa, EGFR binding recovered to within 1.1-fold of WT.

[0576] Example 20: Target co-engagement and bridging of anti-CD3 trispecific variants

[0577] To determine whether PD-L1, Her2, and CD3 can be simultaneously engaged by the anti-CD3 variants described in Examples 1-9, Her2-PD-L1 co-engagement and T cell bridging studies were conducted as follows.

[0578] Method

[0579] Simultaneous binding assessment of Her2 and PD-L1 according to flow cytometry

[0580] JIMT-1 (Leibniz Institute, Braunschweig, Germany) cultured in growth medium consisting of DMEM medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA) was horizontally maintained in a T-175 flask (Corning, Coming, NY) in an incubator containing 5% carbon dioxide at 37 °C. The antibodies were titrated from 100 nM to 1.7 pM at a 1:3 dilution in a total of 20 μL / well in a 96-well V-bottom plate (Thermo Fisher Scientific, Waltham, MA) in FACS buffer, which is PBS containing 2% FBS (Thermo Fisher Scientific, Waltham, MA). The tumor cells were rinsed with PBS (Thermo Fisher Scientific, Waltham, MA), harvested with TrypLE Express (Thermo Fisher Scientific, Waltham, MA), diluted in medium, and counted using a Countess automated cell counter (Thermo Fisher Scientific, Waltham, MA). The tumor cells were washed and resuspended in FACS buffer and added to the 96-well plate at 50,000 cells per well. The cells were incubated with the variants at 4 °C for 1 hour. After incubation, the cells were washed 2 times with FACS buffer, and then 1 mg / mL secondary antibody AF647 goat anti-human IgG Fc (Jackson ImmunoResearch, West Grove, PA) and a 1000-fold diluted viability dye (Thermo Fisher Scientific, Waltham, MA) were added to the wells. The plate was incubated at room temperature for 30 minutes. The cells were washed 2 times in FACS buffer and resuspended in 100 μL FACS buffer. For assay readings, the geometric mean of APC fluorescence was measured by flow cytometry on a BD Celesta (BD Biosciences, San Jose, CA). The raw data were analyzed on FlowJo, LLC software (Becton, Dickinson & Company, Ashland, OR). Graphs were generated using GraphPad Prism version 8.1.2 for Mac OS X (GraphPad Software, La Jolla, CA).

[0581] CD3 / Her2 / PD-L1 bridging assay

[0582] JIMT-1 (Leibniz Institute, Braunschweig, Germany) cultured in growth medium consisting of DMEM medium (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Waltham, MA) was horizontally maintained in a T-175 flask (Coming, Corning, NY) in an incubator containing 5% carbon dioxide at 37 °C. The tumor cells were rinsed with PBS (Thermo Fisher Scientific, Waltham, MA), harvested with TrypLE Express (Thermo Fisher Scientific, Waltham, MA), diluted in PBS, and washed twice in PBS. A vial of primary human pan T cells (BioIVT, Westbury, NY) was thawed in a 37 °C water bath, washed in growth medium consisting of RPMI-1640 ATCC Modified (Thermo Fisher Scientific, Waltham, MA) supplemented with 10% fetal bovine serum, subsequently washed in PBS, and resuspended in PBS. T cells and tumor cells were counted using a Countess automated cell counter (Thermo Fisher Scientific, Waltham, MA), and the cells were resuspended in PBS at 5M / mL. The cell proliferation dye - eF670 (Thermo Fisher Scientific, Waltham, MA) was added to the tumor cells at 1.25 uM. CellTracker Green (Thermo Fisher Scientific, Waltham, MA) was added to the T cells at 2 uM. The T cells and tumor cells were incubated in the dark at 37 °C for 20 minutes and washed twice in FACS buffer - PBS (Thermo Fisher Scientific, Waltham, MA) containing 2% FBS. The antibodies were titrated down from 10 nM to 0.2 pM at a 1:6 dilution in FACS buffer in a V-bottom 96-well plate (Thermo Fisher Scientific, Waltham, MA) at a total of 50 uL / well. The pan T cells were mixed with the tumor cells at 1.44E6 cells / mL at a 5:1 effector:target ratio. 50 uL of the mixed cell suspension was added to the plate containing the titrated variants. The cells were incubated with the va...

Claims

1. A fusion protein comprising: A biofunctional protein, a ligand-receptor pair, a first peptide linker, and a second peptide linker; wherein The biofunctional protein comprises an antibody or an antigen-binding antibody fragment, the antibody or antigen-binding antibody fragment comprising a first VH polypeptide and a first VL polypeptide, the first VH and VL polypeptides forming a first antigen-binding domain; and The ligand-receptor pair comprises an immunoglobulin superfamily receptor and the extracellular portion of its cognate ligand or a receptor-binding fragment thereof; wherein The ligand is fused via the first peptide linker to one of the first VH or VL polypeptides, and the receptor is fused via the second peptide linker to the other of the first VH or VL polypeptides, wherein the ligand-receptor pair sterically hinders the binding of the first antigen-binding domain to its cognate antigen; wherein The first and second peptide linkers have sufficient length to allow the ligand and receptor to pair, and at least one of the first and second peptide linkers comprises a protease cleavage site, wherein At least one of the ligand or the receptor in the ligand-receptor pair is capable of binding to an immunomodulatory target, and wherein The ligand-receptor pair is selected from the group consisting of: PD1-PDL1, PD1-PDL2, CTLA4-CD80, CD28-CD80, CD28-CD86, CTLA4-CD86, PDL1-CD80, ICOS-ICOSL, NCRSRLG1-NKp30, and CD47-SIRPa.

2. The fusion protein according to claim 1, wherein the biofunctional protein comprises a polypeptide scaffold.

3. The fusion protein according to claim 2, wherein the polypeptide scaffold comprises a dimeric Fc region.

4. The fusion protein according to claim 3, wherein the dimeric Fc region is a heterodimeric Fc.

5. The fusion protein according to claim 1, wherein the receptor comprises one or more mutations that increase or decrease the binding affinity of the receptor for its cognate ligand compared to the wild-type receptor.

6. The fusion protein according to claim 1, wherein the ligand comprises one or more mutations that increase or decrease the binding affinity of the ligand for its cognate receptor compared to the wild-type ligand.

7. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is PD1-PDL1.

8. The fusion protein according to claim 7, wherein the ligand PDL1 comprises the amino acid sequence according to SEQ ID NO:

8.

9. The fusion protein according to claim 7, wherein the receptor PD1 comprises the amino acid sequence according to SEQ ID NO:

9.

10. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is CTLA4-CD80.

11. The fusion protein according to claim 10, wherein the ligand CD80 comprises the amino acid sequence according to SEQ ID NO: 25, SEQ ID NO: 185, SEQ ID NO: 187 or SEQ ID NO:

189.

12. The fusion protein according to claim 10, wherein the receptor CTLA4 comprises the amino acid sequence according to SEQ ID NO:

26.

13. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is selected from the group consisting of: CTLA4-CD80, PDL1-CD80 and CD28-CD80, and wherein the ligand CD80 comprises the amino acid sequence according to SEQ ID NO: 25, and the amino acid sequence has mutations selected from the group consisting of: (a) H18Y, A26E, E35D, M47S, I61S and D90G; (b) E35D, M47S, N48K, I61S, K89N; (c) E35D, D46V, M47S, I61S, D90G, K93E; (d) H18Y, A26E, E35D, M47S, I61S, V68M, A71G, D90G; (e) I58S, V68S, L70S; (f) M47S, I61S or (g) V22S.

14. The fusion protein according to any one of claims 1 to 6, wherein the receptor and the ligand are fused to the corresponding N-termini of the first VH and VL polypeptides.

15. The fusion protein according to any one of claims 1 to 6, wherein one of the first or second peptide linkers comprises more than one protease cleavage site.

16. The fusion protein according to any one of claims 1 to 6, wherein the protease is selected from the group consisting of: serine protease, MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, MMP17, MMP18 (collagenase 4), MMP19, MMP20, MMP21, disintegrin, Serratia protease, astaxanthin, caspase 1, caspase 2, caspase 3, caspase 4, caspase 5, caspase 6, caspase 7, caspase 8, caspase 9, caspase 10, caspase 11, caspase 12, caspase 13, caspase 14, cathepsin A, cathepsin B, cathepsin D, cathepsin E, cathepsin K, cathepsin S, granzyme B, guanidobenzoic acid enzyme (GB), heparinase, elastase, podoplanin, proteolytic enzyme, proteolytic enzyme 2, methyldopa, neuropilin, MT-SP1, neprilysin, plasmin, PSA, PSMA, TACE, TMPRSS3, TMPRSS4, uPA, calpain, FAP and KLK.

17. The fusion protein according to any one of claims 1 to 6, wherein the protease is uPA or a protein lysing enzyme.

18. The fusion protein according to any one of claims 1 to 6, wherein the peptide linker has a length of 3 to 50 or 5 to 20 amino acids.

19. The fusion protein according to any one of claims 1 to 6, wherein one of the first or second peptide linkers does not have a protease cleavage site.

20. The fusion protein according to any one of claims 1 to 6, wherein the peptide linker is a (Gly n Ser) linker, and the (Gly n Ser) linker comprises an amino acid sequence selected from the group consisting of: (Gly3Ser) n (Gly4Ser)1, (Gly3Ser)1(Gly4Ser) n , (Gly3Ser) n (Gly4Ser) n and (Gly4Ser) n , where n is an integer from 1 to 5.

21. The fusion protein according to any one of claims 1 to 6, wherein the peptide linker is (EAAAK) n linker, where n is an integer between 1 and 5.

22. The fusion protein according to claim 21, wherein the peptide linker comprises the amino acid sequence EAAAKEAAAK (SEQ ID.NO:38).

23. The fusion protein according to any one of claims 1 to 6, wherein the peptide linker is a polyproline linker; or a glycine-proline linker.

24. The fusion protein according to claim 23, wherein the polyproline linker is PPP or PPPP; the glycine-proline linker is GPPPG, GGPPPGG, GPPPPG or GGPPPPGG 25. The fusion protein according to any one of claims 1 to 6, wherein the peptide linker comprises an immunoglobulin hinge region sequence, and the immunoglobulin hinge region sequence comprises an amino acid sequence having at most 30% difference in amino acid sequence identity compared to the wild-type immunoglobulin hinge region amino acid sequence.

26. The fusion protein according to any one of claims 1 to 6, wherein the peptide linker comprises a protease cleavage site, and the protease cleavage site comprises the amino acid sequence MSGRSANA (SEQ ID NO:28).

27. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is PDL1-CD80, and the PDL1 comprises the amino acid sequence according to SEQ ID NO:

8.

28. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is CD28-CD80, and the CD28 comprises the amino acid sequence according to SEQ ID NO:

254.

29. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is CD28-PDL1.

30. The fusion protein according to claim 29, wherein the CD28 comprises the amino acid sequence according to SEQ ID NO:

254.

31. The fusion protein according to claim 29, wherein the PDL1 comprises the amino acid sequence according to SEQ ID NO:

8.

32. The fusion protein according to any one of claims 1 to 6, wherein the ligand-receptor pair is CD47-SIRPa.

33. The fusion protein according to claim 32, wherein the SIRPa comprises the amino acid sequence according to SEQ ID NO:

255.

34. The fusion protein according to claim 32, wherein the CD47 comprises the amino acid sequence according to SEQ ID NO:

254.

35. The fusion protein according to any one of claims 1 to 6, wherein one of the ligand or the receptor is engineered to comprise one or more additional protease cleavage sites, and wherein the one or more protease cleavage sites in the ligand or the receptor and the protease cleavage site in the first or second peptide linker are cleavable by the same protease or different proteases.

36. The fusion protein according to any one of claims 1 to 6, wherein the binding of the first antigen-binding domain to its cognate antigen is reduced by 10-fold or more compared to the parental antigen-binding domain not fused to the ligand-receptor pair.

37. The fusion protein according to any one of claims 1 to 6, wherein cleavage of the protease cleavage site in a cellular environment releases a member of the ligand-receptor pair from the fusion protein, thereby allowing the antigen-binding domain to bind its cognate antigen.

38. The fusion protein according to any one of claims 1 to 6, wherein the first antigen-binding domain is a Fab.

39. The fusion protein according to any one of claims 1 to 6, wherein the first antigen-binding domain binds an antigen expressed on a cancer cell or an immune cell.

40. The fusion protein according to any one of claims 1 to 6, wherein the first antigen-binding domain binds an antigen expressed on a T cell.

41. The fusion protein according to any one of claims 1 to 6, wherein the first antigen-binding domain binds to a tumor-associated antigen (TAA).

42. The fusion protein according to any one of claims 1 to 6, wherein the first antigen-binding domain binds to an antigen selected from the group consisting of cluster of differentiation 3 (CD3), human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), mesothelin (MSLN), tissue factor (TF), cluster of differentiation 19 (CD19), tyrosine-protein kinase Met (c-Met), and cadherin 3 (CDH3).

43. The fusion protein according to any one of claims 1 to 6, wherein the fusion protein comprises a second antigen-binding domain, the second antigen-binding domain comprising a second VH polypeptide and a second VL polypeptide.

44. The fusion protein according to claim 43, wherein the fusion protein comprises a second ligand-receptor pair, wherein the ligand of the second ligand-receptor pair is fused to one of the second VH or VL polypeptides via a third peptide linker, and the receptor of the second ligand-receptor pair is fused to the other of the second VH or VL polypeptides via a fourth peptide linker, wherein at least one of the third and fourth peptide linkers comprises a protease cleavage site, and wherein the ligand-receptor pair sterically hinders the binding of the second antigen-binding domain to its cognate antigen.

45. The fusion protein according to claim 43, wherein the fusion protein binds to two different antigens.

46. The fusion protein according to claim 45, wherein one antigen is an antigen expressed by T cells and the other antigen is an antigen expressed by cancer cells.

47. The fusion protein according to claim 46, wherein the antigen expressed by T cells is CD3.

48. The fusion protein according to claim 47, which comprises anti-CD3 complementarity determining regions comprising VH and VL, wherein the VH comprises three CDRs, namely HCDR1, HCDR2 and HCDR3, and the VL comprises three CDRs, namely LCDR1, LCDR2 and LCDR3, wherein (a) HCDR1, HCDR2 and HCDR3 are SEQ ID NO:207, 208 and 209 respectively, and LCDR1, LCDR2 and LCDR3 are SEQ ID NO:211, 212 and 214 respectively; (b) HCDR1, HCDR2 and HCDR3 are SEQ ID NO:224, 225 and 226 respectively, and LCDR1, LCDR2 and LCDR3 are SEQ ID NO:228, 229 and 230 respectively; (c) HCDR1, HCDR2 and HCDR3 are SEQ ID NO:232, 233 and 234 respectively, and LCDR1, LCDR2 and LCDR3 are SEQ ID NO:236, 237 and 238 respectively; or (d) HCDR1, HCDR2 and HCDR3 are SEQ ID NO:240, 241 and 242 respectively, and LCDR1, LCDR2 and LCDR3 are SEQ ID NO:244, 245 and 246 respectively.

49. The fusion protein according to claim 45, wherein the fusion protein binds to CD3 and HER2.

50. A fusion protein, comprising: a Fab region and an Fc region; wherein the Fab region comprises a VH polypeptide and a VL polypeptide that form an antigen-binding domain, and a ligand-receptor pair, which comprises an immunoglobulin superfamily receptor and the extracellular portion of its cognate ligand or a receptor-binding fragment thereof; wherein the ligand is fused to the N-terminus of one of the VH or VL polypeptides via a first peptide linker, and the receptor is fused to the N-terminus of the other VH or VL polypeptide via a second peptide linker; wherein the first and second peptide linkers have a length sufficient to allow pairing of the ligand and the receptor; wherein at least one of the first and second peptide linkers comprises a protease cleavage site; wherein the ligand-receptor pair sterically hinders binding of the antigen-binding domain to its cognate antigen, and wherein the ligand-receptor pair is selected from the group consisting of: PD1-PDL1, PD1-PDL2, CTLA4-CD80, CD28-CD80, CD28-CD86, CTLA4-CD86, PDL1-CD80, ICOS-ICOSL, NCRSRLG1-NKp30 and CD47-SIRPa.

51. The fusion protein according to claim 50, further comprising an additional Fab region or scFv.

52. Use of the fusion protein according to any one of claims 1 - 51 in the preparation of a medicament for the treatment of cancer.

53. Use of the fusion protein according to any one of claims 1 - 51 in the preparation of a medicament for modulating an immune response.

54. The use according to claim 53, wherein the immune response is selected from the group consisting of: inhibition of immune checkpoints, stimulation of immune checkpoints, immune cell activation, stimulation of T cell receptor signal transduction, T cell - dependent cytotoxicity (TDCC), antibody - dependent cell phagocytosis (ADCP), and stimulation of antibody - dependent cell cytotoxicity (ADCC).

55. The use according to any one of claims 52 to 54, wherein the fusion protein is administered intravenously.

56. A vector encoding an amino acid sequence, comprising at least one polypeptide of the fusion protein according to any one of claims 1 - 51.

57. A cell comprising the vector according to claim 56.

58. A kit comprising the vector according to claim 56, the cell according to claim 57, the fusion protein according to any one of claims 1 to 51, or a combination thereof, and instructions for use.

59. The fusion protein according to any one of claims 1 to 6, wherein cleavage of the protease cleavage site in a cellular environment releases one member of the ligand - receptor pair from the fusion protein, thereby allowing the other member of the ligand - receptor pair to bind its cognate partner on the cell surface.

Images