EA046634B1 - A NEW FUSION PROTEIN SPECIFIC TO CD137 AND PD-L1 - Google Patents

Description

I. State of the art

Programmed cell death ligand 1 or PD-L1 (also known as cluster of differentiation 274 or CD274 and B7 homolog 1 or B7-H1) is a type I single-pass membrane protein belonging to the B7 family of co-stimulatory/co-inhibitory antigen presentation molecules. The extracellular portion of PD-L1 contains two domains, an N-terminal IgV-like domain and an IgC-like domain. PD-L1 has a short cytoplasmic domain without any obvious signal transduction motif, leading to the initial belief that PD-L1 as a receptor lacks signal transduction. However, recent data indicate that the cytoplasmic domain of PD-L1 contains non-classical conserved signal transduction motifs capable of inhibiting interferon (IFN) transduction and protecting cancer cells from IFN cytotoxicity (Gato-Canas et al., Cell Rep, 2017).

PD-L1 plays a key role in suppressing the immune system during pregnancy, chronic infections, tissue allotransplantation, autoimmune diseases and cancer. PD-L1 is expressed on a variety of cell types, including B cells, T cells, macrophages, myeloid dendritic cells, mast cells, epithelial cells, and vascular endothelial cells. It is also expressed in several types of cancer, including but not limited to melanoma, lung, bladder, colorectal, and breast cancers. High levels of PD-L1 expression are associated with increased tumor aggressiveness, mediating exhaustion and anergy of tumor-infiltrating T cells, secretion of immunosuppressive cytokines, and protection from lysis by cytotoxic T cells.

PD-L1 is a ligand for programmed cell death protein 1 (PD-1), a key inhibitory immune checkpoint receptor that is primarily expressed on activated T cells, but also on other cells of the immune system, including B cells and monocytes. PD-1 is a member of the immunoglobulin family containing an IgV-like extracellular domain, a transmembrane domain, and a cytoplasmic tail with ITIM (immunoreceptor tyrosine-based inhibitory motif) and ITSM (immunoreceptor tyrosine-based switching motif). Binding of PD-1 to PD-L1 results in the recruitment of Src homology domain 2-containing tyrosine phosphatases 1 and 2 (SHP 1 and 2) to intracellular PD-1 switch motifs and the expression of CBL family E3 ubiquitin ligases. These ubiquitin ligases then ubiquitinate and inactivate key TCR signal transduction mediators, resulting in the removal of the TCR from the cell surface (Karwacz et al., EMBO Mol Med, 2011). Phosphatases SHP 1 and SHP 2 inhibit TCR signaling directly by stopping phosphorylation of ZAP70 and PI3K. In addition, PD-1 bound through PD-L1 can cause inhibition of TCR signaling pathways by affecting the expression and activity of CK2 and cyclin-dependent kinases (CDKs) (Arasanz et al., Oncotarget, 2017). PD-1 binding has also been shown to reprogram T cell metabolism from increased glycolysis, which is required to produce energy for effector functions, to fatty acid β-oxidation, which is associated with long-lived cells. This may also explain the survival and persistence of cells with high levels of PD-1 expression in patients with chronic infections and cancer (Patsoukis et al., Nat Commun, 2015).

Blocking the PD-1/PD-L1 interaction with anti-PD-1 or anti-PD-L1-targeted agents can reverse immune checkpoint function and attenuate the inhibition of T cell responses. There are currently three anti-PD-L1 antibodies approved for cancer treatment: atezolizumab (TECENTRIQ, MPDL3280A, RG7466), avelumab (BAVENCIO, MSB0010718C) and durvalumab (IMFINZI, MEDI4736). Several successful clinical studies of these antibodies have shown high objective response rates, sustained response, or improved survival rates in bladder cancer, skin cancer, and lung cancer (Xu-Monette et al., Front Immunol, 2017).

Cluster of differentiation 137 or CD137 (also known as 4-1BB or TNFRS9) is a costimulatory immune receptor and a member of the tumor necrosis factor receptor (TNFR) superfamily. It is mainly expressed on activated CD4+ and CD8+ T cells, activated B cells and natural killer (NK) cells, but can also be found on resting monocytes and dendritic cells (Li and Liu, Clin Pharmacol, 2013) or endothelial cells (Snell et al., Immunol Rev, 2011). CD137 plays an important role in regulating the immune response and is thus a target for cancer immunotherapy. CD137 ligand (CD137L) is the only known natural CD137 ligand and is constitutively expressed on several types of antigen presenting cells, such as activated B cells, monocytes and splenic dendritic cells, and can be induced on T lymphocytes.

CD137L is a trimeric protein that exists as a membrane-bound form and a soluble variant. However, the ability of soluble CD137L to activate CD137, for example on CD137-expressing lymphocytes, is limited and large concentrations are required to produce an effect (Wyzgol et al., J Immunol, 2009). The natural pathway for CD137 activation is the interaction of a CD137-positive cell with a CD137L-positive cell. It is believed that CD137 activation is then induced by clustering via CD137L on the opposing cell, leading to signaling by TRAF1, 2 and 3 (Yao et al., Nat Rev Drug Discov, 2013, Snell et al., Immunol

- 1 046634

Rev, 2011) and further concomitant downstream effects in the CD137-positive T cell. In the case of T cells activated by recognition of their respective cognate targets, the effects caused by CD137 costimulation are further enhanced activation, increased survival and proliferation, production of proinflammatory cytokines, and improved killing capacity.

The benefits of CD137 costimulation for killing cancer cells have been demonstrated in a number of in vivo models. For example, forced expression of CD137L on a tumor results in tumor rejection (Melero et al., Eur J Immunol, 1998). Similarly, forced expression of anti-CD137 ScFv on tumors results in CD4+ T cell and NK cell dependent tumor clearance (Yang et al., Cancer Res. 2007, Zhang et al., Mol Cancer Ther. 2006, Ye et al. al., Nat Med, 2002). Systemically administered anti-CD137 antibody has been shown to slow tumor growth (Martinet et al., Gene Ther, 2002).

It has been shown that CD137 is an excellent marker for natural tumor-reactive T cells in human tumors (Ye et al., Clin Cancer Res, 2014), and that anti-CD137 antibodies can be used to improve the expansion and activity of CD8' tumor-infiltrating lymphocytes melanoma for use in adoptive T-cell therapy (Chacon et al., PLoS One, 2013).

Preclinical demonstration of the potential therapeutic efficacy of CD137 costimulation has stimulated the development of therapeutic antibodies targeting CD137, including BMS663513 (described in US Pat. No. 7,288,638) and PF-05082566 (Fisher et al., Cancer Immunol Immunother, 2012).

The present disclosure provides, among other things, new approaches for simultaneously binding CD137 and PD-L1 using one or more fusion proteins having CD137 binding specificity and PD-L1 binding specificity.

II. Definitions

The following list defines the terms, phrases and abbreviations used herein. All terms listed and defined herein cover all grammatical forms.

As used herein, unless otherwise stated, CD137 means human CD137 (huCD137). Human CD137 means the full-length protein as defined by UniProt Q07011, a fragment or variant thereof. CD137 is also known as 4-1BB, tumor necrosis factor receptor superfamily member 9 (TNFRSF9), and inducible lymphocyte activation (ILA). In some particular embodiments, CD137 from a non-human species is used, for example, cynomolgus CD137 and murine CD137.

As used herein, unless otherwise stated, programmed cell death ligand 1 or PD-L1 means human PD-L1 (huPD-L1). Human PD-L1 means the full-length protein defined by UniProt Q9NZQ7, a fragment or variant thereof. Human PD-L1 is encoded by the CD274 gene. PD-L1 is also known as cluster of differentiation 274 (CD274) or B7 homolog type 1 (B7-H1). In some particular embodiments, PD-L1 from a non-human species is used, for example, cynomolgus PD-L1 and murine PD-L1.

As used herein, binding affinity describes the ability of a biomolecule (eg, a polypeptide or protein) of the invention (eg, a lipocalin mutein, antibody, fusion protein, or any other peptide or protein) to bind a specific target and form a complex. Binding affinity is measured using a number of methods known to those skilled in the art, including, but not limited to, fluorescence titration, enzyme-linked immunosorbent assay (ELISA) assays, including direct and competitive ELISA, calorimetric methods, such as isothermal titration calorimetry (ITC), and surface plasmon resonance (SPR). These methods are well known in the art, and some examples of such methods are further described herein. Binding affinity is thus expressed by the value of the dissociation constant (KD), the half-maximal effective concentration (EC ₅₀ ) or the half-maximal inhibition concentration (IC ₅₀ ) measured using such methods. A lower _KD , EC ₅₀ or IC ₅₀ value reflects better (higher) binding capacity (affinity). Accordingly, the binding affinity of two biomolecules to a specific target can be measured and compared. When comparing the binding affinities of two biomolecules to a particular target, the term approximately the same, substantially the same, or substantially similar means that one biomolecule has a binding affinity, expressed by a _KD , EC ₅₀ or IC ₅₀ value, identical or similar to another molecule within variability in experimental results measuring binding affinity. The variability in experimental results for measuring binding affinity depends on the specific method used and is known to those skilled in the art.

As used herein, the term essentially may also refer to the qualitative condition of demonstrating the full or nearly full extent or magnitude of a characteristic or property of interest. It will be clear to a specialist in the field of biology that biological and chemical

- 2 046634 logical phenomena rarely, if ever, reach completion and/or proceed to completion, or achieve or avoid an absolute result. Thus, the term is essentially used herein to refer to the potential lack of completeness inherent in many biological and chemical phenomena.

As used herein, the term detect, detect, detectable, or perform detection is understood at both a quantitative and qualitative level, or a combination of both. Thus, it includes quantitative, semi-quantitative and qualitative measurements carried out on the biomolecule according to the invention.

As used herein, detectable affinity generally means that the binding capacity between a biomolecule and its target, expressed as a _KD , EC50 or _IC50 value, is no more than about 10 ^-5 M, or lower. Binding affinity, expressed by a K _D , EC ₅₀ or IC ₅₀ value greater than 10 ^-5 M, is generally no longer measurable by conventional methods such as ELISA and SPR and is therefore of secondary importance.

It should be noted that the formation of a complex between the biomolecule of the invention and its target depends on many different factors, such as the concentration of the relevant target, the presence of competitors, the pH and ionic strength of the buffer system used, the experimental method used to determine binding affinity (e.g., fluorescence titration, competitive ELISA (also called competitive ELISA) and surface plasmon resonance), and even a mathematical algorithm used to evaluate experimental data. Therefore, one skilled in the art will appreciate that binding affinity, expressed as K _D , EC ₅₀ or IC ₅₀ , may vary within a certain experimental range, depending on the method used and the experimental conditions. This means that there may be a slight variation in the measured KD, EC50 or IC50 values or range of acceptable values depending, for example, on whether such values were determined by ELISA (including direct or competitive ELISA), SPR, or other method.

As used herein, the terms specific to, binding specific, bind specifically, or binding specificity refer to the ability of a biomolecule to distinguish between a desired target (e.g., CD137 and PD-L1) and one or more reference targets (e.g., cellular neutrophil gelatinase-associated lipocalin receptor). . It is understood that such specificity is not an absolute but a relative property and can be determined, for example, by SPR, Western blotting, ELISA, fluorescence-activated cell sorting (FACS), radioimmunoassay (RIA), electrochemiluminescence (ECL), immunoradiometric analysis ( IRMA), immunohistochemistry (IHC) and peptide scanning.

When used herein in the context of a fusion protein of the present invention that binds to CD137 and PD-L1, the term specific binding, specific binding, or binding specificity means that the fusion protein binds to, reacts with, or is directed against CD137 and PD-L1. L1 as described herein, but does not substantially bind another protein. The term other protein includes any proteins that are not CD137 or PD-L1, or proteins related or homologous to CD137 or PD-L1. However, CD137 or PD-L1 of non-human species, and fragments and/or variants of CD137 or PD-L1 are not excluded by the term other protein. The term substantially does not bind means that the fusion proteins of the present invention bind another protein with lower binding affinity than CD137 and/or PD-L1, i.e. exhibit cross-reactivity of less than 30%, preferably 20%, more preferably 10%, particularly preferably less than 9, 8, 7, 6 or 5%. The ability of a fusion protein to specifically respond as defined hereinabove can be readily tested, among other things, by comparing the response of the fusion protein of the present invention with CD137 and/or PD-L1 and the response of said fusion protein with other protein(s).

As used herein, the term lipocalin refers to a monomeric protein of approximately 18-20 kDa having a supersecondary structural region of a cylindrical β-pleated sheet containing multiple β chains (preferably eight β chains, designated A-H) linked in pairs by a plurality of loops (preferably four) at one end, thus containing the ligand-binding pocket and forming the entrance to the ligand-binding pocket. Preferably, the loops containing the ligand binding pocket used in the present invention are loops connecting the open ends of β-strands A and B, C and D, E and F, and G and H, and are designated loops AB, CD, EF and G.H. It is well established that the diversity of these loops within the otherwise rigid lipocalin scaffold gives rise to different binding mechanisms among members of the lipocalin family, each capable of accommodating targets of different sizes, shapes and chemical natures (reviewed, for example, in Skerra, Biochim Biophys Acta , 2000, Flower et al., Biochim Biophys Acta, 2000, Flower, Biochem J, 1996). The lipocalin family of proteins is reported to have naturally evolved to bind a wide range of ligands, characterized by unusually low levels of overall sequence conservation.

- 3,046,634 features (often with less than 20% sequence identity), while maintaining a highly conserved overall folding pattern. The correspondence between positions in different lipocalins is also well known to those skilled in the art (see, for example, US Pat. No. 7,250,297). Proteins falling within the definition of lipocalins as used herein include, but are not limited to, human lipocalins including tear lipocalin (Tlc, Lcn1), lipocalin-2 (Lcn2) or neutrophil gelatinase-associated lipocalin (NGAL), apolipoprotein D ( ApoD), apolipoprotein M, αι-acid glycoprotein 1, a ₁ -acid glycoprotein 2, a ₁ -microglobulin, complement component 8γ, retinol binding protein (RBP), epididymal retinoic acid binding protein, glycodelin, odorant binding protein IIa , odorant binding protein IIb, lipocalin-15 (Lcn15) and prostaglandin D synthase.

As used herein, unless otherwise specified, tear lipocalin refers to human tear lipocalin (hTIc) and also refers to mature human tear lipocalin. The term mature, when used to characterize a protein, means a protein substantially free of signal peptide. Mature hTIc in the present invention refers to the mature form of human tear lipocalin that does not contain a signal peptide. Mature hTIc is described by residues 19-176 of the sequence deposited in the SWISS-PROT data bank under accession number P31025, and its amino acid sequence is shown in SEQ ID NO: 1.

As used herein, lipocalin-2 or neutrophil gelatinase-associated lipocalin refers to human lipocalin-2 (hLcn2) or human neutrophil gelatinase-associated lipocalin (hNGAL), and also refers to mature human lipocalin-2 or mature human lipocalin-associated lipocalin with neutrophil gelatinase. The term mature, when used to characterize a protein, means a protein substantially free of signal peptide. Mature hNGAL in the present invention refers to the mature form of human lipocalin associated with neutrophil gelatinase, which does not contain a signal peptide. Mature hNGAL is described by residues 21-198 of the sequence deposited in the SWISS-PROT data bank under accession number P80188, and its amino acid sequence is shown in SEQ ID NO: 2.

As used herein, a native sequence refers to a protein or polypeptide having a naturally occurring or wild-type sequence, regardless of how it is produced. Such a native sequence protein or polypeptide may be isolated from natural sources or may be produced by other means, such as recombinant or synthetic methods.

Native sequence lipocalin refers to lipocalin having the same amino acid sequence as the corresponding naturally occurring polypeptide. Thus, a native sequence lipocalin may have the amino acid sequence of the corresponding natural lipocalin (wild type) from any organism, particularly a mammal. The term native sequence, as applied to lipocalin, particularly encompasses naturally occurring truncated or secreted forms of lipocalin, naturally occurring variant forms such as alternatively spliced forms, and naturally occurring allelic variants of lipocalin. The terms native sequence lipocalin and wild-type lipocalin are used interchangeably herein.

As used herein, a mutein, a mutated fragment (whether a protein or nucleic acid fragment) or mutant refers to the exchange, deletion or insertion of one or more amino acids or nucleotides compared to the naturally occurring protein or nucleic acid (wild type). The term also includes mutein fragments as described herein. The present invention explicitly covers lipocalin muteins, as described herein, having a supersecondary structural region of a cylindrical β-pleated sheet containing eight β-strands connected in pairs by four loops at one end, thereby containing a ligand-binding pocket and forming an entrance into a ligand-binding pocket where at least one amino acid of each of at least three of the four loops has been mutated from the native sequence lipocalin. The lipocalin muteins of the present invention preferably have CD137 binding function as described herein.

As used herein, the term fragment in relation to the lipocalin muteins of the invention refers to proteins or polypeptides derived from full-length mature hTIc or hNGAL, or lipocalin muteins truncated at the N-terminus and/or the C-terminus, i.e. lacking at least one of the N-terminal and/or C-terminal amino acids. Such fragments may comprise at least 10 or more, for example, 20 or 30 or more consecutive amino acids of the primary sequence of mature hTIc or hNGAL or the lipocalin mutein from which they are derived, and are typically detectable in a mature hTIc or hNGAL immunoassay. In such

- 4 046634 fragment may be missing up to 2, up to 3, up to 4, up to 5, up to 10, up to 15, up to 20, up to 25 or up to 30 (including all numbers in between) N-terminal and/or C-terminal amino acids. As an illustrative example, such a fragment may lack one, two, three or four N-terminal (His-His-Leu-Leu) and/or one or two C-terminal amino acids (Ser-Asp) of mature hTIc. The fragment is preferably intended to be a functional fragment of the mature hTIc or hNGAL or lipocalin mutein from which it is derived, meaning that it preferably retains the binding specificity, preferably to CD137, of the mature hTIc/hNGAL or lipocalin mutein from which it is derived. As an illustrative example, such a functional fragment may contain at least amino acid positions 5-153, 5-150, 9-148, 12-140, 20-135 or 26-133 corresponding to the linear polypeptide sequence of mature hTIc. As yet another illustrative example, such a functional fragment may comprise at least amino acid positions 13-157, 15-150, 18-141, 20-134, 25-134, or 28-134, corresponding to the linear polypeptide sequence of mature hNGAL.

A fragment, in relation to the corresponding target CD137 or PD-L1 fusion protein of the invention, refers to N-terminally and/or C-terminally truncated CD137 or PD-L1, or CD137 or PD-L1 protein domains. CD137 fragments or PD-L1 fragments, as described herein, retain the ability of full-length CD137 or PD-L1 to be recognized and/or bound by the fusion protein of the invention. As an illustrative example, the fragment may be the extracellular domain of CD137 or PD-L1. As an illustrative example, such an extracellular domain may comprise the amino acids of CD137 extracellular subdomains, such as the individual or combined amino acid sequences of domain 1 (residues 24-45 of UniProt Q07011), domain 2 (residues 46-86), domain 3 (87-118), and domain 4 (residues 119-159). As another illustrative example, such an extracellular domain may comprise amino acid residues 19-238 of UniProt Q9NZQ7.

As used herein, the term variant refers to derivatives of a protein or polypeptide that include mutations, for example, by substitutions, deletions, insertions and/or chemical modifications of an amino acid sequence or nucleotide sequence. In some embodiments, such mutations and/or chemical modifications do not reduce the functionality of the protein or peptide. Such substitutions may be conservative, i.e. an amino acid residue may be replaced by a chemically similar amino acid residue. Examples of conservative substitutions are substitutions among members of the following groups: 1) alanine, serine, threonine and valine; 2) aspartic acid, glutamic acid, glutamine, asparagine and histidine; 3) arginine, lysine, glutamine, asparagine and histidine; 4) isoleucine, leucine, methionine, valine, alanine, phenylalanine, threonine and proline; and 5) isoleucine, leucine, methionine, phenylalanine, tyrosine and tryptophan. Such variants include proteins or polypeptides in which one or more amino acids have been replaced with their corresponding D-stereoisomers or amino acids other than the 20 naturally occurring amino acids, such as, for example, ornithine, hydroxyproline, citrulline, homoserine, hydroxylysine, norvaline. Such variants also include, for example, proteins or polypeptides in which one or more amino acid residues at the N- and/or C-terminus are added or deleted. Typically, the variant has at least about 50%, 60%, 70%, 75%, 80%, 85%, 90%, 92%, 95%, or at least about 98% amino acid sequence identity to a protein or polypeptide with native sequence. The variant preferably retains biological activity, such as binding to the same target protein or polypeptide from which it is derived.

The term variant as used herein in relation to the corresponding CD137 ligand or PD-L1 fusion protein of the invention refers to CD137 or PD-L1 or a fragment thereof, respectively, having one or more, for example, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 40, 50, 60, 70, 80 or more amino acid substitutions, deletions and/or insertions compared to the native sequence of CD137 or PD-L1 (CD137 or wild-type PD-L1), for example, CD137 deposited as UniProt Q07011, or PD-L1 deposited as UniProt Q9NZQ7, as described herein. The CD137 variant or PD-L1 variant, respectively, preferably has at least 50%, 60%, 70%, 80%, 85%, 90%, or 95% amino acid sequence identity to wild-type CD137 or PD-L1. The CD137 variant or PD-L1 variant, as described herein, retains the ability to bind the CD137-PD-L1 specific fusion proteins disclosed in the present invention.

The term variant as used herein in relation to a lipocalin mutein refers to a lipocalin mutein or fragment thereof according to the invention, wherein the sequence has mutations, including substitutions, deletions and insertions, and/or chemical modifications. The lipocalin mutein variant described herein retains the biological activity, eg, binding to CD137, of the lipocalin mutein from which it is derived. Typically, a lipocaine mutein variant has at least about 50%, 60%, 70%, 75%, 80%, 85%, 90%, 92%, 95%, 98% amino acid sequence identity with the lipocaine mutein from which it is derived. received.

As used herein, the term mutagenesis refers to the introduction of mutations into a polynucleotide or amino acid sequence. Mutations are preferably introduced in such experiments

- 5 046634 experimental conditions that the natural amino acid at a given position in the protein or polypeptide sequence can be changed, for example, replaced by at least one amino acid. The term mutagenesis also includes (additional) modification of the length of sequence segments by deletion or insertion of one or more amino acids. Thus, within the scope of the invention, for example, one amino acid at a selected position in the sequence may be replaced by a stretch of three amino acids, resulting in the addition of two amino acid residues compared to the length of the corresponding segment of the native amino acid sequence of the protein or polypeptide. Such an insertion or deletion can be introduced independently of each other into any of the sequence segments that can be mutagenized in the invention. In one illustrative embodiment of the present invention, the insert may be introduced into a segment of the amino acid sequence corresponding to the AB loop of the native sequence lipocalin (see International Patent Application Publication WO 2005/019256, incorporated herein by reference in its entirety).

As used herein, the term random mutagenesis means that there is no predetermined mutation (amino acid change) at a particular sequence position, but that at least two amino acids are likely to be included at a predetermined sequence position during mutagenesis.

As used herein, the term sequence identity or identity means a property of sequences that is a measure of their similarity or relatedness. The term sequence identity or identity as used herein means the percentage of pairwise identical residues - following a (homologous) alignment of the sequence of a protein or polypeptide of the invention with the sequence in question - relative to the number of residues in the longer of the two sequences. Sequence identity is measured by dividing the number of identical amino acid residues by the total number of residues and multiplying the result by 100.

As used herein, the term sequence homology or homology has its usual meaning, and a homologous amino acid includes identical amino acids, as well as amino acids that are considered conservative substitutions, at equivalent positions in the linear amino acid sequence of a protein or polypeptide according to the invention (for example, any fusion proteins or muteins lipocalin according to the invention).

One skilled in the art will be familiar with available computer programs such as BLAST (Altschul et al., Nucleic Acids Res, 1997), BLAST2 (Altschul et al., J Mol Biol, 1990) and Smith-Waterman (Smith and Waterman, J Mol Biol, 1981). , to determine sequence homology or sequence identity using standard parameters. The percentage of sequence homology or sequence identity can, for example, be determined herein using the BLASTP program, version 2.2.5 (November 16, 2002; (Altschul et al., Nucleic Acids Res, 1997). In this embodiment, the percentage of homology is based on the alignment of entire protein or polypeptide sequences (matrix: BLOSUM 62; gap weight: 11.1; threshold set to 10 ^-3 ), including propeptide sequences, preferably using the wild-type protein scaffold as a reference in pairwise comparisons. It is calculated. as the percentage of the number of positive (homologous amino acids) reported as a result in the BLASTP program output, divided by the total number of amino acids selected by the program for alignment.

In particular, to determine whether an amino acid residue in the amino acid sequence of a lipocalin mutein differs from wild-type lipocalin corresponding to a particular position in the amino acid sequence of wild-type lipocalin, one skilled in the art may use tools and methods well known in the art, such as alignments, manual or using computer programs such as BLAST 2.0, which stands for Basic Local Alignment Search Tool, or ClustalW, or any other suitable program that is suitable for generating sequence alignments. Accordingly, a wild-type lipocalin sequence may serve as a target sequence or reference sequence, while an amino acid sequence of a lipocalin mutein other than the wild-type lipocalin described herein serves as a target sequence. The terms wild-type sequence, reference sequence, and subject sequence are used interchangeably herein. The preferred wild-type lipocalin sequence is the hTLc sequence set forth in SEQ ID NO: 1 or hNGAL set forth in SEQ ID NO: 2.

Gaps are gaps in the alignment that result from additions or deletions of amino acids. Thus, two copies of the same sequence have 100% identity, but sequences that are not as highly conserved and have deletions, additions, or substitutions may have a lower degree of identity.

As used herein, the term position means the position of either an amino acid in the amino acid sequence shown herein or the position of a nucleotide in

- 6046634 nucleic acid sequence shown herein. It should be understood that when the term corresponds or corresponds, as used herein in the context of positions in the amino acid sequence of one or more lipocalin muteins, the corresponding position is determined not only by the number of preceding nucleotides or amino acids. Accordingly, the absolute position of a given amino acid according to the invention may vary relative to the corresponding position due to deletions or additions of amino acids elsewhere in lipocalin (mutant or wild type). Likewise, the absolute position of a given nucleotide according to the present invention may differ from the corresponding position due to deletions or additional nucleotides elsewhere in the 5' untranslated region (UTR) of a lipocalin mutein or wild-type lipocalin, including the promoter and/or any other regulatory sequences or gene (including exons and introns).

The corresponding position according to the invention may be a sequence position that coincides with the sequence position to which it corresponds in a pairwise or multiple sequence alignment according to the present invention. It is preferably understood that for a corresponding position according to the invention, the absolute positions of the nucleotides or amino acids may differ from adjacent nucleotides or amino acids, but said adjacent nucleotides or amino acids that may have been replaced, deleted, or added may be covered by one or more corresponding positions.

In addition, for a corresponding position in a lipocalin mutein based on a reference sequence according to the invention, it is preferably understood that the nucleotide or amino acid positions of the lipocalin mutein may structurally correspond to positions elsewhere in the reference lipocalin (wild type lipocalin) or other lipocalin mutein, even though they may differ in absolute numbers of positions, which is clear to specialists taking into account the highly conservative general nature of folding inherent in lipocalins.

Used interchangeably herein, the terms conjugate, conjugation, hybridize, fusion, or linked refer to the joining together of two or more subunits through all forms of covalent or non-covalent binding, including, but not limited to, genetic fusion, chemical conjugation, linking by a linker or cross-linking agent, and non-covalent association.

As used herein, the term fusion polypeptide or fusion protein refers to a polypeptide or protein containing two or more subunits. In some embodiments, a fusion protein as described herein comprises two or more subunits, at least one of which is capable of specifically binding to CD137 and the other subunit is capable of specifically binding to PD-L1. Within the fusion protein, these subunits can be linked covalently or non-covalently. Preferably, the fusion protein is a translational fusion of two or more subunits. A translational fusion can be produced by genetically modifying the coding sequence for one subunit in frame with the coding sequence of another subunit. Both subunits can be separated by a nucleotide sequence encoding a linker. However, the fusion protein subunits of the present invention can also be linked through chemical conjugation. The subunits forming a fusion protein are typically linked to each other by linking the C-terminus of one subunit to the N-terminus of another subunit, or the C-terminus of one subunit to the C-terminus of another subunit, or the N-terminus of one subunit to the N-terminus another subunit, or the N-terminus of one subunit with the C-terminus of another subunit. The fusion protein subunits may be linked in any order and may include more than one of any of the constituent subunits. If one or more subunits are part of a protein (complex) that consists of more than one polypeptide chain, the term fusion protein may also refer to a protein containing the fusion sequences and all other polypeptide chains (chain) of the protein (complex). As an illustrative example, when a full-length immunoglobulin is fused to a lipocalin mutein via an immunoglobulin heavy or light chain, the term fusion protein may refer to a single polypeptide chain comprising a lipocalin mutein and an immunoglobulin heavy or light chain. The term fusion protein can also refer to the entire immunoglobulin (both light and heavy chains) and mutein lipocalin fused to one or both of its heavy and/or light chains.

As used herein, the term subunit of a fusion protein described herein refers to a single protein or single polypeptide chain that can itself form a stable folded structure and have a unique function of providing a binding motif to a target. In some embodiments, a preferred subunit of the invention is a lipocalin mutein. In some other embodiments, a preferred subunit of the invention is a full-length immunoglobulin or an antigen-binding domain thereof.

A linker, which may be contained in a fusion protein of the present invention, joins two or more subunits of the fusion protein as described herein. Communication may be

- 7 046634 covalent or non-covalent. A preferred covalent bond is formed through a peptide bond, such as a peptide bond between amino acids. A preferred linker is a peptide linker. Accordingly, in a preferred embodiment, said linker contains one or more amino acids, for example, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or more amino acids. Preferred peptide linkers are described herein, including glycine-serine (GS) linkers, glycosylated GS linkers, and proline-alanine-serine polymer (PAS) linkers. In some preferred embodiments, a GS linker of (G ₄ S) ₃ is used to join together the subunits of the fusion protein, as described in SEQ ID NO: 13. Other preferred linkers include chemical linkers.

As used herein, the term albumin includes all mammalian albumins, such as human serum albumin, or bovine serum albumin, or rat serum albumin.

As used herein, the term organic molecule or small organic molecule means an organic molecule containing at least two carbon atoms, but preferably no more than 7 or 12 rotatable carbon bonds, having a molecular weight in the range of 100 to 2000 daltons, preferably from 100 to 1000 dalton, and optionally including one or two metal atoms.

A sample is defined as a biological sample taken from any subject. Biological samples include, but are not limited to, blood, serum, urine, feces, semen, or tissue, including tumor tissue.

The subject is a vertebrate animal, preferably a mammal, more preferably a human. The term mammal is used herein to mean any animal classified as a mammal, including, but not limited to, humans, domestic animals, farm animals, zoo animals, sporting animals, or pets such as sheep, dogs, horses. , cats, cows, rats, pigs, monkeys such as cynomolgus macaques. Preferably, the mammal as used herein is a human.

An effective amount is an amount sufficient to produce beneficial or desired results. An effective amount may be administered in one or more separate administrations or doses.

As used herein, an antibody includes whole antibodies or any antigen-binding fragment (ie, antigen-binding portion) or one chain thereof. A whole antibody refers to a glycoprotein containing at least two heavy chains (HC) and two light chains (LC) linked by disulfide bonds. Each heavy chain consists of a heavy chain variable domain (VH or HCVR) and a heavy chain constant region (CH). The heavy chain constant region consists of three domains: C _H1 , C'ii2 and C _H3 . Each light chain consists of a light chain variable domain (V _L or LCVR) and a light chain constant region ( _CL ). The light chain constant region consists of a single domain, C _L . The V _H and V _L regions can be further subdivided into regions of hypervariability called complementarity determining regions (CDRs), interspersed with more conserved regions called framework regions (FR). Each VH and V _L consists of three CDRs and four FRs, arranged in the following order from amino terminus to carboxy terminus: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The heavy and light chain variable regions contain a binding domain that interacts with an antigen (eg, PD-L1). Antibody constant regions may optionally mediate binding of the immunoglobulin to host tissues or factors, including various cells of the immune system (eg, effector cells) and the first component (C1q) of the classical complement system.

As used herein, an antigen-binding antibody fragment refers to one or more antibody fragments that retain the ability to specifically bind an antigen (eg, PD-L1). It has been shown that the antigen-binding function of an antibody can be performed by fragments of a full-length antibody. Examples of binding fragments encompassed by the term antigen-binding antibody fragment include (i) a Fab fragment consisting of the _VH , _VL , _CL, and _CH1 domains; (ii) an F(ab') ₂ fragment containing two Fab fragments linked by a disulfide bridge in the hinge region; (iii) a Fab' fragment consisting of the V _H , V _L , C _L and C _H1 domains and the region between the C _H1 and C _H2 domains; (iv) Fd fragment consisting of VH and C _H1 domains; (v) a single chain Fv fragment consisting of the VH and V _L domains of one arm of the antibody, (vi) a dAb fragment (Ward et al., Nature, 1989) consisting of a VH domain; and (vii) a dedicated complementarity determining region (CDR) or a combination of two or more dedicated CDRs, which may optionally be connected by a synthetic linker; (VIII) a diabody containing _VH and _VL joined in the same polypeptide chain by a short linker (see, for example, patent documents EP 404,097; WO 93/11161; and Holliger et al., Proc Natl Acad Sci USA, 1993); (ix) a domain antibody fragment containing only _VH or _VL , where in some cases two or more _VH regions are covalently linked.

Antibodies can be polyclonal or monoclonal;

- 8 046634 xenogeneic, allogeneic or syngeneic; or modified forms thereof (eg, humanized, chimeric or polyspecific). Antibodies can also be entirely human.

As used herein, framework or FR refers to variable domain residues other than hypervariable region residues (CDR).

The crystallizable fragment region or Fc region refers to the C-terminal region of an immunoglobulin heavy chain, including native sequence Fc regions and variant Fc regions. Although the boundaries of the Fc region of the immunoglobulin heavy chain can vary, the Fc region of the human IgG heavy chain is generally defined as lying from the amino acid residue at position Cys226, or Pro230, to its carboxyl terminus, where the numbering corresponds to the EU Kabat index (Johnson and Wu, Nucleic Acids Res, 2000). The C-terminal lysine (residue 447 according to the EU Kabat index) of the Fc region can be removed, for example, during production or purification of the antibody or by genetic modification of the nucleic acid encoding the heavy chain of the antibody. Accordingly, an intact antibody composition may include antibody populations with K447 residues completely deleted, antibody populations in which K447 residues have not been deleted, and antibody populations containing a mixture of antibodies with and without K447 residue. Suitable native sequence Fc regions for use in antibodies of the invention include human IgG1, IgG2 (IgG2A, IgG2B), IgG3 and IgG4.

Fc receptor or FcR refers to a receptor that binds to the Fc region of an antibody.

As used herein, an isolated antibody refers to an antibody substantially free from its natural environment. For example, an isolated antibody contains substantially no cellular material or other proteins from the cellular or tissue source from which it is derived. An isolated antibody also refers to an antibody that is substantially free of other antibodies having a different antigenic specificity. Here, the isolated antibody that specifically binds PD-L1 is substantially free of antibodies that specifically bind antigens other than PD-L1. However, an isolated antibody that specifically binds PD-L1 may have cross-reactivity with other antigens, such as PD-L1 molecules from other species.

As used herein, a monoclonal antibody refers to a preparation of antibody molecules with one type of molecule in the composition. A monoclonal antibody composition exhibits a single binding specificity and affinity for a particular epitope.

As used herein, a humanized antibody refers to an antibody consisting of a CDR of an antibody derived from a non-human mammal and an FR region and a constant region of a human antibody or derived from a human antibody. In some embodiments, a humanized antibody comprises a variable domain that has a variable region amino acid sequence that is generally closer to human than to other species, as assessed using the DomainGapAlign tool from the Immunogenetics Information System (IMGT), as described by Ehrenmann et al . (2010). In some embodiments, a humanized antibody may be used as an effective component of a therapeutic agent due to reduced antigenicity. As used herein, the term therapeutic agent or therapeutically active agent refers to an agent that is therapeutically useful. A therapeutic agent can be any agent for the prevention, amelioration or treatment of a disease, physiological condition, symptom, or for the evaluation or diagnosis thereof.

As used herein, a human antibody includes antibodies having variable regions in which both the framework region and the CDR regions are derived from human germline immunoglobulin sequences. In addition, if the antibody contains a constant region, the constant region is also derived from human germline immunoglobulin sequences. Human antibodies of the invention may include amino acid residues not encoded by human germline immunoglobulin sequences (eg, mutations introduced by random or site-specific mutagenesis in vitro or somatic mutation in vivo). However, the term human antibody as used herein does not imply antibodies in which CDR sequences derived from the germ line of another mammalian species, such as a mouse, have been grafted onto human framework sequences.

III. Description of graphic materials

Fig. 1: is an overview of the design of exemplary fusion proteins described in this invention that are bispecific for the targets CD137 and PD-L1. Exemplary fusion proteins were generated from an antibody specific for PD-L1 (e.g., antibodies whose heavy chains are represented by SEQ ID NO: 86, or contain the heavy chain variable domain of SEQ ID NO: 77, or contain the CDR sequences GFSLSNYD (HCDR1, SEQ ID NO: 60), IWTGGAT (HCDR2; SEQ ID NO: 61) and VRDSNYRYDEPFTY (HCDR3; SEQ ID NO: 62), and the light chains represented by SEQ ID NO: 87, or contain the heavy chain variable domain SEQ ID NO: 82, or contain the CDR sequences QSIGTN (LCDR1, SEQ ID NO: 63), YAS (LCDR2), and QQSNWPYT (LCDR3; SEQ ID NO: 64)), and one or more CD137-specific lipocalin muteins (e.g., lipocalin mutein

- 9 046634 line with SEQ ID NO: 42). One or more lipocalin muteins have been genetically fused to the C and/or N terminus of the heavy chain or light chain of a PD-L1 specific antibody, as shown in FIG. 1A-1I, resulting in fusion proteins such as SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NO: 94 and 87 and SEQ ID NOs: 90 and 91. The resulting fusion proteins may be CD137 bivalent (eg, as demonstrated in FIGS. 1A-1D), or CD137 tetravalent (eg, as demonstrated in FIGS. 1E-1H), or have an even higher valence to CD137 (eg, as demonstrated in Fig. 1I). Additional monospecific fusion proteins were made by fusing one or more CD137-specific lipocalin muteins (eg, as demonstrated in Figures 1J-1K) to the C-terminus of the Fc region of an antibody prepared as described herein via a peptide linker. The resulting monospecific fusion proteins are shown, for example, in SEQ ID NO: 88 and SEQ ID NO: 89.

Fig. 2: Shows the results of ELISA experiments in which binding to PD-L1 or CD137 of exemplary fusion proteins was determined as described in Example 4. PD-L1 or CD137 (with a C-terminal His or Fc tag) was applied to a microtiter plate and titrated to test agents starting at the highest concentration of 100 nM. The bound agents studied were detected using anti-human IgG Fc-HRP or anti-NGAL-HRP antibodies, respectively. The data were fitted to a 1:1 binding model with EC50 and maximum signal as free parameters, and the slope was set to one. The obtained EC50 values are presented in table. 4.

Fig. 3: Illustrates the results of an ELISA experiment in which the ability of exemplary fusion proteins to simultaneously bind both targets, PD-L1 and CD137, as described in Example 5 was determined. Recombinant huPD-L1-His or huCD137-His was applied to a microtiter plate followed by titration of the fusions proteins, starting with the highest concentration of 100 nM. A constant concentration of biotinylated huCD137-His or biotinylated huPD-L1-His, respectively, was then added and detected using ExtrAvidin peroxidase.

Fig. 4: Shows results of flow cytometry assessment of target binding of fusion proteins using Flp-In-CHO cells expressing human or cynomolgus CD137 (Figures 4A-4B) and human or cynomolgus PD-L1 (Figure 4C -4D) as described in Example 6. No binding was observed using mock-transfected Flp-InCHO cells (Fig. 4E). Geometric mean fluorescence intensities were used to calculate EC50 values using nonlinear regression (overall lower asymptote, slope =1). EC ₅₀ values are presented in table. 6.

Fig. 5: Demonstrates binding of fusion proteins to PD-L1 positive tumor cells assessed by flow cytometry by incubating PKO cells and fusion proteins as described in Example 7.

Fig. 6: Shows examples of a multi-binding SPR experiment designed to examine whether binding of CD137L to CD137 impairs the interactions of the fusion protein with CD137 as described in Example 8. This is assessed by creating a complex of huCD137 (C-terminal Fc fusion) and huCD137L (with a C-terminal His tag) on the SPR sensor chip and testing whether the fusion proteins could still bind the huCD137 and CD137L complex. For comparison, huCD137 in the absence of huCD137L was also incubated with the test fusion proteins. The SPR curve for binding of the corresponding fusion protein to huCD137 only is indicated by a solid arrow. The SPR curve for binding of the corresponding huCD137 fusion protein saturated with huCD137L is indicated by a dotted arrow. Blanks without fusion proteins were used as controls. The experiment shows that all fusion proteins tested are able to bind CD137 in the presence of CD137L.

Fig. 7: demonstrates that the fusion proteins compete with PD-L1 for binding to PD-1, as illustrated in competitive ELISA studies as described in Example 9. A constant concentration of huPD-1-His was applied to a microtiter plate followed by the addition of the test mixture molecules at various concentrations and the huPD-L1-Fc indicator at a fixed concentration. The bound tracer was detected using an HRP-labeled anti-Fc IgG antibody. Dose-dependent inhibition of huPD-L1-Fc binding to PD-1 by bispecific fusion proteins to CD137 and PD-L1 or specific antibodies to PD-L1 was observed.

Fig. 8: The potential of exemplary fusion proteins to costimulate T cell activation in a PD-L1 target-dependent manner was assessed using a CD137 bioassay. NFkBluc2/CD137 Jurkat cells were cocultured with the PDL1-expressing RKO tumor cell line in the presence of varying concentrations of fusion proteins or controls. After 4 h, luciferase assay reagent was added and luminescent signals were measured. Four-parameter logistic curve analysis was performed using GraphPad Prism® to calculate EC ₅₀ values (see Table 9). The fusion proteins co-stimulate T cell activation only in the presence of PD-L1 (Figures 8A and 8C) but not in the absence of PD-L1 (Figures 8B and 8D). In contrast, the reference mAb to CD137 (SEQ ID

- 10 046634

NO: 28 and 29) show similar activation in the presence and absence of PD-L1 positive RKO cells.

Fig. 9: Shows the results of an illustrative experiment in which the ability of selected fusion proteins to induce T cell activation was examined. Anti-PD-L1 antibodies, including the corresponding anti-PD-L1 building block antibody, CD137-binding lipocalin muteins as Fc fusions, and the reference anti-CD137 antibody, were tested individually and in combination as an anti-PD-L1/anti cocktail. -CD137. In the experiment, human peripheral blood mononuclear cells (PBMCs) were incubated with fusion proteins, antibodies, lipocalin mutein Fc fusions, cocktails, or control in the presence of 1 ng/ml staphylococcal enterotoxin B (SEB). Levels of secreted interleukin 2 (IL-2), reflecting T cell activation, were determined using an electrochemoluminescence-based assay as a readout value reflecting T cell activation and normalized to the levels of the corresponding control IgG4 as described in Example 11. All fusions proteins are capable of inducing T cell activation, and induce to a greater extent or at least comparable to individual building blocks or a mixture of anti-PD-L1/anti-CD137 reference antibodies.

Fig. 10: Demonstrates the ability of the exemplary fusion proteins to co-stimulate T cell activation in a PD-L1 target-dependent manner. Anti-PD-L1 antibodies, including the corresponding anti-PD-L1 building block antibody, CD137-binding lipocalin muteins as Fc fusions, and the reference anti-CD137 antibody, were tested individually and in combination as an anti-PD-L1/αntuCD137 cocktail . Different tumor cell lines expressing different levels of PD-L1 (high: RKO; medium: HCC827; negative: HepG) were seeded onto anti-human CD3-coated plates. Pan T cells and various concentrations of fusion proteins and individual building blocks were added and incubated for 3 days. Levels of secreted IL-2 were determined using an electrochemoluminescence assay as described in Example 12. All fusion proteins are capable of increasing IL-2 secretion in a PD-L1 dependent manner.

Fig. 11: Illustrates the stability of fusion proteins when stored in PBS or 25 mM histidine, 60 mM NaCl, 200 mM arginine at pH 6 after 1-, 2-, 3- or 4-week incubation at 37°C or 40°C at concentration 1 mg/ml or 20 mg/ml. Stability is assessed by the percentage of monomer peak in analytical size exclusion chromatography or the percentage of functional proteins detected in a quantitative ELISA as described in Example 13.

Fig. 12: Demonstrates the ability of the exemplary fusion protein (SEQ ID NOs: 90 and 87) to stimulate IL-2 secretion in a mixed lymphocyte culture reaction (MLR) with CD4+ T cells. Fusion proteins, anti-PD-L1 building block antibody (SEQ ID NOs: 86 and 87), CD137-binding lipocalin mutein Fc fusion (SEQ ID NO: 89) and reference anti-CD137 antibody or reference anti-PD-L1 antibody , alone or in combination as an anti-PD-L1/anti-CD137 cocktail, were tested at equimolar concentrations as described in Example 14. IL-2 secretion was measured in supernatants after 6 days of incubation with total human CD4+ T cells and monocytic dendritic cells (moDC) from different healthy donors. Fig. 12A illustrates that the fusion protein with SEQ ID NOs: 90 and 87 showed a significant increase in IL-2 secretion compared to the building block alone (anti-PD-L1 antibody with SEQ ID NOs: 86 and 87 or CD137-specific lipocalin mutein with the sequence SEQ ID NO: 89) and a reference antibody to PD-L1 or CD137 (SEQ ID NOs: 26 and 27 or SEQ ID NO: 28 and 29, respectively), at equimolar concentrations (equivalent to 10 or 0.1 μg/ml of test fusion protein). Data from 8 independent experiments are presented. Fig. 12B demonstrates that the fusion protein with the sequences SEQ ID NO: 90 and 87 was able to induce dose-dependent secretion of IL-2 in the concentration range from 0.001 to 20 μg/ml. IL-2 levels induced by the fusion protein were higher compared to equimolar concentrations of a cocktail of reference anti-PD-L1 antibody (SEQ ID NOs: 26 and 27) and reference anti-CD137 antibody (SEQ ID NOs: 28 and 29). Data are shown for an illustrative donor.

Fig. 13: Demonstrates the ability of the exemplary fusion protein (SEQ ID NOs: 90 and 87) to induce the secretion of CD8+ T cell effector molecules. The fusion protein was cultured with moDC and CD8+ T cells from unmatched healthy donors for 6 days, after which the secretion of IL-2 and CD8' T cell effector molecules in the supernatants was quantified using the Luminex assay as described in Example 15. Fusion Protein with SEQ ID NOs: 90 and 87 showed increased secretion of IL-2 and cytotoxic factors (perforin, granzyme B and granzyme A) at 10 μg/ml compared to the reference anti-PD-L1 antibody (SEQ ID NOs: 26 and 27) and the reference antibody to CD137 (SEQ ID NOs: 28 and 29), when used alone or as a cocktail.

Fig. 14: demonstrates that the fusion proteins bind overlapping epitopes to the clinically active anti-CD137 antibody (SEQ ID NOs: 28 and 29), as illustrated in competitive ELISA studies as described in Example 17. A constant concentration of SEQ ID NOs: 28 and 29 was applied onto a microtiter plate followed by the addition of a mixture of test molecules at various concentrations and the biotinylated huCD137-Fc indicator at a fixed concentration. The bound indicator was detected using ExtrAvidin peroxidase. Show jumping fused squirrels

- 11 046634 are used with the anti-CD137 antibody for binding to CD137.

Fig. 15: Demonstrates the potential of the exemplary fusion proteins to block an inhibitory signal mediated by the PD-1/PD-L1 interaction, assessed using a PD-1/PD-L1 blockade bioassay as described in Example 18. PD-1-NFAT-luc T Cells Jurkat (a Jurkat cell line expressing PD-1 and the NFAT-mediated luciferase gene under the control of the NFAT promoter) was co-cultured with PD-L1 aAPC/CHO-K1 cells in the presence of various concentrations of test molecules. After 6 h, luciferase assay reagent was added and luminescent signals were measured. The background signal is PD-1-NFAT-luc Jurkat T cells cocultured only with PD-L1 aAPC/CHO-K1 cells. The fusion protein of SEQ ID NOs: 90 and 87 blocks the PD-1/PD-L1 signaling pathway at levels comparable to tested anti-PDL1 antibodies, including the anti-PD-Ll building block antibody shown in SEQ ID NOs: 86 and 87 , and the reference anti-PD-L1 antibody shown in SEQ ID NOs: 26 and 27.

Fig. 16: Demonstrates the ability of an exemplary fusion protein to induce T cell activation. An anti-PD-Ll building block antibody and a reference anti-CD137 antibody were also tested when used alone and in combination with an anti-PD-L1 antibody. In the experiment, human PBMCs were incubated with fusion protein, antibodies, cocktails, or control in the presence of 0.1 ng/ml SEB. Levels of secreted IL-2 were determined using an electrochemiluminescence-based assay as a readout value reflecting T cell activation as described in Example 19 and depicted in FIG. 16A. Fig. 16B shows the fold increase in IL2 secretion levels induced by test molecules compared to background IL-2 secretion levels (PBMCs stimulated with 0.1 ng/ml SEB and without any test molecules). The fusion protein causes a dose-dependent increase in IL-2 secretion at levels greater than anti-PD-L1 or anti-CD137, alone or in combination.

Fig. 17: Demonstrates the ability of an exemplary fusion protein to co-stimulate T cell activation in the presence of PD-L1. In parallel, we tested an anti-PD-L1 building block antibody, a reference anti-CD137 antibody, and a cocktail of an anti-CD137 antibody and a reference anti-PD-L1 antibody. CHO cells transfected with human PD-L1 (Fig. 17A) or mock transfected (negative for human PD-L1, Fig. 17B) were seeded on plates coated with anti-human CD3 antibody. Pan T cells were added, as well as various concentrations of test molecules, and incubated for 2 days. Levels of secreted IL-2 in the supernatant were determined using an electrochemiluminescence-based assay as described in Example 20. Levels of IL-2 secretion were normalized to background levels (pan-T cells + anti-CD3 + CHO cells) to reflect the fold increase secretion of IL-2 in the presence of CHO cells expressing human PD-L1 (Fig. 17C) or mock transfected CHO cells (Fig. 17D). The fusion protein induces a strong dose-dependent increase in IL-2 secretion only in the presence of PD-L1, at levels above the reference CD137 antibody alone or in combination with the reference PD-L1 antibody.

Fig. 18: Shows the result of pharmacokinetic assays of bispecific fusion proteins and anti-PD-L1 building block antibodies (SEQ ID NOs: 86 and 87) in mice as described in Example 21. Male CD-1 mice (3 mice per time point) were intravenously administered fusion proteins at a dose of 10 mg/kg. Drug levels were determined using a sandwich ELISA, detecting the whole molecule through the targets PD-L1 and CD137. Plasma anti-PD-L1 antibody levels were determined using a sandwich ELISA targeting PD-L1 and human Fc.

Fig. 19: Shows the results of a pharmacokinetic analysis of an exemplary fusion protein (SEQ ID NO: 90 and 87) compared to two previously described CD137- and PD-L1-binding fusion proteins (SEQ ID NO: 147 and SEQ ID NO: 148) in mice, as described in Example 22. Male CD-1 mice (2 mice per time point) were intravenously administered test molecules at a dose of 2 mg/kg. Drug levels were determined by ELISA at the indicated time points. The data were plotted on a graph of concentration versus time. SEQ ID NOs: 90 and 87 described herein, but not SEQ ID NO: 147 or SEQ ID NO: 148, demonstrate a favorable pharmacokinetic profile or antibody-like pharmacokinetics.

IV. Detailed Description of the Invention

As described herein, the present invention includes the recognition that a bivalent CD137-binding molecule, such as an antibody, by itself may not be sufficient to cluster CD137 on T cells or NK cells and provide effective activation, similar to the lack of activity of a trivalent soluble CD137L. Recent publications using preclinical in vivo mouse models have provided evidence that the mechanism of action of other anti-TNFR antibodies requires interaction of the antibodies through their Fc portion with Fc-gamma receptors on cells expressing Fc-gamma receptors (Bulliard et al., Immunol Cell Biol, 2014, Bulliard et al., J Exp Med, 2013). Thus, the mechanism of action of these TNFR antibodies may be dominated by off-target clustering through Fc-gamma receptors, depending on the presence of cells expressing the Fc-gamma receptor, which may not necessarily be overexpressed in the target tumor microenvironment compared to normal tissues .

- 12 046634

Thus, there is an unmet need to develop therapeutics that cluster and activate CD137 with a specific, tumor-targeting mechanism of action.

To address this unmet need, the present disclosure provides, among other things, novel approaches for simultaneously binding CD137 and PD-L1 using one or more fusion proteins having CD137 binding specificity and PD-L1 binding specificity. The proposed fusion proteins are designed to promote CD137 clustering by linking CD137-positive T cells to PD-L1 expressed in the tumor microenvironment. Such bispecific molecules may combine CD137-induced T cell activation and expansion with anti-PD-LT-mediated immune checkpoint blockade and thus may overcome certain limitations of monotherapy and provide benefits, for example, to resistant or refractory patients. Fusion proteins are also designed to provide combination therapy capabilities in a single molecule, while allowing localized induction of antigen-specific T cells in the tumor microenvironment, potentially reducing peripheral toxicity.

Some aspects of the present invention provide fusion proteins that bind CD137 and PD-L1, as well as methods and useful applications thereof. The invention also provides methods for producing the CD137-PD-L1 binding fusion proteins described herein, as well as compositions containing such proteins. The CD137 and PD-L1 binding fusion proteins of the invention, as well as compositions containing them, can be used in methods of detecting CD137 and/or PDL1 in a sample, in methods of binding CD137 and/or PD-L1 in a subject, or in methods of modulating immune responses. responses from the subject. Similar fusion proteins having these characteristics relevant to the applications contemplated by the present invention have not been previously described. In contrast to the fusion proteins proposed herein, previously known fusion proteins targeting both CD137 and PD-L1 had one or more of the following disadvantages: poor pharmacokinetics, unacceptable degree of off-target binding, reduced or otherwise impaired ability bind one or both targets of a single fusion protein (eg, PD-L1 and/or CD137) and/or an unacceptable degree of nonspecific (eg, PD-L1-independent) activation of, for example, the immune system.

A. Exemplary CD137 and PD-L1 specific fusion proteins of the invention.

In some embodiments, the proposed fusion protein contains at least two subunits in any order: (1) a first subunit containing a full-length immunoglobulin or its antigen binding domain specific for PD-L1, and (2) a second subunit containing a lipocalin mutein specific for PD-L1. CD 137.

In some embodiments, the proposed fusion protein may also contain at least one additional subunit, for example, a third subunit. For example, the fusion protein may contain a third subunit specific for CD137. In some embodiments, the third subunit may be or comprise a CD137-specific lipocalin mutein. For example, two lipocalin muteins can be fused to the first immunoglobulin subunit, one at the C-terminus and one at the N-terminus of the immunoglobulin. In some embodiments, lipocalin muteins may be fused to an immunoglobulin heavy chain or light chain.

In some embodiments, the proposed fusion proteins may contain one or more additional subunits (eg, a fourth, fifth or sixth subunit).

In some embodiments, at least one subunit may be fused at its N-terminus and/or its C-terminus to another subunit.

In some embodiments, at least one subunit may be linked to another subunit via a linker. In some further embodiments, the linker is a peptide linker, such as a non-structured glycine-serine (GS) linker, a glycosylated GS linker, or a proline-alanine-serine polymer (PAS) linker. In some embodiments, the GS linker is the (Gly ₄ Ser) ₃ ((G ₄ S) ₃ ) linker set forth in SEQ ID NO: 13. Other exemplary linkers are set forth in SEQ ID NO: 14-23. In some embodiments, the peptide linker may contain from 1 to 50 amino acids, for example, 1, 2, 3, 4, 5, 10, 11, 12, 13, 14, 15, 16, 17 18, 19, 20, 25, 30 , 35, 40, 45 or 50 amino acids. For example, when the first subunit contains a full-length immunoglobulin, the second subunit can be linked through a peptide linker between the N-terminus of the second subunit and the C-terminus of the heavy chain constant region (CH) of said immunoglobulin. In some further embodiments, the third subunit may be linked via a peptide linker between the N-terminus of the third subunit and the C-terminus of the light chain constant region (CL) of the immunoglobulin.

In some embodiments, one subunit may be linked to another subunit substantially as described in FIG. 1. Typically, one subunit may be fused at its N-terminus and/or its C-terminus to another subunit. For example, in some embodiments, a lipocalin mutein subunit may be fused at its N-terminus and/or at its C-terminus to an immunoglobe subunit

- 13 046634 lin. As another example, one lipocalin mutein may be linked, preferably via a peptide bond, to the C terminus of an immunoglobulin heavy chain (HC) domain, the N terminus of an HC domain, the C terminus of an immunoglobulin light chain (LC), and/or the N terminus LC (Fig. 1A-1D).

In some embodiments, a lipocalin mutein subunit may be fused at its N-terminus and/or at its C-terminus to an immunoglobulin moiety. For example, in some embodiments, a lipocalin mutein may be linked, preferably via a peptide linker, to the C-terminus of a heavy chain constant region (CH) or the C-terminus of a light chain constant region (CL) of an immunoglobulin.

In some embodiments, when one subunit contains a full-length immunoglobulin, the second subunit may be linked between the N-terminus of the second subunit and the C-terminus of the heavy chain constant region (CH) of said immunoglobulin.

In some embodiments, the third subunit may be linked between the N-terminus of the third subunit and the C-terminus of the light chain constant region (CL) of said immunoglobulin.

In some embodiments, with respect to a fusion protein of the invention, wherein at least one subunit may be or comprise a full-length immunoglobulin, the Fc function of the Fc region of the full-length immunoglobulin relative to the Fc receptor-positive cell may be retained, and the fusion protein may simultaneously bind CD137 and PD-L1.

In some embodiments, where at least one subunit of the proposed fusion protein may be or contain a full-length immunoglobulin, the Fc function of the Fc region of the full-length immunoglobulin relative to the Fc receptor-positive cell can be reduced or completely suppressed through protein engineering, and the fusion protein can simultaneously bind CD137 and PD-L1. In some embodiments, this may be achieved, for example, by replacing the IgG1 backbone with IgG4, since IgG4 is known to exhibit reduced interactions with the Fc gamma receptor compared to IgG1. In some embodiments, mutations such as F234A and L235A can be introduced into the IgG4 backbone to further reduce residual binding to Fc-gamma receptors. In some embodiments, the S228P mutation may also be introduced into the IgG4 backbone to minimize the exchange of IgG4 halves (Silva et al., J Biol Chem, 2015). In some embodiments, the F234A and L235A mutations may be introduced to reduce ADCC (antibody-dependent cell-mediated cytotoxicity) and ADCP (antibody-dependent cell-mediated phagocytosis) (Glaesner et al., Diabetes Metab Res Rev, 2010) and/or the M428L and N434S mutations or the M252Y mutation, S254T and T256E to increase serum half-life (Dall'Acqua et al., J Biol Chem, 2006, Zalevsky et al., Nat Biotechnol, 2010). In some embodiments, an additional N297A mutation may be present in the immunoglobulin heavy chain of the fusion protein to remove the natural glycosylation motif.

In some embodiments, the Fc portion of an immunoglobulin included in a fusion protein of the invention may help maintain serum levels of the fusion protein. For example, when the Fc moiety binds to Fc receptors on endothelial cells and phagocytes, the fusion protein can be internalized and released back into the bloodstream, increasing its half-life in the body.

In one aspect, the fusion proteins of the invention bind CD137 with high affinity. In yet another aspect, the proposed fusion proteins bind PD-L1 with high affinity. In some preferred embodiments, the proposed fusion proteins simultaneously bind CD137 and PD-L1. In some embodiments, simultaneous binding to CD137 and PD-L1 allows the proposed fusion proteins to demonstrate a sustained antitumor or anti-infective response.

In some embodiments, the fusion protein of the invention may be capable of binding PD-L1 with a KD value of no more than about 2 nM or even lower, such as about 1.5 nM or lower, about 1 nM or lower, about 0.6 nM, or below or about 0.4 nM or below. In some embodiments, a fusion protein of the invention may be capable of binding PD-L1 with a KD value comparable to or lower than the KD value of a PD-L1-specific immunoglobulin included in such fusion protein, such as an antibody having a heavy and a light chains represented by SEQ ID NOs: 86 and 87. The KD values of the proposed fusion proteins can be measured, for example, in a surface plasmon resonance (SPR) assay, such as an SPR assay as substantially described in Example 3.

In some embodiments, the fusion protein of the invention may be capable of binding CD137 with a KD value of no more than about 10 nM or even lower, for example, about 7 nM, about 6 nM, or about 5 nM, about 4 nM, about 3 nM, about 2 nM or even lower. In some embodiments, the fusion protein of the invention may be capable of binding CD137 with a K _D value comparable to or lower than the K _D value of the CD137-specific lipocalin mutein included in the particular fusion protein, e.g., SEQ ID NO: 42, or lipocalin mutein fused to the Fc region of an antibody,

- 14 046634 for example, SEQ ID NO: 89. The KD values of the proposed fusion proteins can be measured, for example, in an SPR assay, such as an SPR assay as substantially described in Example 3.

In some embodiments, the fusion protein of the invention may be capable of binding PD-L1 with an EC50 value of no more than about 0.5 nM or even lower, such as about 0.3 nM or lower, about 0.2 nM or lower, about 0 .15 nM or lower or approximately 0.1 nM or lower. In some embodiments, a fusion protein of the invention may be capable of binding PD-L1 with an EC ₅₀ value comparable to or lower than the EC ₅₀ value of a PD-L1-specific immunoglobulin contained in the particular fusion protein, such as an antibody having a severe and the light chain represented by SEQ ID NOs: 86 and 87. The EC50 values of the proposed fusion proteins can be measured, for example, in an enzyme-linked immunosorbent assay (ELISA) assay, such as an ELISA assay as substantially described in Example 4.

In some embodiments, the fusion protein of the invention may be capable of binding CD137 with an EC ₅₀ value of no more than about 0.6 nM or even lower, such as about 0.5 nM or lower, about 0.2 nM or lower, about 0. 15 nM or lower or approximately 0.1 nM or lower. In some embodiments, the fusion protein of the invention may be capable of binding CD137 with an EC ₅₀ value comparable to or lower than the EC ₅₀ value of the CD137-specific lipocalin mutein included in the particular fusion protein, for example, SEQ ID NO: 42, or a lipocalin mutein fused to the Fc region of an antibody, for example, SEQ ID NO: 89. The EC ₅₀ values of the proposed fusion proteins can be measured, for example, in an ELISA assay, such as an ELISA assay as substantially described in Example 4.

In some embodiments, the fusion proteins of the invention are cross-reactive with cynomolgus PD-L1. In some embodiments, the proposed fusion protein may be capable of binding cynomolgus PD-L1 with an EC50 value of no more than about 0.5 nM or even lower, such as about 0.2 nM or lower, about 0.1 nM or lower, or about 0.05 nM. or lower. The EC50 values of the proposed fusion proteins can be measured, for example, measured in an ELISA assay, such as an ELISA assay as substantially described in Example 4.

In some embodiments, the fusion proteins of the invention are cross-reactive with cynomolgus CD137. In some embodiments, the proposed fusion protein may be capable of binding cynomolgus CD137 with an EC ₅₀ value of no more than about 15 pM or even lower, e.g., about 10 nM or lower, about 8 nM or lower, about 6 nM or lower, about 3 nM or lower, about 1 nM or lower, about 0.5 nM or lower, about 3 nM or lower, or about 0.1 nM or lower. The EC ₅₀ values of the proposed fusion proteins may be measured, for example, in an ELISA assay, such as the ELISA assay as substantially described in Example 4. In some embodiments, the binding of the proposed fusion protein to cynomolgus CD137 may be enhanced by an avidity effect, as described in example 22.

In some embodiments, fusion proteins of the invention may be capable of simultaneously binding CD137 and PD-L1. In some embodiments, the proposed fusion protein may be capable of simultaneously binding CD137 and PD-L1 with an EC ₅₀ value of no more than about 1 nM or even lower, such as 0.8 nM or lower, 0.6 nM or lower, or 0.4 nM or lower. In some other embodiments, the proposed fusion protein may be capable of simultaneously binding CD137 and PD-L1 with an EC ₅₀ value of no more than about 10 nM or even lower, such as 8 nM or lower, 6 nM or lower, 3 nM or lower, or 2 nM or lower. Simultaneous binding can be determined, for example, in an ELISA assay, such as an ELISA assay as substantially described in Example 5.

In some embodiments, the fusion protein of the invention may be capable of binding CD137 expressed on a cell to an EC ₅₀ value of no more than about 60 nM or even lower, such as about 50 nM or even lower than about 40 nM or even lower than about 30 nM or lower, about 10 nM or lower, about 7 nM or lower, about 5 nM or lower, about 3 nM or lower, or about 1 nM or even lower. The EC ₅₀ value of the proposed fusion protein can be measured, for example, in a flow cytometry assay as essentially described in Example 6. The cell expressing CD137 can be, for example, a CHO cell transfected with human CD137 or cynomolgus CD137.

In some embodiments, the fusion protein of the invention may be capable of binding PD-L1 expressed on a cell to an EC50 value of no more than about 10 nM or even lower, such as about 8 nM or even lower than about 6 nM or even lower than about 4 nM or lower, about 2 nM or lower, or about 1 nM or even lower. The EC50 value of the proposed fusion protein can be measured, for example, in a flow cytometry assay as essentially described in Example 6. The cell expressing PD-L1 can be, for example, a CHO cell transfected with human PD-L1 or PD-L1 Javan

- 15 046634 sky macaque.

In some embodiments, fusion proteins of the invention may be capable of binding PD-L1 expressed on tumor cells. In some embodiments, the proposed fusion protein may be capable of binding PD-L1 expressed on a tumor cell to an EC50 value of no more than about 2 nM or even lower, such as about 1.5 nM or lower, about 1 nM or lower, about 0.6 nM or lower or approximately 0.3 nM or even lower. The EC ₅₀ value of the fusion protein for binding tumor cells expressing PD-L1 can be measured, for example, in a flow cytometry assay as substantially described in Example 7. Tumor cells expressing PD-L1 can be, for example, RKO.

In some embodiments, the fusion proteins of the invention do not substantially affect the binding of CD137 to CD137L. In some embodiments, fusion proteins of the invention may be capable of binding CD137 complexed with CD137L. In some embodiments, the fusion proteins of the invention may be capable of binding CD137 through a mechanism similar to the anti-CD137 antibody having the heavy and light chains represented by SEQ ID NOs: 28 and 29. The CD137 binding mechanism of the fusion protein can be determined, for example, by an SPR assay, such as an SPR assay as substantially described in Example 8.

In some embodiments, fusion proteins of the invention may be able to compete with PD-1 for binding to PD-L1. In some embodiments, the proposed fusion protein may be capable of competing with PD-1 for binding to PD-L1 with an IC ₅₀ value of no more than about 5 nM or even lower, such as about 3 nM or lower, about 2 nM or lower, or about 1 or even lower. The mechanism of inhibitory action can be determined, for example, through an ELISA assay, such as an ELISA assay as substantially described in Example 9.

In some embodiments, the fusion proteins of the invention may be capable of competing with the anti-CD137 antibody set forth in SEQ ID NOs: 28 and 29 for binding to CD137. Such competition can be assessed using an ELISA assay as substantially described in Example 17. In some embodiments, the proposed fusion protein may have an overlapping epitope with the anti-CD137 antibody set forth in SEQ ID NOs: 28 and 29.

In some embodiments, the fusion proteins of the invention may be capable of co-stimulating T cell responses. In some embodiments, the proposed fusion proteins result in comparable or greater T cell activation compared to an anti-PDL1 antibody, such as an anti-PD-L1 building block antibody of SEQ ID NOs: 86 and 87 or a reference anti-PD-L1 antibody with the sequences of SEQ ID NOs: 26 and 27, or an anti-CD137 antibody, such as a reference antibody with the sequences of SEQ ID NOs: 28 and 29. In some embodiments, the proposed fusion proteins lead to T cell activation with comparable or better efficiency compared to a combination of an anti-PD-L1 antibody and a molecule targeting CD137, such as an anti-CD137 antibody or the previously known CD137-specific lipocalin mutein. Stimulated T cell response or T cell activation can be measured, for example, in a CD137 bioassay as essentially described in Example 10, in a PD-1/PD-L1 blockade bioassay as described in Example 18, or in a functional T cell activation assay as essentially described in Example 11, Example 12, Example 19 and Example 20.

In some embodiments, fusion proteins of the invention may be capable of inducing increased secretion of IL-2. In some preferred embodiments, the proposed fusion proteins may be capable of inducing concentration-dependent secretion of IL-2 and/or tend to induce increased secretion of IL-2 at higher concentrations, preferably concentrations when applied as a coating. In some embodiments, the proposed fusion proteins may result in increased IL-2 secretion with comparable or better efficacy than a combination of an anti-PD-L1 antibody and a CD137-targeting molecule, such as an anti-CD137 antibody or the previously known CD137-specific lipocalin mutein. IL-2 secretion can be measured, for example, in a functional T-cell activation assay as substantially described in Example 11 and Example 19.

In some embodiments, fusion proteins of the invention may be capable of co-stimulating T cell responses in a PD-L1 dependent manner. In some embodiments, the proposed fusion proteins may result in local induction of IL-2 production by T cells in the vicinity of PD-L1 positive cells, such as PD-L1 transfected cells or PD-L1 positive tumor cells. Proximity PD-L1 positive cells as used herein refers to a T cell and a PD-L1 positive cell brought into proximity to each other by a proposed fusion protein that simultaneously binds CD137 and PD-L1. PD-L1-dependent T cell activation by the proposed fusion proteins can be determined, for example, in a CD137 bioassay substantially described in Example 10, in a PD-1/PD-L1 blockade bioassay substantially described in Example 18, or in a functional T cell activation assay essentially described in Note

- 16 046634 re 12 and example 20.

In some preferred embodiments, the proposed fusion proteins may be capable of co-stimulating T cell responses in the presence of tumor cells expressing PD-L1 and/or in the tumor microenvironment. In some embodiments, the proposed fusion protein may be capable of co-stimulating T cell responses in the presence of PD-L1 positive tumor cells with an EC50 value of about 1 nM or less, about 0.5 nM or less, about 0.3 nM or less, about 0.1 nM or lower or approximately 0.05 nM or lower. Activation of T cells by the proposed fusion proteins in the presence of tumor cells expressing PD-L1 and/or in the tumor microenvironment can be assessed, for example, in a CD137 bioassay, essentially described in example 10, or in a functional T cell activation assay, essentially described in example 12.

In some embodiments, the proposed fusion proteins are unable to co-stimulate T cell responses in the absence of PD-L1. In some embodiments, the proposed fusion proteins are unable to co-stimulate T cell responses in the absence of cells expressing PD-L1. In some embodiments, the proposed fusion protein may be better able to recognize the presence of PD-L1 and lead to appropriate T cell activation than the anti-CD137 antibody set forth in SEQ ID NOs: 28 and 29. The PD-L1-dependent effect of the fusion proteins may be as determined, for example, in the CD137 bioassay substantially described in Example 10, in the PD1/PD-L1 blockade bioassay substantially described in Example 18, or in the functional T cell activation assay substantially described in Example 12 and Example 20.

In some embodiments, the proposed fusion proteins may be capable of blocking an inhibitory signal mediated by PD-1 binding to PD-L1. In some embodiments, the proposed fusion protein may be able to relieve the brake on T cell activation or result in successful T cell activation by blocking the PD-1/PD-L1 interaction. Blockade of the PD-1 inhibitory signal can be measured, for example, in a PD-1/PD-L1 blockade bioassay as described in Example 18.

In some embodiments, the fusion proteins of the invention may be capable of promoting T cell proliferation and/or activation. In some embodiments, the proposed fusion proteins may be capable of promoting proliferation and/or activation of CD4+ T cells. In some embodiments, the proposed fusion proteins may be capable of inducing the secretion of IL-2, preferably a dose-dependent secretion of IL-2. In some embodiments, the proposed fusion proteins may be capable of inducing a higher level of IL-2 secretion compared to a combination of an anti-PD-L1 antibody and a CD137-targeting molecule, such as an anti-CD137 antibody or the previously known CD137-specific lipocalin mutein. Secretion of IL-2 as used herein may be a measure of T cell activation. The proliferation and/or activation of CD4+ T cells stimulated by the proposed fusion proteins can be assessed using, for example, a mixed lymphocyte response (MLR) assay, as essentially described in Example 14.

In some embodiments, the fusion proteins of the invention may be capable of stimulating the proliferation and/or activation of CD8+ T cells. In some embodiments, the proposed fusion proteins may be capable of inducing the production of IL-2 and effector molecules such as perforin, granzyme A, and granzyme B. In some embodiments, the proposed fusion proteins may be capable of inducing increased production of IL-2 and cytotoxic factors such as as perforin, granzyme B and granzyme A, compared with a combination of an anti-PD-L1 antibody and a molecule targeting CD137, such as an anti-CD137 antibody or the previously known CD137-specific lipocalin mutein. The proliferation and/or activation of CD8+ T cells stimulated by the proposed fusion proteins can be assessed using, for example, an MLR assay, as essentially described in Example 15.

In some embodiments, the proposed fusion proteins have a favorable stability profile and pharmacokinetic profile. In some embodiments, the proposed fusion protein has a pharmacokinetic profile comparable to the building block antibody of SEQ ID NOs: 86 and 87. In some embodiments, the proposed fusion protein has antibody-like pharmacokinetics. In some embodiments, the proposed fusion protein has a terminal half-life of about 200 hours or longer, about 250 hours or longer, about 300 hours or longer, about 350 hours or longer, about 400 hours or longer, or even longer. In some embodiments, the proposed fusion protein has a more favorable pharmacokinetic profile than SEQ ID NO: 147. In some embodiments, the proposed fusion protein has a more favorable pharmacokinetic profile than SEQ ID NO: 148. The pharmacokinetic profiles of the proposed fusion proteins can be analyzed as described in Example 21 and Example 22. In some embodiments, a favorable pharmacokinetic profile or antibody-like pharmacokinetics may be considered to have been achieved if the % _cmax was greater than 10% after 336 hours.

- 17 046634

In some embodiments, the proposed fusion protein contains the amino acid sequence presented in any of SEQ ID NO: 88-94.

In some embodiments, the proposed fusion protein contains an amino acid sequence having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 92%, at least at least 95%, at least 97%, at least 98% or even higher sequence identity with the amino acid sequences presented in any of SEQ ID NOs: 88-94.

In some embodiments, the proposed fusion protein contains the amino acids shown in SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87 or SEQ ID NO: 90 and 91.

In some embodiments, the proposed fusion protein contains amino acid sequences having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 92%, at least at least 95%, at least 97%, at least 98%, or even greater sequence identity to the amino acid sequences set forth in SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87 , SEQ ID NO: 86 and 93, SEQ ID NO: 94 and 87, or SEQ ID NO: 90 and 91.

B. Illustrative immunoglobulins included in fusion proteins.

In some embodiments, with respect to the proposed fusion protein, the first subunit may be or contain a full-length immunoglobulin or a PD-L1-specific antigen binding domain thereof. In some embodiments, the immunoglobulin, for example, may be IgG1, IgG2, or IgG4. In some embodiments, the immunoglobulin is or contains IgG4. In some embodiments, the immunoglobulin is an anti-PD-L1 monoclonal antibody.

Illustrative examples of PD-L1 binding antibodies of the invention may comprise an antigen binding region that cross-blocks or binds to the same epitope as a PD-L1 binding antibody comprising variable heavy chain (VH) and variable light chain (VL) regions ) a known antibody such as atezolizumab (also known as MPDL3280A or RG7446, trade name Tecentriq®), avelumab (also known as MSB0010718C, trade name Bavencio®), durvalumab (formerly known as MEDI4736, trade name Imfinzi®) and BMS-936559 (also known as MDX-1105), 5C10 (including humanized 5C10), 5F10 (including humanized 5F10) and 9F6 (including humanized 9F6). In some embodiments, a PD-L1 binding antibody of the invention may comprise an antigen binding region, such as any of three heavy chain complementarity determining regions (CDRs) (HCDR1, HCDR2, and HCDR3) and three light chain CDRs (LCDR1, LCDR2, and LCDR3) of an antibody selected from the group consisting of atezolizumab, avelumab, durvalumab, BMS-936559, 5C10, 5F10 and 9F6.

In some embodiments, a proposed anti-PD-L1 antibody or antigen binding domain thereof may have a heavy chain variable region (HCVR) selected from the group consisting of SEQ ID NOs: 75-79 and/or a light chain variable region (LCVR) selected from the group consisting of SEQ ID NO: 80-84.

In some embodiments, a provided anti-PD-L1 antibody or antigen binding domain thereof may have a heavy chain of any of SEQ ID NO: 85-86 and/or a light chain of SEQ ID NO: 87.

In some embodiments, the heavy chain and light chain pair of the proposed anti-PD-L1 antibody or antigen binding domain thereof is or comprises the following HCVR and LCVR, respectively: SEQ ID NOs: 75 and 80, SEQ ID NOs: 76 and 81, SEQ ID NOs : 77 and 82, SEQ ID NO: 78 and 83 or SEQ ID NO: 79 and 84.

In some embodiments, the heavy chain and light chain pair of the proposed anti-PD-L1 antibody is or contains the amino acid sequences set forth in SEQ ID NOs: 85 and 87 or SEQ ID NOs: 86 and 87.

In some embodiments, a proposed anti-PD-L1 antibody or antigen binding domain thereof may have an HCVR having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 92%, at least 95%, at least 97%, at least 98% or even higher sequence identity to an amino acid sequence selected from the group consisting of SEQ ID NOs: 75-79, and/or LCVR having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 92%, at least 95%, at least 97%, at least 98% or even higher sequence identity with an amino acid sequence selected from the group consisting of SEQ ID NO: 80-84. In other embodiments, the proposed anti-PD-L1 antibody or antigen binding domain thereof may have a heavy chain having at least 70%, at least 75%, at least 80%,

- 18 046634 at least 85%, at least 90%, at least 92%, at least 95%, at least 97%, at least 98% or even higher sequence identity with an amino acid sequence selected from the group consisting of SEQ ID NO: 85-86, and/or a light chain having at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least at least 92%, at least 95%, at least 97%, at least 98%, or even greater sequence identity to the amino acid sequence of SEQ ID NO: 87.

In some embodiments, the heavy chain variable region of a proposed anti-PD-L1 antibody or antigen binding domain thereof may have three CDRs having the following sequences: GFSLSNYD (HCDR1, SEQ ID NO: 60), IWTGGAT (HCDR2, SEQ ID NO: 61), VRDSNYRYDEPFTY (HCDR3; SEQ ID NO: 62). In some embodiments, the heavy chain variable region of a proposed anti-PD-L1 antibody or antigen binding domain thereof may have three CDRs having the following sequences: GFDIKDTY (HCDR1, SEQ ID NO: 65), IDPADGNT (HCDR2, SEQ ID NO: 66), ARGLGAWFAS (HCDR3; SEQ ID NO: 67). In some embodiments, the heavy chain variable region of a proposed anti-PD-L1 antibody or antigen binding domain thereof may have three CDRs having the following sequences: GFNIKDTY (HCDR1, SEQ ID NO: 70), IDPANGNT (HCDR2, SEQ ID NO: 71), SRGPPGGIGEYIYAMDY (HCDR3; SEQ ID NO: 72).

In some embodiments, the light chain variable region of a proposed anti-PD-L1 antibody or antigen binding domain thereof may have three CDRs having the following sequences: QSIGTN (LCDR1, SEQ ID NO: 63), YAS (LCDR2), QQSNSWPYT (LCDR3; SEQ ID NO: 63) : 64). In some embodiments, the light chain variable region of a proposed anti-PDL1 antibody or antigen binding domain thereof may have three CDRs having the following sequences: QDITNS (LCDR1, SEQ ID NO: 68), YTS (LCDR2), QQGHTLPPT (LCDR3; SEQ ID NO: 69 ). In some embodiments, the light chain variable region of a proposed anti-PD-L1 antibody or antigen binding domain thereof may have three CDRs having the following sequences: SSVSSSY (LCDR1, SEQ ID NO: 73), STS (LCDR2), HQYHRSPPT (LCDR3; SEQ ID NO: 73), STS (LCDR2), HQYHRSPPT (LCDR3; SEQ ID NO: 73). : 74).

In some embodiments, the proposed anti-PD-L1 antibody or antigen binding domain thereof comprises a heavy chain variable region having three CDRs having the following sequences: GFSLSNYD (HCDR1, SEQ ID NO: 60), IWTGGAT (HCDR2, SEQ ID NO: 61), VRDSNYRYDEPFTY (HCDR3; SEQ ID NO: 62), and a light chain variable region having three CDRs having the following sequences: QSIGTN (LCDR1, SEQ ID NO: 63), YAS (LCDR2), QQSNSWPYT (LCDR3; SEQ ID NO: 64 ). In some embodiments, the proposed anti-PD-L1 antibody or antigen binding domain thereof comprises a heavy chain variable region having three CDRs having the following sequences: GFDIKDTY (HCDR1, SEQ ID NO: 65), IDPADGNT (HCDR2, SEQ ID NO: 66), ARGLGAWFAS (HCDR3; SEQ ID NO: 67), and a light chain variable region having three CDRs having the following sequences: QDITNS (LCDR1, SEQ ID NO: 68), YTS (LCDR2), QQGHTLPPT (LCDR3; SEQ ID NO: 69 ). In some embodiments, the proposed anti-PD-L1 antibody or antigen binding domain thereof comprises a heavy chain variable region having three CDRs having the following sequences: GFNIKDTY (HCDR1, SEQ ID NO: 70), IDPANGNT (HCDR2, SEQ ID NO: 71), SRGPPGGIGEYIYAMDY (HCDR3; SEQ ID NO: 72), and a light chain variable region having three CDRs having the following sequences: SSVSSSY (LCDR1, SEQ ID NO: 73), STS (LCDR2), HQYHRSPPT (LCDR3; SEQ ID NO: 74 ).

Unless otherwise stated, all CDR sequences disclosed herein are determined according to the IMGT method as described in Lefranc, M.-R., The Immunologist, 7, 132-136 (1999). CDR1 consists of positions 27 to 38, CDR2 consists of positions 56 to 65, CDR3 for germline V genes consists of positions 105 to 116, CDR3 for rearranged V-J genes or V-D-J genes consists of positions 105 to 117 (position preceding J-PHE or J-TRP 118) with gaps at the top of the loop for a rearranged CDR3-IMGT with less than 13 amino acids, or with additional positions 112.1, 111.1, 112.2, 111.2, etc. for rearranged CDR3-IMGT with more than 13 amino acids. The provisions given in this paragraph correspond to the IMGT numbering described in Lefranc, M.-R., The Immunologist, 7, 132-136 (1999).

Antibodies that specifically bind to PD-L1 included in the fusion proteins of the invention may contain an Fc moiety that allows the in vivo half-life of the bispecific binding molecule of the invention to be extended. In some embodiments, such Fc portion is preferably of human origin, more preferably a human IgG1 or IgG4 antibody Fc portion, even more preferably a genetically modified human IgG1 or IgG4 Fc portion with activated or suppressed effector functions. In some embodiments, suppression of effector functions may be preferable to activation of effector functions. In some embodiments, such an F portion is genetically modified to suppress effector functions through mutation(s) at positions 234 and/or 235, numbered according to the EU Kabat index (Johnson and Wu, Nucleic Acids Res, 2000). In some embodiments, agents may be administered to suppress effector functions.

- 19 046634 we have mutations in positions F234 and L235 of the proposed anti-PD-L1 antibody. In other embodiments, mutations may be introduced at positions D265 and P329 of the proposed anti-PD-L1 antibody to suppress effector function. The numbering for both sets of these possible mutations is given according to the EU Kabat index (Shields et al., J Biol Chem, 2001).

Various methods for producing antibodies and fragments thereof are well known in the art and are described, for example, in Altshuler et al. (2010). Thus, for example, polyclonal antibodies can be obtained from the blood of an animal after immunization with an antigen in a mixture with additives and adjuvants, and monoclonal antibodies can be obtained by any method that provides antibodies produced by cultures of stable cell lines. Examples of such methods are described, for example, by Harlow and Lane (1999), (1988), and include hybridoma technology, first described by Kohler and Milstein, 1975, triome technology, human B cell hybridoma technology (see, for example, Li et al ., Proc Natl Acad Sci USA, 2006, Kozbor and Roder, Immunol Today, 1983) and EBV-hybridoma technology for the production of human monoclonal antibodies (Cole et al., Cancer Res, 1984). In addition, recombinant antibodies can be produced from monoclonal antibodies or can be produced de novo using various display methods such as phage, ribosomal, mRNA or cellular display. In some embodiments, a suitable system for expressing recombinant (humanized) antibodies or fragments thereof may be selected, for example, from bacterial, yeast, insect, mammalian, or transgenic animal or plant cell lines, see, for example, US Pat. No. 6,080,560; Holliger and Hudson, Nat Biotechnol, 2005). In addition, methods described for the production of single chain antibodies (see, inter alia, US Pat. No. 4,946,778) can be adapted to produce single chain antibodies specific for the target of this invention. Surface plasmon resonance used in the BIAcore system can be used to improve the effectiveness of phage antibodies.

C. Exemplary lipocalin muteins according to the invention.

Lipocalins are protein binding molecules that have naturally evolved to bind ligands. Lipocalins are found in many organisms, including vertebrates, insects, plants and bacteria. Members of the lipocalin family of proteins (Pervaiz and Brew, FASEB J, 1987) are typically small secreted proteins and have a single polypeptide chain. They are characterized by a number of different molecular recognition properties: their binding to various, predominantly hydrophobic small molecules (such as retinoids, fatty acids, cholesterols, prostaglandins, biliverdins, pheromones, flavoring agents and odorants) and their binding to certain cell surface receptors and their formation of macromolecular complexes. Although in the past they were classified primarily as transport proteins, it is now clear that lipocalins have a variety of physiological functions. These include roles in retinol transport, olfaction, pheromone signaling, and prostaglandin synthesis. Lipocalins are also involved in regulating the immune response and mediating cellular homeostasis (reviewed in, for example, Flower et al., Biochim Biophys Acta, 2000, Flower, Biochem J, 1996).

Lipocalins have an unusually low level of overall sequence conservation, often with less than 20% sequence identity. In contrast, their general folding pattern is highly conservative. The central part of the lipocalin structure consists of a single eight-stranded antiparallel β-sheet forming a closed figure representing a continuously hydrogen-bonded β-cylinder. This β-cylinder forms the central cavity. One end of the barrel is sterically blocked by an N-terminal peptide segment that extends through its bottom, as well as by three peptide loops connecting the β-strands. The other end of the p-barrel is open to the solvent and surrounds the target binding site, which is formed by four flexible peptide loops (AB, CD, EF and GH). It is the diversity of loops within the otherwise rigid lipocalin framework that gives rise to different binding mechanisms, each capable of accommodating targets of different size, shape and chemical nature (reviewed, for example, in Skerra, Biochim Biophys Acta, 2000, Flower et al., Biochim Biophys Acta, 2000, Flower, Biochem J, 1996).

The lipocalin mutein of the present invention may be any lipocalin mutein. Examples of suitable lipocalins (also sometimes referred to as reference lipocalin, wild-type lipocalin, reference protein scaffolds, or simply scaffolds) whose mutein may be used include, but are not limited to, tear lipocalin (lipocalin-1, Tlc, or von Ebner gland protein), retinol -binding protein, neutrophil prostaglandin E-lipocalin-type synthase, β-lactoglobulin, bilin-binding protein (BBP), apolipoprotein D (APOD), neutrophil gelatinase-associated lipocalin (NGAL), α2microglobulin-associated protein (A2m), 24p3/uterocalin (24p3), von Ebner gland protein 1 (VEGP 1), von Ebner gland protein 2 (VEGP 2) and major allergen Can f 1 (ALL-1). In related embodiments, the lipocalin mutein is derived from the group of lipocalins consisting of human tear lipocalin (hTIc), human neutrophil gelatinase-associated lipocalin (hNGAL),

- 20 046634 human apolipoprotein D (hAPOD) and bilin binding protein from Pieris brassicae.

The amino acid sequence of a lipocalin mutein according to the invention may have high sequence identity compared to the reference lipocalin (or wild type lipocalin) from which it is derived, for example hTIc or hNGAL, compared to sequence identity to another lipocalin (see also above). In this general context, the amino acid sequence of a lipocalin mutein according to the invention is at least substantially similar to the amino acid sequence of the corresponding reference lipocalin (wild type), provided that the alignment may utilize gaps (as defined herein) that result from additions or deletions amino acids. The corresponding lipocalin mutein sequence of the invention, while being substantially similar to the sequences of the corresponding reference lipocalin (wild type), in some embodiments has at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 82%, at least 85%, at least 87%, at least 90% identity, including at least 95% sequence identity to the corresponding lipocalin. However, the lipocalin mutein of the invention may of course contain substitutions, as described herein, that impart the ability to bind CD137 to the lipocalin mutein.

Typically, a lipocalin mutein contains one or more mutated amino acid residues - relative to the amino acid sequence of wild-type lipocalin or reference lipocalin, such as hTIc and hNGAL - in four loops at the open end that contain the ligand-binding pocket and form the entrance to the ligand-binding pocket (see . higher). As explained above, these regions are key to determining the binding specificity of the lipocalin mutein to the desired target. In some embodiments, the lipocalin mutein of the invention may also contain regions of mutated amino acid residues beyond the four loops. In some embodiments, the lipocalin mutein of the invention may contain one or more mutated amino acid residues in one or more of the three peptide loops (designated BC, DE and FG) connecting the beta strands at the closed end of the lipocalin. In some embodiments, a mutein derived from tear lipocalin, NGAL lipocalin, or a homologue thereof may have 1, 2, 3, 4 or more mutated amino acid residues at any sequence position in the N-terminal region and/or the three BC peptide loops, DE and FG located at the end of the β-barrel structure opposite the natural lipocalin binding pocket. In some embodiments, a mutein derived from tear lipocalin, NGAL lipocalin, or a homolog thereof may not have mutated amino acid residues in the DE peptide loop located at the end of the β-barrel structure compared to the wild-type tear lipocalin sequence.

In some embodiments, the lipocalin mutein of the invention may include one or more, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or even more mutated amino acid residues compared to the amino acid sequence of the corresponding reference lipocalin (wild type), provided that such lipocalin mutein must be able to bind to CD137. In some embodiments, the lipocalin mutein of the invention includes at least two, including 2, 3, 4, 5 or even more, mutated amino acid residues, wherein the native amino acid residue of the corresponding reference lipocalin (wild type) is replaced by an arginine residue.

Any type and number of mutations are contemplated, including substitutions, deletions and insertions, as long as the resulting lipocalin mutein retains its ability to bind CD137 and/or has at least 60%, e.g., at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, or higher sequence identity to the amino acid sequence of reference lipocalin (wild type), for example, mature hTIc or mature hNGAL.

In some embodiments, the replacement is a conservative replacement. In some embodiments, the substitution is a non-conservative substitution or one or more of the following exemplary substitutions.

In particular, to determine whether an amino acid residue in the amino acid sequence of a lipocalin mutein that differs from the reference lipocalin (wild type) corresponds to a particular position in the amino acid sequence of the reference lipocalin (wild type), one of ordinary skill in the art may use tools and methods well known in the art. , for example, alignments, manually or using computer programs such as BLAST 2.0, which stands for Basic Local Alignment Search Tool, or ClustalW, or any other suitable program that is suitable for generating sequence alignments. Accordingly, the amino acid sequence of the reference lipocalin (wild type) may serve as the subject sequence or reference sequence, while

- 21 046634 time as the amino acid sequence of lipocalin mutein serves as the desired sequence (see also above).

Conservative substitutions are typically the following substitutions, listed according to the amino acid being mutated, each followed by one or more substitutions that may be considered conservative: Ala^Ser, Thr, or Val; Arg^Lys, Gln, Asn or His; Asn^Gln, Glu, Asp or His; Asp^Glu, Gln, Asn or His; Gln^Asn, Asp, Glu or His; Glu^Asp, Asn, Gln or His; His^Arg, Lys, Asn, Gln, Asp or Glu; Ile^Thr, Leu, Met, Phe, Val, Trp, Tyr, Ala or Pro; Leu^Thr, Ile, Val, Met, Ala, Phe, Pro, Tyr or Trp; Lys^Arg. His, Gln or Asn; Met^Thr. Leu, Tyr, Ile, Phe, Val, Ala, Pro or Trp; Phe^Thr, Met, Leu, Tyr, Ile, Pro, Trp, Val or Ala; Ser^Thr, Ala or Val; Thr^Ser, Ala, Val, Ile, Met, Val, Phe, Pro or Leu; Trp^Tyr, Phe, Met, Ile or Leu; Tyr^Trp, Phe, Ile, Leu or Met; Val^Thr, Ile, Leu, Met, Phe, Ala, Ser or Pro. Other substitutions are also acceptable and may be determined empirically or in accordance with other known conservative or non-conservative substitutions. As an additional guide, each of the following groups contains amino acids that can generally be used to determine conservative substitutions for each other:

(a) Alanine (Ala), serine (Ser), threonine (Thr), valine (Val).

(b) Aspartic acid (Asp), glutamic acid (Glu), glutamine (Gln), asparagine (Asn), histidine (His).

(c) Arginine (Arg), lysine (Lys), glutamine (Gln), asparagine (Asn), histidine (His).

(d) Isoleucine (Ile), leucine (Leu), methionine (Met), valine (Val), alanine (Ala), phenylalanine (Phe), threonine (Thr), proline (Pro).

(e) Isoleucine (Ile), leucine (Leu), methionine (Met), phenylalanine (Phe), tyrosine (Tyr), tryptophan (Trp).

If such conservative substitutions result in a change in biological activity, more significant changes, such as the following, or as further described below with reference to amino acid classes, can be made, and products can be screened for the desired characteristics. Examples of such more significant changes are: Ala^Leu or Phe; Arg^Glu; Asn^Ile, Val or Trp; Asp^Met; Cys^Pro; Gln^Phe; Glu^Arg; His^Gly; Ile^Lys, Glu or Gln; Leu^Lys or Ser; Lys^Tyr; Met^Glu; Phe^Glu, Gln or Asp; Trp^Cys; Tyr^Glu or Asp; Val^Lys, Arg, His.

In some embodiments, significant modifications to the physical and biological properties of the lipocalin (mutein) are achieved by selecting substitutions that vary significantly in their effect on maintaining (a) the structure of the polypeptide backbone at the site of the substitution, e.g., a sheet or helical conformation, (b) charge or the hydrophobicity of the molecule at the target site, or (c) the volume of the side chain.

Natural residues are divided into groups based on common side chain properties: (1) hydrophobic: methionine, alanine, valine, leucine, isoleucine; (2) neutral hydrophilic: cysteine, serine, threonine, asparagine, glutamine; (3) acidic: aspartic acid, glutamic acid; (4) basic: histidine, lysine, arginine; (5) residues affecting chain orientation: glycine, proline; and (6) aromatic: tryptophan, tyrosine, phenylalanine. In some embodiments, replacements may involve replacing a member of one of these classes with another class.

Any cysteine residue not involved in maintaining the proper conformation of the corresponding lipocalin can also be replaced, usually with serine, to improve the oxidative stability of the molecule and prevent aberrant cross-linking. Conversely, cysteine linkage(s) can be added to lipocalin to improve its stability.

D. Exemplary CD137-specific lipocalin muteins according to the invention.

As noted above, lipocalin is a polypeptide defined by its supersecondary structure, namely, a supersecondary structural region of a cylindrical β-pleated sheet containing eight β-strands linked in pairs by four loops at one end, thus forming a binding pocket. The present description is not limited to the lipocalin muteins specifically disclosed herein. In this regard, the invention relates to a lipocalin mutein having a supersecondary structural region of a cylindrical β-pleated sheet containing eight β-strands linked in pairs by four loops at one end, thereby forming a binding pocket in which at least one amino acid of each at least three of the four loops have been mutated, and wherein said lipocalin is effective for binding CD137 with detectable affinity.

In some embodiments, the lipocalin muteins disclosed herein may be or contain a mature human tear lipocalin (hTIc) mutein. Mature hTIc mutein may be referred to herein as hTIc mutein. In some other embodiments, the lipocalin mutein disclosed herein is a mature human neutrophil gelatinase-associated lipocalin (hNGAL) mutein. Mature hNGAL mutein may be referred to herein as hNGAL mutein.

- 22 046634

In one aspect, the present invention includes any number of lipocalin muteins derived from reference lipocalin (wild type), preferably derived from mature hTIc or mature hNGAL, that bind CD137 with detectable affinity. In a related aspect, the invention includes various lipocaine muteins capable of activating CD137 downstream signaling pathways by binding to CD137. In this sense, CD137 can be considered as a non-natural target of the reference lipocalin (wild type), preferably hTIc or hNGAL, where the non-natural target refers to a substance that does not bind to the reference lipocalin (wild type) under physiological conditions. By genetically modifying reference lipocalins (wild type) with one or more mutations at specific sequence positions, we have demonstrated that it is possible to achieve high affinity and high specificity for a non-natural target, CD137. In some embodiments, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 or even more triplets of nucleotides encoding specific sequence positions in wild-type lipocalins may be subject to random substitution mutagenesis subsets of nucleotide triplets at these positions to create a lipocalin mutein capable of binding CD137.

In some embodiments, the lipocalin muteins of the invention may have a mutation, including substitution, deletion, and insertion of amino acid residue(s) at one or more sequence positions corresponding to the linear polypeptide sequence of the reference lipocalin, preferably hTIc or hNGAL. In some embodiments, the number of amino acid residues of a lipocalin mutein according to the invention that are mutated from the amino acid sequence of a reference lipocalin, preferably hTIc or hNGAL, is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14, 15, 16, 17, 18, 19, 20 or more, for example, 25, 30, 35, 40, 45 or 50, where 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 or 11 is preferred, and 9, 10 or 11 is even more preferred. However, it is preferred that the lipocalin mutein of the invention remains capable of binding CD137.

In some embodiments, the lipocalin mutein of the present invention may lack 1, 2, 3, 4 or more amino acids at its N-terminus and/or 1, 2 or more amino acids at its C-terminus, compared to the corresponding reference lipocalin (wild type). type); for example, SEQ ID NO: 34-40. In some embodiments, the present invention covers hTIc muteins as defined above, wherein the first one, two, three, or four N-terminal amino acid residues of the mature hTIc sequence (His-His-Leu-Leu; positions 1-4) and/or the last one or two C-terminal amino acid residues (Ser-Asp; positions 157-158) of the mature hTIc linear polypeptide sequence have been deleted (eg, SEQ ID NO: 34-40). In some embodiments, the present invention covers hNGAL muteins, as defined above, in which amino acid residues (Lys-Asp-Pro, positions 46-48) of the linear polypeptide sequence of mature hNGAL have been deleted (SEQ ID NO: 45). In addition, the lipocalin mutein of the invention may include the (natural) wild-type amino acid sequence of the reference lipocalin (wild-type), preferably hTIc or hNGAL, outside of mutated amino acid sequence positions.

In some embodiments, one or more mutated amino acid residues included in the lipocalin mutein of the invention do not significantly impede or interfere with the binding activity to the selected target and folding of the mutein. Such mutations, including substitution, deletion and insertion, can be performed at the DNA level using established standard methods (Sambrook and Russell, 2001, Molecular cloning: a laboratory manual). In some embodiments, the mutated amino acid residue(s) at one or more sequence positions corresponding to the linear polypeptide sequence of the reference lipocalin (wild type), preferably hTIc or hNGAL, is introduced by random mutagenesis by replacing the triplet(s) of nucleotides encoding the corresponding reference sequence positions lipocalin, a subset of nucleotide triplets.

In some embodiments, a proposed lipocalin mutein that binds CD137 with detectable affinity may include at least one amino acid substitution of a native cysteine residue with another amino acid, such as a serine residue. In some embodiments, a lipocalin mutein that binds CD137 with detectable affinity may include one or more non-native cysteine residues replacing one or more amino acids of a reference lipocalin (wild type), preferably hTIc or hNGAL. In some embodiments, the lipocalin mutein of the invention includes at least two amino acid substitutions of a native amino acid with a cysteine residue to thereby form one or more cysteine bridges. In some embodiments, said cysteine bridge may connect at least two loop regions. The definition of these regions is used herein according to (Biochim Biophys Acta, 2000), Flower (1996) and Breustedt et al. (2005).

Typically, the lipocalin mutein of the invention may be at least 70%, including at least about 80%, such as at least about 85%

- 23 046634 amino acid sequence identity with the amino acid sequence of mature hTIc (SEQ ID NO: 1) or mature hNGAL (SEQ ID NO: 2).

In some aspects of the present invention, CD137-binding hTIc muteins are provided. In this regard, the invention provides one or more hTIc muteins capable of binding CD137 with an affinity, measured by KD, of approximately 300 nM, 200 nM, 150 nM, 100 nM, or lower. In some embodiments, the proposed hTIc muteins are capable of binding CD137 with an EC ₅₀ value of approximately 250 nM, 150 nM, 100 nM, 50 nM, 20 nM, or even lower. In some other embodiments, hTIc CD137-binding muteins may be cross-reactive with cynomolgus CD137 (cyCD137).

In some embodiments, the hTIc mutein of the invention may interfere with the binding of CD137L to CD137.

In some embodiments, the proposed hTIc muteins may contain a mutated amino acid residue at one or more positions corresponding to positions 5, 26-31, 33-34, 42, 46, 52, 56, 58, 60-61, 65, 71, 85, 94, 101, 104-106, 108, 111, 114, 121, 133, 148, 150 and 153 linear polypeptide sequence of mature hTIc (SEQ ID NO: 1).

In some embodiments, the proposed hTIc muteins may contain a mutated amino acid residue at one or more positions corresponding to positions 26-34, 5558, 60-61, 65, 104-106, and 108 of the linear mature hTIc polypeptide sequence (SEQ ID NO: 1).

In some embodiments, the proposed hTIc muteins may further comprise a mutated amino acid residue at one or more positions corresponding to positions 101, 111, 114 and 153 of the linear mature hTIc polypeptide sequence (SEQ ID NO: 1).

In some embodiments, the proposed hTIc muteins may contain a mutated amino acid residue at 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 , 20, 21, 22, 23, 24, 25, 26 or even more provisions corresponding to provisions 5, 26-31, 33-34, 42, 46, 52, 56, 58, 6061, 65, 71, 85, 94 , 101, 104-106, 108, 111, 114, 121, 133, 148, 150 and 153 linear polypeptide sequence of mature hTIc (SEQ ID NO: 1). In some preferred embodiments, the proposed hTIc muteins are capable of binding CD137, in particular human CD137.

In some embodiments, the proposed hTIc muteins may contain a mutated amino acid residue at positions 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, or even more, corresponding to positions 26-34, 55-58 , 60-61, 65, 104-106 and 108 linear polypeptide sequence of mature hTIc (SEQ ID NO: 1). In some preferred embodiments, the proposed hTIc muteins are capable of binding CD137, in particular human CD137.

In some embodiments, the lipocalin mutein of the invention may include at least one amino acid substitution of a native cysteine residue, such as a serine residue. In some embodiments, the hTIc mutein of the invention comprises an amino acid replacement of a native cysteine residue at positions corresponding to positions 61 and/or 153 of the linear polypeptide sequence of mature hTIc (SEQ ID NO: 1) with another amino acid, such as a serine residue. In this context, it should be noted that removal of the structural disulfide bond (at the level of the corresponding naïve nucleic acid library) of wild-type hTIc, which is formed by cysteine residues 61 and 153 (cf. Breustedt et al., J Biol Chem, 2005), has been found to can provide hTIc muteins that are not only stably folded but also capable of binding a selected non-natural target with high affinity. In some embodiments, elimination of the structural disulfide bond may provide the additional benefit of allowing non-natural disulfide bonds to be created or intentionally introduced into the muteins of the invention, thereby increasing the stability of the muteins. However, hTIc muteins that bind CD137 and that have a disulfide bridge formed between Cys 61 and Cys 153 are also part of the present invention.

In some particular embodiments, the hTIc mutein of the invention may include one or more amino acid substitutions Cys 61 ^-Ala, Phe, Lys, Arg, Thr, Asn, Gly, Gln, Asp, Asn, Leu, Tyr, Met, Ser, Pro, or Trp and/or Cys 153^-Ser or Ala, at positions corresponding to positions 61 and/or 153 of the linear polypeptide sequence of mature hTIc (SEQ ID NO: 1).

In some embodiments, either two or all three cysteine codons at positions corresponding to positions 61, 101 and 153 of the mature hTIc linear polypeptide sequence (SEQ ID NO: 1) are replaced by a codon of another amino acid. Additionally, in some embodiments, the hTIc mutein of the invention comprises an amino acid replacement of the native cysteine residue at position corresponding to position 101 of the mature hTIc linear polypeptide sequence (SEQ ID NO: 1) with a serine or histidine residue.

In some embodiments, the mutein of the invention comprises an amino acid substitution of a native amino acid with a cysteine residue at positions corresponding to positions 28 or 105 of the linear polypeptide sequence of mature hTIc (SEQ ID NO: 1). Additionally, in some embodiments, the mutein of the invention comprises an amino acid substitution of the native

- 24 046634 arginine residue at a position corresponding to position 111 of the linear polypeptide sequence of mature hTIc (SEQ ID NO: 1), a proline residue. Additionally, in some embodiments, the mutein of the invention comprises an amino acid substitution of a native lysine residue at a position corresponding to position 114 of the linear polypeptide sequence of mature hTIc (SEQ ID NO: 1) with a tryptophan or glutamic acid residue.

In some embodiments, the proposed CD137-binding hTIc muteins may contain at one or more positions corresponding to positions 5, 26-31, 33-34, 42, 46, 52, 56, 58, 60-61, 65, 71, 85, 94, 101, 104-106, 108, 111, 114, 121, 133, 148, 150 and 153 of the mature hTIc linear polypeptide sequence (SEQ ID NO: 1), one or more of the following mutated amino acid residues: Ala 5—Val or Thr; Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Thr 42—Ser; Gly 46—Asp; Lys 52—Glu; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61 Ala; Lys 65—Arg or Asn; Thr 71—Ala; Val 85—Asp; Lys 94—Arg or Glu; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Lys 121—Glu; Ala 133—Thr; Arg 148—Ser; Ser 150—Ile; and Cys 153—Ser. In some embodiments, the hTIc mutein of the invention comprises two or more, for example, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 , 21, 22, 23, 24, 25, 26 or more, or even all of the mutated amino acid residues at these positions in the mature hTIc sequence (SEQ ID NO: 1).

In some embodiments, the proposed CD137-binding hTIc muteins may contain one of the following sets of mutated amino acid residues compared to the linear polypeptide sequence of mature hTIc (SEQ ID NO: 1):

(a) Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; and Cys 153—Ser;

(b) Ala 5—Thr; Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Lys 65—Arg; Val 85—Asp; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Lys 121—Glu; Ala 133- —Thr; and Cys 153- —Ser;

(c) Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Lys 65—Asn; Lys 94—Arg; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Lys 121—Glu; Ala 133—Thr; and Cys 153—Ser;

(d) Ala 5—Val; Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Lys 65—Arg; Lys 94—Glu; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Lys 121—Glu; Ala 133—Thr; and Cys 153—Ser;

(e) Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Thr 42—Ser; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Ser 150—Ile; and Cys 153—Ser;

(f) Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Lys 52—Glu; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Thr 71—Ala; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Ala 133—Thr; Arg 148—Ser; Ser 150—Ile; and Cys 153—Ser; and (g) Ala 5—Thr; Arg 26—Glu; Glu 27—Gly; Phe 28—Cys; Pro 29—Arg; Glu 30—Pro; Met 31—Trp; Leu 33—Ile; Glu 34—Phe; Gly 46—Asp; Leu 56—Ala; Ser 58—Asp; Arg 60—Pro; Cys 61—Ala; Thr 71—Ala; Cys 101—Ser; Glu 104—Val; Leu 105—Cys; His 106—Asp; Lys 108—Ser; Arg 111—Pro; Lys 114—Trp; Ser 150—Ile; and Cys 153—Ser.

In some embodiments, a residual region, that is, a region other than positions corresponding to positions 5, 26-31, 33-34, 42, 46, 52, 56, 58, 60-61, 65, 71, 85, 94, 101 , 104-106, 108, 111, 114, 121, 133, 148, 150 and 153 of the mature hTIc linear polypeptide sequence (SEQ ID NO: 1), in the hTIc mutein according to the invention may contain the (natural) wild-type amino acid sequence of the linear polypeptide sequence mature hTIc outside the mutated amino acid sequence positions.

In some embodiments, the hTIc mutein of the invention has at least 70% sequence identity or at least 70% sequence homology to the mature hTIc sequence (SEQ ID NO: 1). As an illustrative example, the mutein of SEQ ID NO: 34 has an amino acid sequence identity or sequence homology of approximately 84% with the amino acid sequence of mature hTIc.

In some embodiments, the hTIc mutein of the invention comprises the amino acid sequence set forth in any of SEQ ID NOs: 34-40, or a fragment or variant thereof.

In some embodiments, the hTIc mutein of the invention has at least

- 25 046634 at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 98% or higher sequence identity with an amino acid sequence selected from the group consisting from SEQ ID NO: 34-40.

The present invention also includes structural homologues of an hTIc mutein having an amino acid sequence selected from the group consisting of SEQ ID NOs: 34-40, wherein the structural homologues have a homology or amino acid sequence identity to said hTIc mutein of greater than about 60%, preferably greater than 65% %, greater than 70%, greater than 75%, greater than 80%, greater than 85%, greater than 90%, greater than 92%, and most preferably greater than 95%.

In some aspects of the present invention, CD137-binding hNGAL muteins are provided. In this regard, the invention provides one or more hNGAL muteins capable of binding CD137 with an affinity, measured by K _D , of about 800 nM, 700 nM, 200 nM, 140 nM, 100 nM or lower, preferably about 70 nM, 50 nM, 30 nM, 10 nM, 5 nM, 2 nM, or even lower. In some embodiments, the proposed hNGAL muteins are capable of binding CD137 with an EC ₅₀ value of approximately 1000 nM, 500 nM, 100 nM, 80 nM, 50 nM, 25 nM, 18 nM, 15 nM, 10 nM, 5 nM, or lower.

In some embodiments, hNGAL CD137-binding muteins may be cross-reactive with cynomolgus CD137. In some embodiments, the proposed hNGAL muteins are capable of binding cynomolgus CD137 with an affinity, measured by KD, of about 50 nM, 20 nM, 10 nM, 5 nM, 2 nM, or even lower. In some embodiments, the proposed hNGAL muteins are capable of binding cynomolgus CD137 with an EC ₅₀ value of approximately 100 nM, 80 nM, 50 nM, 30 nM, or even lower.

In some embodiments, the hNGAL mutein of the invention may interfere with or compete for the binding of CD137L to CD137. In some other embodiments, the hNGAL mutein of the invention may be capable of binding CD137 in the presence of CD137L and/or binding the CD137/CD137L complex.

In some embodiments, the proposed hNGAL muteins may contain a mutated amino acid residue at one or more positions corresponding to positions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 83, 87, 94, 96, 100, 103, 106, 125, 127, 132 and 134 linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2).

In some embodiments, the proposed hNGAL muteins may contain a mutated amino acid residue at 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 , 20, 21 or even more provisions corresponding to provisions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 83, 87, 94, 96, 100, 103 , 106, 125, 127, 132 and 134 linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2). In some preferred embodiments, the proposed hNGAL muteins are capable of binding CD137, in particular human CD137.

In some embodiments, the proposed hNGAL muteins may contain a mutated amino acid residue at one or more positions corresponding to positions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 87, 96, 100, 103, 106, 125, 127, 132 and 134 linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2). In some preferred embodiments, the proposed hNGAL muteins are capable of binding CD137, in particular human CD137.

In some embodiments, the proposed hNGAL muteins may contain a mutated amino acid residue at one or more positions corresponding to positions 36, 87 and 96 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2), and at one or more positions corresponding to positions 28, 40 -41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 83, 94, 100, 103, 106, 125, 127, 132 and 134 linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2).

In some other embodiments, the proposed hNGAL muteins may contain a mutated amino acid residue at one or more positions corresponding to positions 20, 25, 28, 33, 36, 40-41, 44, 49, 52, 59, 68, 70-73, 77 -82, 87, 92, 96, 98, 100, 101, 103, 122, 125, 127, 132 and 134 linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2).

In other embodiments, the proposed hNGAL muteins may contain a mutated amino acid residue at one or more positions 36, 40, 41, 49, 52, 68, 70, 72, 73, 77, 79, 81, 96, 100, 103, 125, 127 , 132 and 134 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2), and at one or more positions corresponding to positions 20, 25, 33, 44, 59, 71, 78, 80, 82, 87, 92, 98, 101 and 122 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2).

In some embodiments, the lipocalin mutein of the invention may contain at least one amino acid substitution of a native cysteine residue, for example, a serine residue. In some embodiments, the hNGAL mutein of the invention may comprise an amino acid replacement of the native cysteine residue at positions corresponding to positions 76 and/or 175 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2) with another amino acid, such as a serine residue. In this context, it should be noted that it has been found that the removal of a structural disulfide bond (at the level of the corresponding naive nucleic acid library

- 26 046634 slot) wild-type hNGAL, which is formed by cysteine residues 76 and 175 (cf. Breustedt et al., J Biol Chem, 2005), can provide hNGAL muteins that are not only stably folded, but also capable of binding a selected non-natural target to high affinity. In some embodiments, elimination of the structural disulfide bond may provide the additional benefit of allowing non-natural disulfide bonds to be created or intentionally introduced into the muteins of the invention, thereby increasing the stability of the muteins. However, hNGAL muteins that bind CD137 and that have a disulfide bridge formed between Cys 76 and Cys 175 are also part of the present invention.

In some embodiments, the proposed CD137-binding hNGAL muteins may contain at one or more positions corresponding to positions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 83, 87, 94, 96, 100, 103, 106, 125, 127, 132 and 134 of the mature hNGAL linear polypeptide sequence (SEQ ID NO: 2), one or more of the following mutated amino acid residues. Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg or Lys; Gln 49>Val, Ile, His, Ser or Asn; Tyr 52>Met; Asn 65>Asp; Ser 68>Met, Ala or Gly; Leu 70>Ala, Lys, Ser or Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Met, Arg, Thr or Asn; Trp 79>Ala or Asp; Arg 81>Met, Trp or Ser; Phe 83>Leu; Cys 87>Ser; Leu 94>Phe; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu and Lys 134>Tyr. In some embodiments, the hNGAL mutein of the invention contains two or more, for example 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, even more, for example 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, or all mutated amino acid residues at these positions in the mature hNGAL sequence (SEQ ID NO: 2).

In some embodiments, the proposed CD137-binding muteins hNGAL may contain at one or more positions corresponding to positions 20, 25, 28, 33, 36, 40-41, 44, 49, 52, 59, 68, 70-73, 77- 82, 87, 92, 96, 98, 100, 101, 103, 122, 125, 127, 132 and 134 of the mature hNGAL linear polypeptide sequence (SEQ ID NO: 2), one or more of the following mutated amino acid residues. Gln 20>Arg; Asn 25>Tyr or Asp; Gln 28>His; Val 33>Ile; Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Glu 44>Val or Asp; Gln 49>His; Tyr 52>Ser or Gly; Lys 59>Asn; Ser 68>Asp; Leu 70>Met; Phe 71>Leu; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln or His; Tyr 78>His; Trp 79>Ile; Ile 80>Asn; Arg 81>Trp or Gln; Thr 82>Pro; Cys 87>Ser; Phe 92>Leu or Ser; Asn 96>Phe; Lys 98>Arg; Tyr 100>Asp; Pro 101>Leu; Leu 103>His or Pro; Phe 122>Tyr; Lys 125>Ser; Ser 127>Ile; Tyr 132 >Trp; and Lys 134 >Gly

In some embodiments, the proposed CD137-binding hNGAL muteins may contain at one or more positions corresponding to positions 36, 40, 41, 49, 52, 68, 70, 72, 73, 77, 79, 81, 96, 100, 103, 125, 127, 132 and 134 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2), one or more of the following mutated amino acid residues. Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Ser or Gly; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln or His; Trp 79>Ile; Arg 81>Trp or Gln; Asn 96>Phe; Tyr 100>Asp; Leu 103>His or Pro; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly. In some embodiments, the proposed CD137-binding hNGAL muteins may be present at one or more positions corresponding to positions 20, 25, 33, 44, 59, 71, 78, 80, 82, 87, 92, 98, 101, and 122 of the linear polypeptide sequence mature hNGAL (SEQ ID NO: 2), one or more of the following mutated amino acid residues. Gln 20>Arg; Asn 25>Tyr or Asp; Val 33>Ile; Glu 44>Val or Asp; Lys 59>Asn; Phe 71>Leu; Tyr 78>His; Ile 80>Asn; Thr 82>Pro; Phe 92>Leu or Ser; Lys 98>Arg; Pro 101>Leu; and Phe 122>Tyr.

In some embodiments, the proposed CD137-binding hNGAL muteins may contain one of the following sets of mutated amino acid residues compared to the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2):

(a) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Lys; Gln 49>Asn; Tyr 52>Met; Ser 68>Gly; Leu 70>Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Thr; Trp 79>Ala; Arg 81>Ser; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr;

(b) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg; Gln 49>Ile; Tyr 52>Met; Asn 65>Asp; Ser 68>Met; Leu 70>Lys; Arg 72>Asp; Lys 73>Asp; Asp 77>Met; Trp 79>Asp; Arg 81>Trp; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr;

(c) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg; Gln 49>Asn; Tyr 52>Met; Asn 65>Asp; Ser 68>Ala; Leu 70>Ala; Arg 72>Asp; Lys 73>Asp; Asp 77>Thr; Trp 79>Asp; Arg 81>Trp; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr;

(d) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Lys; Gln 49>Asn; Tyr 52>Met; Asn 65>Asp; Ser 68>Ala; Leu 70>Ala; Arg 72>Asp; Lys 73>Asp; Asp 77>Thr; Trp 79>Asp; Arg 81>Trp; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys

- 27 046634

134 >Tyr;

(e) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Lys; Gln 49>Ser; Tyr 52>Met; Asn 65>Asp; Ser 68>Gly; Leu 70>Ser; Arg 72>Asp; Lys 73>Asp; Asp 77>Thr; Trp 79>Ala; Arg 81>Met; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr;

(f) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Lys; Gln 49>Val; Tyr 52>Met; Asn 65>Asp; Ser 68>Gly; Leu 70>Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Arg; Trp 79>Asp; Arg 81>Ser; Cys 87>Ser; Leu 94>Phe; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr;

(g) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg; Gln 49>His; Tyr 52>Met; Asn 65>Asp; Ser 68>Gly; Leu 70>Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Thr; Trp 79>Ala; Arg 81>Ser; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr;

(h) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Lys; Gln 49>Asn; Tyr 52>Met; Asn 65>Asp; Ser 68>Gly; Leu 70>Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Thr; Trp 79>Ala; Arg 81>Ser; Phe 83>Leu; Cys 87>Ser; Leu 94>Phe; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr; or (i) Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg; Gln 49>Ser; Tyr 52>Met; Asn 65>Asp; Ser 68>Ala; Leu 70>Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Asn; Trp 79>Ala; Arg 81>Ser; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134>Tyr.

In some additional embodiments, in the residual region, that is, a region other than positions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 83, 87, 94, 96, 100, 103, 106, 125, 127, 132 and 134 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2), the hNGAL mutein of the invention may include the (natural) wild-type amino acid sequence of mature hNGAL outside of mutated amino acid sequence positions.

In some other embodiments, the proposed CD137-binding hNGAL muteins may contain one of the following sets of mutated amino acid residues compared to the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2):

(a) Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Ser; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Trp; Asn 96>Phe; Tyr 100>Asp; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(b) Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Ser; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Trp; Phe 92>Leu; Asn 96>Phe; Lys 98>Arg; Tyr 100>Asp; Pro 101>Leu; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(c) Asn 25>Tyr; Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Gly; Ser 68>Asp; Leu 70>Met; Phe 71>Leu; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Gln; Phe 92>Ser; Asn 96>Phe; Tyr 100>Asp; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(d) Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Gly; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Tyr 78>His; Trp 79>Ile; Arg 81>Trp; Phe 92>Leu; Asn 96>Phe; Tyr 100>Asp; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(e) Asn 25>Asp; Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Gly; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Trp; Phe 92>Leu; Asn 96 >Phe; Tyr 100>Asp; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(f) Val 33>Ile; Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Gly; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Trp; Phe 92>Leu; Asn 96>Phe; Tyr 100>Asp; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(g) Gln 20>Arg; Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Glu 44>Val; Gln 49>His; Tyr 52>Gly; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Trp; Phe 92>Leu; Asn 96>Phe; Tyr 100>Asp; Leu 103>His; Phe 122>Tyr; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(h) Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Ser; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Ile 80>Asn; Arg 81>Trp; Thr 82>Pro; Asn 96>Phe; Tyr 100>Asp; Pro 101>Leu; Leu 103>Pro; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly;

(i) Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Gln 49>His; Tyr 52>Gly; Lys 59>Asn; Ser 68>Asp; Leu 70>Met; Arg 72>Leu; Lys 73>Asp; Asp 77>Gln; Trp 79>Ile; Arg 81>Trp; Phe 92>Leu; Asn 96>Phe; Tyr 100>Asp; Leu 103>His; Lys 125>Ser; Ser 127>Ile; Tyr 132>Trp; and Lys 134>Gly; and (j) Leu 36>Met; Ala 40>Asn; Ile 41>Leu; Glu 44>Asp; Gln 49>His; Tyr 52>Ser; Ser 68>Asp; Leu

- 28 046634

70^Met; Phe 71^-Leu; Arg 72^-Leu; Lys 73^-Asp; Asp 77^-His; Trp 79^>Ile; Arg 81^-Trp; Phe 92^Leu; Asn 96^-Phe; Tyr 100^ Asp; Leu 103^-His; Lys 125^-Ser; Ser 127^-Ile; Tyr 132^-Trp; and Lys 134^-Gly.

In some embodiments, in the residual region, that is, a region other than positions 20, 25, 28, 33, 36, 40-41, 44, 49, 52, 59, 68, 70-73, 77-82, 87, 92 96, 98, 100, 101, 103, 122, 125, 127, 132 and 134 of the linear polypeptide sequence of mature hNGAL (SEQ ID NO: 2), the hNGAL mutein of the invention may include the (natural) wild-type amino acid sequence of mature hNGAL outside of the mutated positions of the amino acid sequence.

In some embodiments, the hNGAL mutein of the invention has at least 70% sequence identity or at least 70% sequence homology to the mature hNGAL sequence (SEQ ID NO: 2). As an illustrative example, the mutein of SEQ ID NO: 42 has an amino acid sequence identity or sequence homology of approximately 87% with the amino acid sequence of mature hNGAL.

In some embodiments, the hNGAL mutein of the invention comprises the amino acid sequence set forth in any of SEQ ID NOs: 41-59, or a fragment or variant thereof.

In some embodiments, the hNGAL mutein of the invention has at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 98%, or greater sequence identity to an amino acid sequence selected from the group consisting of SEQ ID NO: 41-59.

The present invention also includes structural homologs of an hNGAL mutein having an amino acid sequence selected from the group consisting of SEQ ID NOs: 41-59, wherein the structural homologs have homology or amino acid sequence identity to said hNGAL mutein of approximately 60%, preferably greater than 65% , greater than 70%, greater than 75%, greater than 80%, greater than 85%, greater than 90%, greater than 92%, and most preferably greater than 95%.

In some embodiments, the present invention provides a lipocalin mutein that binds CD137 with an affinity, measured by a _KD , of about 5 nM or less, wherein the lipocalin mutein has at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 98% or greater sequence identity to the amino acid sequence of SEQ ID NO: 42.

In some embodiments, the lipocalin mutein of the present invention may contain a heterologous amino acid sequence at its N- or C-terminus, preferably the C-terminus, such as a Strep II tag (SEQ ID NO: 12) or a cleavage site sequence for certain restriction enzymes, not affecting the biological activity of lipocalin mutein (binding to its target, for example, CD137).

In some embodiments, additional modifications may be introduced into the lipocalin mutein in order to modulate certain characteristics of the mutein, for example, to improve folding stability, serum stability, protein stability, or water solubility, or to reduce the tendency to aggregate, or to impart mutin new characteristics. In some embodiments, the modification(s) may result in modulation of two or more (eg, 2, 3, 4, 5, 6, 7, 8, 9, or 10) characteristics of the proposed mutein.

For example, one or more positions in the amino acid sequence of the lipocalin mutein can be mutated to introduce new reactive groups, for example, to conjugate with other compounds such as polyethylene glycol (PEG), hydroxyethyl starch (HES), biotin, peptides or proteins, or to form non-occurring in the nature of disulfide bonds. A conjugate such as PEG and HES may, in some cases, increase the serum half-life of the corresponding lipocalin mutein.

In some embodiments, a reactive group of a lipocalin mutein may be naturally present in its amino acid sequence, for example, naturally occurring cysteine residues in said amino acid sequence. In some other embodiments, such a reactive group may be introduced through mutagenesis. In the event that a reactive group is introduced by mutagenesis, one possibility is to mutate the amino acid at the appropriate position to a cysteine residue. Examples of such possible mutations to introduce a cysteine residue into the amino acid sequence of the hTIc mutein include Thr 40^-C'ys substitutions. Glu 73^-Cys, Arg 90^-C’ys. Asp 95^-C'ys and Glu 131^-C'ys wild-type hTIc sequences (SEQ ID NO: 1). Examples of such possible mutations to introduce a cysteine residue into the amino acid sequence of the hNGAL mutein include the introduction of a cysteine residue at one or more sequence positions that correspond to positions 14, 21, 60, 84, 88, 116, 141, 145, 143, 146, or 158 of the hNGAL sequence wild type (SEQ ID NO: 2). The resulting thiol moiety can be used to PEGylate or HESylate a mutein, for example to increase the serum half-life of the corresponding lipocalin mutein.

In some embodiments, to provide suitable amino acid side chains

- 29 046634 as new reactive groups for conjugating one of the above compounds with lipocalin mutein, artificial amino acids can be introduced into the amino acid sequence of lipocalin mutein. Typically, such artificial amino acids are designed to be more reactive and thus facilitate conjugation to the desired compound. Such artificial amino acids can be introduced by mutagenesis, for example para-acetylphenylalanine can be introduced using artificial tRNA.

In some embodiments, the lipocalin mutein of the invention is fused at its N-terminus or its C-terminus to a protein, protein domain, or peptide, such as an antibody, signal sequence, and/or affinity tag. In some other embodiments, the lipocalin mutein of the invention is conjugated at its N-terminus or its C-terminus to a protein, protein domain, or peptide partner, such as an antibody, signal sequence, and/or affinity tag.

Affinity tags such as Strep-tag or Strep-tag II (Schmidt et al., J Mol Biol, 1996), c-myc tag, FLAG tag, His tag or HA tag, or proteins such as Glutathione-transferases, which allow easy detection and/or purification of recombinant proteins, are examples of suitable fusion partners. Proteins with chromogenic or fluorescent properties, such as green fluorescent protein (GFP) or yellow fluorescent protein (YFP), are also suitable fusion partners for the lipocalin muteins of the invention. In general, the lipocalin muteins of the invention can be labeled with any suitable chemical or enzyme that directly or indirectly generates a detectable compound or signal in a chemical, physical, optical or enzymatic reaction. For example, a fluorescent or radioactive label may be conjugated to a lipocalin mutein to generate fluorescence or x-rays as a detectable signal. Alkaline phosphatase, horseradish peroxidase and p-galactosidase are examples of enzyme tags (and at the same time optical tags) that catalyze the formation of chromogenic reaction products. In general, all labels commonly used for antibodies (with the exception of labels used exclusively with the sugar moiety in the Fc portion of immunoglobulins) can also be used for conjugation with the lipocalin muteins of the invention.

In some embodiments, the lipocalin mutein of the invention may be fused or conjugated to a moiety that prolongs the serum half-life of the mutein (in this regard, see also International Patent Application Publication WO 2006/056464, which describes such strategies with reference to human lipocalin muteins, neutrophil gelatinase-associated (hNGAL), with binding affinity to CTLA-4). The serum half-life prolonging moiety may be, for example, a PEG molecule, a HES molecule, a fatty acid molecule such as palmitic acid (Vajo and Duckworth, Pharmacol Rev, 2000), an immunoglobulin Fc portion, an immunoglobulin CH3 domain, a CH4 domain immunoglobulin, albumin-binding peptide, albumin-binding protein or transferrin.

In some embodiments, if PEG is used as the conjugation partner, the PEG molecule may be substituted, unsubstituted, linear, or branched. It may also be an activated polyethylene derivative. Examples of suitable compounds are PEG molecules as described in International Patent Application Publication WO 1999/64016, US Patent No. 6,177,074 or US Patent No. 6,403,564 for interferon, or as described for other proteins such as PEG-modified asparaginase, PEGadenosine deaminase ( PEG-ADA) or PEG superoxide dismutase (Fuertges and Abuchowski, Journal of Controlled Release, 1990). The molecular weight of such a polymer, such as polyethylene glycol, may range from about 300 to about 70,000 daltons, including, for example, polyethylene glycol with a molecular weight of about 10,000, about 20,000, about 30,000, or about 40,000 daltons. Moreover, as, for example, described in US Pat. No. 6,500,930 or 6,620,413, carbohydrate oligomers and polymers, such as HES, can be conjugated to the mutein of the invention to increase serum half-life.

In some embodiments, if the Fc portion of an immunoglobulin is used to increase the serum half-life of the lipocalin muteins of the invention, SynFusion™ technology, commercially available from Syntonix Pharmaceuticals, Inc., can be used. (Massachusetts, USA). The use of this Fc-fusion technology allows the creation of long-acting biopharmaceuticals and can, for example, consist of two copies of a mutein linked to the Fc region of an antibody to improve pharmacokinetics, solubility and manufacturing efficiency.

Examples of albumin-binding peptides that can be used to increase the serum half-life of lipocalin mutein include, for example, peptides having the consensus sequence Cys-XaaiXaa ₂ -Xaa ₃ -Xaa4-Cys, where Xaa1 is Asp, Asn, Ser, Thr or Trp; Xaa ₂ is Asn, Gln, His, Ile, Leu or Lys; Xaa ₃ is Ala, Asp, Phe, Trp or Tyr; and Xaa4 is Asp, Gly, Leu, Phe, Ser or Thr, as described in US patent publication 20030069395 or Dennis et al. (2002). Albumin binding protein, fusion or

- 30 046634 conjugated to a lipocalin mutein to increase serum half-life, may be a bacterial albumin binding protein, an antibody, an antibody fragment including domain antibodies (see, for example, US Pat. No. 6,696,245), or a lipocalin mutein with albumin binding activity . Examples of bacterial albumin-binding proteins include streptococcal protein G (Konig and Skerra, J Immunol Methods, 1998).

In some embodiments, if the albumin binding protein is an antibody fragment, it may be a domain antibody. Domain antibodies (dAbs) are engineered to precisely control biophysical properties and in vivo half-life to create an optimal product safety and efficacy profile. Domain antibodies, for example, are commercially available from Domantis Ltd. (Cambridge, UK and Massachusetts, USA).

In some embodiments, albumin itself (Osborn et al., J Pharmacol Exp Ther, 2002) or a biologically active fragment of albumin can be used as a partner of the lipocalin mutein of the invention to increase serum half-life. The term albumin includes all mammalian albumins, such as human serum albumin, or bovine serum albumin, or rat albumin. Albumin or a fragment thereof can be produced recombinantly, as described in US Patent No. 5,728,553 or European Patent Publications EP0330451 and EP0361991. Accordingly, recombinant human albumin (eg, Recombumin® from Novozymes Delta Ltd., Nottingham, UK) can be conjugated or fused to the lipocalin mutein according to the invention.

In some embodiments, when transferrin is used as a partner to increase the serum half-life of the lipocalin muteins of the invention, the muteins can be genetically fused to the N-terminus or the C-terminus, or both, of the non-glycosylated transferrin. Non-glycosylated transferrin has a half-life of 14-17 days, and a transferrin fusion protein will similarly have an increased half-life. Transferrin as a carrier also provides high bioavailability, biodistribution and stability in the bloodstream. This technology is commercially available from BioRexis (BioRexis Pharmaceutical Corporation, Pennsylvania, USA). Recombinant human transferrin (DeltaFerrin™) for use as a protein stabilizer/partner to increase half-life is also commercially available from Novozymes Delta Ltd. (Nottingham, UK).

Another option for extending the half-life of the lipocalin muteins of the invention is to fuse long, unstructured, flexible glycine-rich sequences (eg, polyglycine with approximately 20-80 consecutive glycine residues) to the N- or C-terminus of the mutein. This approach, disclosed for example in international patent publication WO2007/038619, has also been called rPEG (recombinant PEG).

E. Examples of variants and applications of CD137- and PD-L1-specific fusion proteins.

In some embodiments, fusion proteins of the invention may provide a synergistic effect by dual targeting CD137 and PD-L1. In some embodiments, fusion proteins of the invention may provide a synergistic effect through costimulation of CD137 and blockade of the PD-1/PD-L1 signaling pathway. In some embodiments, fusion proteins of the invention may provide localized antitumor effects by dual targeting CD137 and PD-L1. Thus, there are many possible medical uses of the fusion proteins of the invention.

In some embodiments, the present invention relates to the use of one or more fusion proteins disclosed herein, or one or more compositions containing such fusion proteins, to simultaneously bind CD137 and PD-L1.

The present invention also includes the use of one or more of the described fusion proteins to form a complex with CD137 and/or PD-L1.

Therefore, in one aspect of the invention, the proposed fusion proteins can be used to detect CD137 and PD-L1. Such use may include the steps of bringing one or more of these fusion proteins, under suitable conditions, into contact with a sample suspected of containing CD137 and/or PD-L1, thereby allowing the formation of a complex between the fusion proteins and CD137 and/or PD-L1. L1, and detection of the complex using a suitable signal. The detected signal may be caused by a label, as explained above, or by a change in physical properties due to binding, i.e. formation of the complex as such. One example is surface plasmon resonance, the value of which changes during the binding of binding partners, one of which is immobilized on a surface such as gold foil.

The fusion proteins of the invention can also be used to separate CD137 and/or PD-L1. Such use may include the steps of bringing one or more of these fusion proteins, under suitable conditions, into contact with a sample suspected of containing CD137 and/or PDL1, thereby allowing the formation of a complex between the fusion proteins and CD137 and/or PD-L1, and separating the complex from the sample.

- 31 046634

In some aspects of the present invention, diagnostic and/or assay kits are provided containing one or more fusion proteins of the invention.

In addition to diagnostic use, another aspect of the invention provides pharmaceutical compositions comprising one or more fusion proteins of the invention and a pharmaceutically acceptable excipient.

In addition, some embodiments of the present invention provide fusion proteins that simultaneously bind CD137 and/or PD-L1 for use as antitumor and/or anti-infective agents, as well as immunomodulators. In some embodiments, the fusion proteins of the present invention are used in a method for preventing, ameliorating, or treating human diseases, such as various cancers, including PD-L1 positive cancers. Accordingly, also provided are methods of preventing, ameliorating, or treating human diseases, such as various cancers, including PD-L1 positive cancers, in a subject in need thereof, comprising administering to said subject a therapeutically effective amount of one or more fusion proteins of the invention. .

Examples of cancers treatable with the fusion proteins of the invention include liver cancer, bone cancer, pancreatic cancer, skin cancer, head and neck cancer, breast cancer, lung cancer, cutaneous or intraocular malignant melanoma, kidney cancer, uterine cancer , ovarian cancer, colorectal cancer, colon cancer, rectal cancer, anal cancer, stomach cancer, testicular cancer, uterine cancer, fallopian tube carcinoma, endometrial carcinoma, cervical carcinoma, vaginal carcinoma, vulvar carcinoma, non-Hodgkin's lymphoma, esophageal cancer , small bowel cancer, endocrine system cancer, thyroid cancer, parathyroid cancer, adrenal cancer, soft tissue sarcoma, urethral cancer, penile cancer, childhood solid tumors, lymphocytic lymphoma, bladder cancer, kidney or ureter cancer, renal carcinoma pelvis, central nervous system (CNS) neoplasm, primary CNS lymphoma, tumor angiogenesis, spinal axis tumor, brainstem glioma, pituitary adenoma, Kaposi's sarcoma, epidermoid cancer, squamous cell carcinoma, cancers caused by environmental factors, including those caused asbestos, hematologic malignancies including, for example, multiple myeloma, B-cell lymphoma, Hodgkin's lymphoma/primary mediastinal B-cell lymphoma, non-Hodgkin's lymphomas, acute myeloid lymphoma, chronic myeloid leukemia, chronic lymphocytic leukemia, follicular lymphoma, diffuse large B-cell private lymphoma, Burkitt's lymphoma, immunoblastic large cell lymphoma, B-lymphoblastic progenitor cell lymphoma, mantle cell lymphoma, acute lymphoblastic leukemia, fungoid mycosis, anaplastic large cell lymphoma, T-cell lymphoma and T-lymphoblastic progenitor cell lymphoma, as well as any combinations of these cancers. In some embodiments, the present invention is also applicable to the treatment of metastatic cancers.

In some embodiments, the fusion proteins of the invention can simultaneously target tumor cells that express PD-L1 and activate host immune system lymphocytes in the vicinity of such tumor cells. In some embodiments, the fusion proteins of the invention may enhance T cell targeting antitumor activity, enhance antitumor immunity, and/or have a direct tumor growth inhibitory effect, thereby providing synergistic antitumor results. In some embodiments, the fusion proteins of the invention can activate immune responses in the tumor microenvironment. In some embodiments, the fusion proteins of the invention can reduce the side effects of effector lymphocytes on healthy cells, i.e. non-target toxicity, for example, through local inhibition of oncogene activity and induction of lymphocyte activation.

In some embodiments, the present invention includes the use of a fusion protein of the invention, or a composition comprising a fusion protein according to the invention, to induce a localized lymphocyte response in the vicinity of PD-L1 positive tumor cells. Accordingly, in some embodiments, the present invention provides methods for inducing a localized lymphocyte response in the vicinity of PD-L1-positive tumor cells, comprising the use of one or more fusion proteins of the invention or one or more compositions containing such fusion proteins. Localized means that when T cells simultaneously bind via CD137 and interact with PD-L1 positive tumor cells, the T cells produce cytokines, particularly IL-2 and/or IFN gamma, in the vicinity of the PD-L1 positive cells. Such cytokines reflect the activation of T cells, which may then be able to kill PD-L1 positive cells, either directly or indirectly by recruiting other killer cells such as T cells or NK cells.

In some embodiments, the present invention includes the use of a fusion protein of the invention or a composition comprising such a fusion protein for the costimulation of T cells

- 32 046634 and/or activation of downstream CD137 signaling pathways. Preferably, the proposed fusion protein co-stimulates T cells and/or activates CD137 downstream signaling pathways upon binding to tumor cells that express PD-L1. Accordingly, the present invention provides methods for inducing T lymphocyte proliferation and/or activation of CD137 downstream signaling pathways, preferably upon binding to tumor cells on which PDL1 is expressed, comprising the use of one or more fusion proteins according to the invention and/or one or more compositions, containing such fusion proteins.

In some embodiments, the present invention includes the use of a fusion protein of the invention or a composition comprising such a fusion protein to induce CD137 clustering and activation on T cells and directing such T cells to tumor cells that express PD-L1.

Additional objects, advantages and features of the present invention will become apparent to those skilled in the art upon examination of the following examples and accompanying drawings, which should not be construed in a limiting manner. Thus, it is to be understood that while the present invention is specifically disclosed by way of illustrative embodiments and optional features, modifications and variations may be made by those skilled in the art to the embodiments of the invention disclosed herein, and that such modifications and variations are considered to be within the scope of scope of this invention.

F. Preparation of exemplary proposed fusion proteins specific for CD137 and PD-L1.

In some embodiments, the present invention provides nucleic acid molecules (DNA and RNA) that include nucleotide sequences encoding the proposed fusion proteins. In some embodiments, the invention includes a host cell containing the proposed nucleic acid molecule. Since the degeneracy of the genetic code allows for the substitution of individual codons by other codons specifying the same amino acid, the invention is not limited to a particular nucleic acid molecule encoding a fusion protein as described herein, but rather embraces all nucleic acid molecules that include nucleotide sequences encoding a functional fusion protein. In this regard, the present invention also relates to nucleotide sequences encoding the proposed fusion proteins.

A nucleic acid molecule, such as DNA, is said to be capable of expressing a nucleic acid molecule or capable of expressing a nucleotide sequence if it includes sequence elements containing information relating to the regulation of transcription and/or translation, and such sequences are operably linked to a nucleotide sequence encoding a protein. A functional link is a link in which the regulatory sequence elements and the expressed sequence are linked in such a way as to allow gene expression. The exact nature of the regulatory regions required for gene expression may vary between species, but in general these regions include the promoter, which in prokaryotes contains both the promoter itself, that is, DNA elements that control the initiation of transcription, and DNA elements that, when transcribed, in RNA will signal the initiation of translation. Such promoter regions typically include 5' non-coding sequences involved in transcription and translation initiation, such as -35/-10 boxes and the ShineDalgarno element in prokaryotes or TATA box, CAAT sequences and 5' capping elements in eukaryotes. These regions may also include enhancer or repressor elements, as well as translated signal and leader sequences for targeting the native protein to a specific compartment of the host cell.

In addition, the 3' non-coding sequences may contain regulatory elements involved in transcription termination, polyadenylation, and the like. If, however, these termination sequences do not function satisfactorily in a particular host cell, then they can be replaced by signals that function in that cell.

Therefore, a nucleic acid molecule of the invention may be operably linked to one or more regulatory sequences, such as a promoter sequence, to enable expression of the nucleic acid molecule. In some embodiments, a nucleic acid molecule of the invention includes a promoter sequence and a transcription termination sequence. Suitable prokaryotic promoters are, for example, the tet promoter, the lacUV5 promoter or the T7 promoter. Examples of promoters suitable for expression in eukaryotic cells are the SV40 promoter or the CMV promoter.

In some embodiments, a nucleic acid molecule encoding a lipocalin mutein disclosed herein may be operably linked to another nucleic acid molecule encoding an immunoglobulin of the invention to enable expression of a fusion protein disclosed herein.

In some embodiments, the proposed methods may include exposure to

- 33 046634 of at least one nucleic acid molecule encoding a mature hTIc, mutagenesis by nucleotide triplets encoding one or more positions corresponding to positions 5, 26-31, 3334, 42, 46, 52, 56, 58, 60-61, 65 , 71, 85, 94, 101, 104-106, 108, 111, 114, 121, 133, 148, 150 and 153 of the linear hTIc polypeptide sequence (SEQ ID NO: 1), to obtain lipocalin muteins included in the proposed fusions proteins. In some embodiments, the proposed methods may include subjecting at least one nucleic acid molecule encoding mature hNGAL to mutagenesis at nucleotide triplets encoding one or more positions corresponding to positions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81, 83, 87, 94, 96, 100, 103, 106, 125, 127, 132 and 134 linear polypeptide sequence hNGAL (SEQ ID NO: 2), to obtain lipocalin muteins, included in the proposed fusion proteins. In some embodiments, the proposed method may include subjecting at least one nucleic acid molecule encoding mature hNGAL to mutagenesis at nucleotide triplets encoding one or more positions corresponding to positions 20, 25, 28, 33, 36, 40-41, 44, 49, 52, 59, 68, 70-73, 77-82, 87, 92, 96, 98, 100, 101, 103, 122, 125, 127, 132 and 134 linear polypeptide sequence of hNGAL (SEQ ID NO: 2) , to obtain lipocalin muteins included in the proposed fusion proteins.

Additionally, for hTIc muteins or hNGAL muteins of the invention included in fusion proteins, in some embodiments, the native disulfide bond between Cys 61 and Cys 153 or Cys 76 and Cys 175, respectively, may be removed. Accordingly, such muteins can be produced in a cellular compartment that has a reducing redox environment, for example, in the cytoplasm of Gram-negative bacteria.

In addition, with respect to the proposed hTIc muteins or hNGAL muteins of the invention included in fusion proteins, the invention also includes nucleic acid molecules encoding such muteins, which in some embodiments may include one or more additional mutations beyond the specified experimental mutagenesis positions in the sequence . Such mutations are often tolerated or may even be beneficial, for example, if they contribute to improved folding efficiency, serum stability, thermostability, or ligand binding affinity of lipocalin muteins and/or fusion proteins.

In some embodiments, the nucleic acid molecules of the invention may also be part of a vector or any other type of cloning vehicle, such as a plasmid, phagemid, phage, baculovirus, cosmid, or artificial chromosome.

In some embodiments, the proposed nucleic acid molecule may be included in a phagemid. As used herein, a phagemid vector means a vector encoding the intergenic region of a temperate phage, such as M13 or f1, or a functional portion thereof fused to a cDNA of interest. For example, in some embodiments, following superinfection of bacterial host cells with such a provided phagemid vector and a suitable helper phage (e.g., M13KO7, VCS-M13, or R408), intact phage particles are obtained, thereby allowing physical linkage of the encoded heterologous cDNA to its corresponding polypeptide displayed on phage surface (Lowman, Annu Rev Biophys Biomol Struct, 1997, Rodi and Makowski, Curr Opin Biotechnol, 1999).

In accordance with various embodiments, the cloning media may include, in addition to the regulatory sequences described above and nucleic acid sequences encoding the fusion protein as described herein, replication and control sequences derived from species compatible with the host cell that is used for expression, as well as selectable markers that confer a selectable phenotype on transformed or transfected cells. A large number of suitable cloning vectors are known in the art and are commercially available.

In some embodiments, the invention also provides methods for producing the fusion proteins of the invention, starting from a nucleic acid encoding the fusion protein or any subunits thereof, using genetic engineering techniques. In some embodiments, the proposed method can be performed in vivo, where the proposed fusion protein can, for example, be produced in a bacterial or eukaryotic host organism and then isolated from that host organism or culture thereof. It is also possible to produce the fusion protein according to the invention in vitro, for example using an in vitro translation system.

When producing a fusion protein in vivo, the nucleic acid encoding the fusion protein can be introduced into a suitable bacterial or eukaryotic host organism using recombinant DNA technology well known in the art. In some embodiments, a DNA molecule encoding a fusion protein as described herein (e.g., SEQ ID NOs: 138-144), and in particular a cloning vector containing the coding sequence of such a fusion protein, can be transformed into a host cell , capable of expressing the gene. The transformation can be carried out using standard methods. Thus, the invention also relates to host cells containing a nucleic acid molecule, as disclosed in

- 34 046634 standing document.

In some embodiments, the transformed host cells can be cultured under conditions suitable for expression of the nucleotide sequence encoding the fusion protein of the invention. In some embodiments, the host cells may be prokaryotic, such as Escherichia coli (E. coli) or Bacillus subtilis, or eukaryotic, such as Saccharomyces cerevisiae, Pichia pastoris, SF9 or High5 insect cells, immortalized mammalian cell lines (e.g., HeLa cells or CHO cells) or primary mammalian cell culture.

In some embodiments, where the lipocalin mutein of the invention, including that of a fusion protein disclosed herein, includes intramolecular disulfide bonds, it may be preferable to target the nascent protein to a cellular compartment having an oxidizing redox environment using an appropriate signal sequence. Such an oxidizing environment may be provided by the periplasm of Gram-negative bacteria such as E. coli, the extracellular environment of Gram-positive bacteria, or the lumen of the endoplasmic reticulum of eukaryotic cells, and typically promotes the formation of structural disulfide bonds.

In some embodiments, it is also possible to produce the fusion protein of the invention in the cytosol of a host cell, preferably E. coli. In this case, the proposed fusion protein can either be directly produced in a soluble and folded state, or isolated in the form of inclusion bodies followed by renaturation in vitro. Another option is the use of certain host strains that have an oxidizing intracellular environment, which may thus allow the formation of disulfide bonds in the cytosol (Venturi et al., J Mol Biol, 2002).

In some embodiments, the fusion protein of the invention described herein need not be created or produced, in whole or in part, using genetic engineering.

Rather, such a protein can also be produced by any of many traditional and well-known methods, such as simple organic synthesis strategies, solid phase synthesis methods, commercially available automated synthesizers, or by in vitro transcription and translation. For example, it is possible that promising lipocalin fusion proteins or muteins included in such fusion proteins will be identified by molecular modeling, synthesized in vitro, and screened for binding activity to the target(s) of interest. Methods for solid-phase and/or liquid-phase protein synthesis are well known in the art (see, for example, Bruckdorfer et al., Curr Pharm Biotechnol, 2004).

In some embodiments, the fusion protein of the invention can be produced by in vitro transcription/translation using conventional methods known to those skilled in the art.

In some additional embodiments, fusion proteins as described herein can also be produced by conventional recombinant methods, alone or in combination with conventional synthetic methods.

Moreover, in some embodiments, the fusion protein of the present invention can be produced by conjugating together individual subunits, such as immunoglobulins and muteins, of the fusion protein. Such conjugation can be achieved, for example, using all forms of covalent or non-covalent bonding using conventional methods.

One skilled in the art will know methods suitable for producing fusion proteins that are contemplated by the present invention, but whose protein or nucleic acid sequences are not explicitly disclosed herein. In general terms, such amino acid sequence modifications include, for example, site-directed mutagenesis of individual amino acid positions to facilitate subcloning of a protein gene or portions thereof by incorporating cleavage sites for individual restriction enzymes. Additionally, these mutations can be included to further improve the affinity of the fusion protein for its targets (eg, CD137 and PD-L1). In addition, mutations can be introduced to modulate one or more characteristics of the protein, for example, to improve folding stability, serum stability, protein stability or water solubility, or to reduce aggregation propensity, if desired.

V. Examples

Example 1: Expression and Analysis of Exemplary Fusion Proteins.

In this example, exemplary antibody-lipocalin mutein fusion proteins were produced by fusing a PD-L1-specific antibody having the heavy chain represented by SEQ ID NO: 86, or containing the heavy chain variable domain of SEQ ID NO: 77, or containing CDRs with the sequences GFSLSNYD (HCDR1, SEQ ID NO: 60), IWTGGAT (HCDR2, SEQ ID NO: 61), VRDSNYRYDEPFTY (HCDR3; SEQ ID NO: 62), and the light chains represented by SEQ ID NO: 87, or containing the heavy chain variable domain of SEQ ID NO: 82, or containing CDR with sequences

- 35 046634

QSIGTN (LCDR1, SEQ ID NO: 63), YAS (LCDR2), QQSNSWPYT (LCDR3; SEQ ID NO: 64), and CD137 specific lipocalin mutein with the sequence SEQ ID NO: 42 via a linker such as a non-structured linker (G ₄ S) ₃ with the sequence SEQ ID NO: 13, for simultaneous binding of PD-L1 and CD137. The various formats created are shown in FIG. 1. For example, such fusion proteins as SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NOs: 90 and 91 were prepared by fusing one or more lipocalin muteins of the sequence SEQ ID NO: 42 with one or more of the four ends of an antibody containing the heavy chain represented by the heavy chain represented by SEQ ID NO: 86, or containing a heavy chain variable domain with the sequence SEQ ID NO: 77, or containing a CDR with the sequences GFSLSNYD (HCDR1, SEQ ID NO: 60), IWTGGAT (HCDR2, SEQ ID NO: 61), VRDSNYRYDEPFTY (HCDR3; SEQ ID NO: 62), and light chains represented by SEQ ID NO: 87, or containing a heavy chain variable domain with the sequence SEQ ID NO: 82, or containing a CDR with the sequences QSIGTN (LCDR1, SEQ ID NO: 63), YAS (LCDR2), QQSNSWPYT (LCDR3; SEQ ID NO: 64). The resulting fusion proteins may be bivalent to CD137 (eg, as demonstrated in FIGS. 1A-1D), or tetravalent to CD137 (eg, as shown in FIGS. 1E-1H), or have even higher valency to CD137 (eg, as shown in Fig. 11).

The PD-L1-specific antibodies, as well as all antibody-lipocalin mutein fusion proteins described in this example, had an engineered IgG4 backbone that contained the S228P mutation to minimize the exchange of IgG4 halves in vitro and in vivo (Silva et al., J Biol Chem, 2015). Additional mutations in the IgG4 backbone may also exist in all antibodies and fusion proteins described herein, including any one or more of the F234A, L235A, M428L, N434S, M252Y, S254T, and T256E mutations.

The F234A and L235A mutations can be introduced to reduce ADCC and ADCP (Glaesner et al., Diabetes Metab Res Rev, 2010). The M428L and N434S mutations or the M252Y, S254T and T256E mutations can be introduced to increase serum half-life (Dall'Acqua et al., J Biol Chem, 2006, Zalevsky et al., Nat Biotechnol, 2010). All antibodies were expressed without the carboxy-terminal lysine to avoid heterogeneity.

In addition, monospecific lipocalin mutein and Fc fusions have been prepared by fusing one or more CD137-specific lipocalin muteins to SEQ ID NO: 42 via a linker, e.g., an unstructured (G ₄ S) ₃ linker to SEQ ID NO: 13, with The C-terminus of the Fc region of the antibody shown in SEQ ID NO: 30, as shown in FIG. 1J-1K. The resulting design is shown in SEQ ID NO: 88-89.

The present invention also includes asymmetric antibody-lipocalin mutein fusion formats, in which, for example, one light chain of an antibody may be fused to a lipocalin mutein and the other is not.

Fusion protein constructs were generated by gene synthesis and cloned into a mammalian expression vector. They were then transiently expressed in Expi293FTM cells (Life Technologies). The concentration of fusion proteins in the cell culture medium was measured by HPLC (Agilent Technologies) using a POROS® Protein A affinity column (Applied Biosystems). Summary data on titers of fusion proteins are given in Table. 1.

The fusion proteins were purified using protein A chromatography followed by size exclusion chromatography (SEC) in sodium phosphate buffer (PBS). After purification by SEC, fractions containing monomeric protein were pooled and reanalyzed by analytical SEC.

Table 1

Temporal expression titers

SEQ ID NO: Expression titer [m g/ml] SEQ ID NO: 90 and 87 0.05 SEQ ID NO: 86 and 91 0.07 SEQ ID NO: 92 and 87 0.06 SEQ ID NO: 86 and 93 0.09 SEQ ID NO: 94 and 87 0.06 SEQ ID NO: 90 and 91 0.10 SEQ ID NO: 88 0.24 SEQ ID NO: 86 and 87 0.07

Example 2: Expression of Fusion Proteins.

- 36 046634

The exemplary fusion protein constructs were generated by gene synthesis, including codon optimization, and cloned into a mammalian expression vector. They were then stably expressed in Chinese hamster ovary (CHO) cells. The concentration of fusion proteins in cell culture medium was measured using Octet (ForteBio, Pall Corp.) with Protein A sensors and quantified using a human IgG1 standard. Summary data on titers of fusion proteins are given in Table. 2. Data suggest that the geometry of fusion proteins may influence product yield and cell productivity.

table 2

Stable expression titers

SEQ ID NO: Expression titer [m g/ml] SEQ ID NO: 90 and 87 1.423 SEQ ID NO: 86 and 91 1,502 SEQ ID NO: 92 and 87 0.456 SEQ ID NO: 86 and 93 0.294 SEQ ID NO: 94 and 87 0.293 SEQ ID NO: 90 and 91 1,297 SEQ ID NO: 88 0.150 SEQ ID NO: 86 and 87 1.018

Example 3: Binding of fusion proteins to PD-L1 or CD137 determined by surface plasmon resonance (SPR).

The binding kinetics and affinity of exemplary fusion proteins to huPD-L1-His or huCD137His (human PD-L1 or human CD137 with a C-terminal polyhistidine tag, R&D Systems) were determined by surface plasmon resonance (SPR) using a Biacore 8K or Biacore T200 ( GE Healthcare).

Anti-human Fc IgG antibody (GE Healthcare) was immobilized onto the CM5 sensor chip using a standard amine reaction: the carboxyl groups on the chip were activated using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) . A solution of anti-human IgG Fc antibody (GE Healthcare) at a concentration of 25 μg/ml in 10 mM sodium acetate (pH 5.0) was then applied at a flow rate of 5 μl/min to an immobilization level of 6000-10000 resonance units (RU). Residual unreacted NHS esters were blocked by passing a 1 M ethanolamine solution through the surface. The reference channel was processed in a similar way. Then test fusion proteins (SEQ ID NO: 90 and 87, SEQ ID NO: 86 and 91, SEQ ID NO: 92 and 87, SEQ ID NO: 86 and 93, SEQ ID NO: 94 and 87, SEQ ID NO: 90 and 91, SEQ ID NO: 88 and SEQ ID NO: 89) at 0.25 μg/ml or 0.5 μg/ml in HBS-EP+ buffer was captured with anti-human Fc-IgG antibody on the chip surface for 180 sec at speed flow 10 µl/min. After each capture step, the needle was washed. Antibodies to PD-L1, including reference antibody (SEQ ID NOs: 26 and 27) and fusion protein antibody (SEQ ID NOs: 86 and 87), and reference antibody to CD137 (SEQ ID NOs: 28 and 29) also tested as controls.

To determine the affinity, dilutions of huPD-L1-His (10 nM, 5 nM and 2.5 nM) or huCD137-His (900 nM, 300 nM and 100 nM) were prepared in HBS-EP+ buffer and applied to the prepared chip surface. Binding assay carried out at a contact time of 180 sec, a dissociation time of 900 sec and a flow rate of 30 µl/min. All measurements were carried out at 25°C. Regeneration of the chip surface was achieved by injecting 3 M MgCl ₂ for 120 sec. Three start-up cycles were carried out before measuring the protein. for preparation purposes, data were evaluated using Biacore T200 evaluation software (v2.0) or Biacore 8K evaluation software (V1.1.1) to fit the raw data using a double comparison with control and a 1:1 binding model.

Summary values determined for k _on , k _off and the resulting equilibrium dissociation constant (KD) for exemplary fusion proteins are given in Table. 3. All bispecific fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NO: 90 and 91) bind PD-L1 as well as CD137 with affinities in the subnanomolar or low nanomolar range. Monospecific CD137-specific lipocalin-Fc mutein fusions (SEQ ID NO: 88 and SEQ ID NO: 89) bind only CD137 with affinity in the low nanomolar range.

- 37 046634

Table 3

Kinetic constants and affinities of fusion proteins determined by SPR

SEQ ID NO: huPD-L1 huCD137 kon [Μ ¹ x s ¹ ] _kOff [s-1] K _D [nM] kon [Μ ¹ x s ¹ ] _kOff [s ¹ ] K _D [nM] 90 and 87 1.30E+06 7.69E-04 0.592 3.93E+04 2.55E-04 6,484 86 and 91 1.32E+06 7.48E-04 0.568 3.42E+04 1.75E-04 5.106 92 and 87 7.87E+05 8.03E-04 1.021 3.48E+04 2.08E-04 5,966 86 and 93 5.46E+05 7.54E-04 1.381 3.68E+04 1,ZOE-04 3,548 94 and 87 1.51E+06 9.16E-04 0.608 4.07E+04 2.24E-04 5,504 90 and 91 1.60E+06 8.61 E-04 0.537 3.91 E+04 2.24E-04 5,726 88 nd nd nd 3.94E+04 1.63E-04 4.155 86 and 87 1.09E+06 6.49E-04 0.593 nd nd nd 89 nd nd nd 3.66E+04 1.32 E-04 3,595 26 and 27 6.38E+05 2.01 E-04 0.314 nd nd nd 28 and 29 nd nd nd 4.72E+05 2.94E-03 6,237

Example 4 Binding of fusion proteins to PD-L1 or CD137 in an enzyme-linked immunosorbent assay (ELISA).

An enzyme-linked immunosorbent assay (ELISA) was used to determine the binding strength of exemplary fusion proteins to human PD-L1 and cynomolgus PD-L1.

Recombinant huPD-L1-His or cyPD-L1-His (human or cynomolgus PD-L1 with C-terminal polyhistidine tag, R&D Systems or Sino Biologics) at a concentration of 1 μg/ml in PBS was applied overnight to microtiter plates at 4°C . After washing with PBS-0.05%T (PBS supplemented with 0.05% (v/v) Tween 20), the plates were blocked with 2% BSA (w/v) in PBS-0.1%T (PBS with adding 0.1% (v/v) Tween 20) for 1 hour at room temperature. After washing five times with 100 μl PBS-0.05%T, the exemplary fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NO: 94 and 87, SEQ ID NO: 90 and 91), CD137-specific lipocalin mutein Fc fusions (SEQ ID NO: 88 and SEQ ID NO: 89) and anti-PD-L1 antibodies (SEQ ID NO: 26 and 27 , SEQ ID NOs: 86 and 87) at various concentrations were added to the wells and incubated for 1 hour at room temperature, followed by another washing step. Bound molecules of interest were detected by incubation with a 1:5000 dilution of anti-human IgG Fc-HRP (Jackson Laboratory) in PBS-0.1%T-2%BSA. After an additional washing step, the fluorogenic substrate HRP (QuantaBlu, Thermo) was added to each well and the fluorescence intensity was determined using a fluorescence microplate reader.

The same ELISA scheme was also used to determine the binding strength of the fusion proteins to CD137, where huCD137-His (human CD137 with a C-terminal polyhistidine tag, R&D Systems) or cyCD137-Fc (cynomolgus CD137 fused from Fc to C) was reciprocally coated on a microtiter plate. -end). The test agents were titrated in a similar manner and bound agents were detected using an anti-NGAL-HRP antibody.

The results of illustrative experiments are presented in FIG. 2A-2D along with fitting curves obtained from a 1:1 sigmoidal binding curve fit where EC ₅₀ and maximum signal were free parameters and the slope was set to one. The obtained EC ₅₀ values are presented in table. 4.

Observed EC ₅₀ values relative to two human targets of the proposed fusion proteins (SEQ ID NO: 90 and 87, SEQ ID NO: 86 and 91, SEQ ID NO: 92 and 87, SEQ ID NO: 86 and 93, SEQ ID NO: 94 and 87, SEQ ID NOs: 90 and 91, SEQ ID NO: 88, and SEQ ID NO: 89) were very similar or comparable to the anti-PD-L1 antibodies tested (reference anti-PD-L1 antibody with the sequences SEQ ID NO: 26 and 27 and an anti-PD-L1 antibody with the sequence SEQ ID NO: 86 and 87 included in the fusion proteins) and/or CD137-specific lipocalin mutein included in the fusion proteins (SEQ ID NO: 42).

All fusion proteins tested exhibit cross-reactivity with cynomolgus PD-L1, with EC ₅₀ values comparable to the reference anti-PD-L1 antibody (SEQ ID NOs: 26 and 27) or the anti-PD-L1 antibody included in the fusion proteins (SEQ ID NO: 86 and 87). Only fusion proteins that are tetravalent to CD137 (SEQ ID NOs: 94 and 87, SEQ ID NOs: 90 and 91, and SEQ ID NO: 88) show cross-reactivity with cynomolgus CD137 at a level comparable to human CD137, i.e. e. associated cynomolgus CD137 with EC ₅₀ values in the same range as

- 38 046634 corresponding EC50 for human CD137.

Table 4

ELISA data for PD-L1 or CD137 binding

SEQ ID NO: EC ₅ o [nM] Binding to huPD-L1 EC ₅₀ [nM] Binding to cyPD-L1 EC ₅₀ [nM] chuCD137 binding EC ₅₀ [nM] Binding to cyCD137 90 and 87 0.15 0.13 0.28 5.9 86 and 91 0.23 0.18 0.57 13 92 and 87 0.16 0.15 0.41 9.8 86 and 93 0.19 0.17 0.35 8.1 94 and 87 0.18 0.15 0.16 0.19 90 and 91 0.20 0.17 0.14 0.21 88 nd nd 0.15 0.12 86 and 87 0.12 0.1 nd nd 89 nd nd 0.29 0.82 26 and 27 0.09 0.09 nd nd 42 nd nd 0.27 nd

Example 5: Simultaneous binding of fusion proteins to PD-L1 and CD137 in ELISA.

To demonstrate simultaneous binding of exemplary fusion proteins to PD-L1 and CD137, a dual-binding ELISA format was used.

Recombinant huPD-L1-His (R&D Systems) in PBS (1 μg/ml) was applied to microtiter plates overnight at 4°C. The plates were washed five times after each incubation step with 100 μl PBS0.05%T. The plates were blocked with 2% BSA (w/v) in PBS-0.1%T for 1 h at room temperature and then washed again. Various concentrations of test fusion proteins were added to the wells and incubated for 1 h at room temperature, followed by a washing step. Then, biotinylated huCD137-His (huCD137-His-Bio, Sino Biological) was added at a constant concentration of 1 μg/ml in PBS-0.1%T-2%BSA for 1 hour. After washing, a diluted 1:5000 ExtrAvidin- HRP (Sigma-Aldrich) in PBS-0.1 ^o /oT-2% BSA and incubated for 1 h. After an additional washing step, the fluorogenic HRP substrate (QuantaBlu, Thermo) was added to each well and the fluorescence intensity was determined using a fluorescence reader devices for microplates.

Double binding of exemplary fusion proteins was also tested using a reverse design where recombinant 1 μg/ml huCD137-His (R&D Systems) was coated on microtiter plates and bound fusion proteins were detected by adding biotinylated huPDL1-His (R&D Systems).

Dual Fusion Binding Data (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NO: 90 and 91) are shown in Fig. 3A and 3B along with fitting curves obtained from a 1:1 sigmoidal binding curve fit where the EC50 value and maximum signal were free parameters and the slope was set to unity. Summary EC50 values are given in table. 5. All bispecific fusion proteins exhibit clear binding signals, indicating that the fusion proteins are capable of simultaneously binding PD-L1 and CD137. The data also suggest that fusion of CD137-specific lipocalin muteins to the C termini of PD-L1-specific antibodies may be favored over the N termini.

Table 5

ELISA Data for Simultaneous Targeting of PD-L1 and CD37

SEQ ID NO: EC ₅ o [nM] Capture PO-1_1_detection CD137 EC ₅ o [nM] CO137 capture_PD-L1 detection 90 and 87 0.58 2.9 86 and 91 0.59 3.3 92 and 87 1.2 7.2 86 and 93 0.74 6.7 94 and 87 0.48 2.6 90 and 91 0.49 2.3

- 39 046634

Example 6 Flow cytometry analysis of fusion protein binding to cells expressing human and cynomolgus CD137 and PD-L1.

Target specific binding of the fusion proteins to cells expressing human and cynomolgus PD-L1 and cells expressing human and cynomolgus CD137 was assessed by flow cytometry.

CHO cells were stably transfected with human PD-L1, cynomolgus PD-L1, human CD137, cynomolgus CD137, or mock control using the Flp-In system (Life technologies) according to the manufacturer's instructions.

Transfected CHO cells were maintained in Ham's F12 medium (Life technologies) supplemented with 10% fetal calf serum (Biochrom) and 500 μg/ml hygromycin B (Roth). Cells were cultured in cell culture flasks according to the manufacturer's instructions (37°C, 5% CO2 atmosphere).

For flow cytometry analysis, appropriate cell lines were incubated with fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NO: 94 and 87, SEQ ID NOs: 90 and 91, SEQ ID NO: 88, and SEQ ID NO: 89) and detected using fluorescently labeled anti-human IgG antibody in a FACS assay as described below: 5x104 cells per well were incubated for 1 h in ice-cold PBS containing 5% fetal calf serum (PBS-FCS). Serial dilutions of the fusion proteins and control antibodies were added to the cells and incubated for 1 h on ice. Cells were washed twice with PBS and then incubated with Alexa647-labeled goat anti-hIgG antibody for 30 min on ice. Cells were then washed and analyzed using an iQue flow cytometer (Intellicyte Screener). Geometric means of fluorescence signals were plotted and fitted using Graphpad software using nonlinear regression (overall lower asymptote, slope = 1).

The ability of the fusion proteins to bind human and cynomolgus PD-L1 and CD137 is shown in FIG. 4. Binding affinities (EC50) of bispecific fusion proteins (SEQ ID NO: 90 and 87, SEQ ID NO: 86 and 91, SEQ ID NO: 92 and 87, SEQ ID NO: 86 and 93, SEQ ID NO: 94 and 87 , and SEQ ID NOs: 90 and 91) to cells expressing human and cynomolgus PD-L1 are in the single-digit nanomolar range, demonstrating complete cross-reactivity with cynomolgus (summarized in Table 6). The binding affinities of the fusion proteins to cells expressing human CD137 are in the low nanomolar range. The fusion proteins tested fully cross-reacted with cynomolgus CD137 (SEQ ID NOs: 94 and 87, SEQ ID NOs: 90 and 91, and SEQ ID NO: 88), binding cynomolgus CD137 with a 6- to 13-fold reduced affinity compared to the corresponding binding affinities to human CD137 (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91 and SEQ ID NO: 89), or do not bind cynomolgus CD137 (SEQ ID NOs: 92 and 87 and 86 and 93). Neither fusion protein binds to mock-transfected cells.

Table 6

Binding affinities of fusion proteins to cells expressing human and cynomolgus PD-L1 or CD137

SEQ ID NO: EC ₅₀ [nM] Flp-lnCHO::huCD137 EC ₅ o [nM] Flp-lnCHO::cynoCD137 EU ₅ o [nM] Flp-lnCHO::huPDL-1 EU ₅ o [nM] Flp-lnCHO::cynoPDL-1 90 and 87 3.74 51.35 3.64 4.48 86 and 91 6.48 39.15 1.43 1.41 92 and 87 11.91 - 2.95 2.03 86 and 93 12.73 - 4.63 2.03 94 and 87 4.15 6.92 6.99 5.89 90 and 91 4.69 3.66 4.09 3.37 86 and 87 - - 2.96 3.55 89 5.74 26.70 - - 88 4.33 3.81 - - 28 and 29 1.31 - - -

Example 7 Binding affinities of fusion proteins to PD-L1 positive tumor cells.

The binding of fusion proteins to tumor cells expressing PD-L1 was assessed by flow cytometry.

The PD-L1-expressing colorectal cancer cell line RKO was maintained in RPMI1640 (Life technologies) supplemented with 10% FCS at 37°C in a humidified atmosphere with 5% CO2.

- 40 046634

For flow cytometry analysis, RKO cells were incubated with the fusion proteins and detected using a fluorescently labeled anti-human IgG antibody as described in Example 6.

The ability of fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NO: 90 and 91) bind PD-L1 positive tumor cells is shown in FIG. 5, and a summary of the corresponding binding affinities (EC ₅₀ ) is presented in table. 7. The binding affinities of the fusion proteins to PD-L1-expressing RKO cells were in the low nanomolar or sub-nanomolar range and were comparable to the anti-PD-L1 antibody included in the fusion proteins (SEQ ID NOs: 86 and 87).

Table 7

Binding affinities of fusion proteins to PD-L1-positive tumor cells

Example 8 Determination of competition between CD137L and fusion proteins for binding to CD137 by SPR method.

An SPR assay was used to examine the competition between human CD137L (huCD137L-His, R&D Systems) and exemplary fusion proteins for binding to human CD137. Competition assays were performed at 25°C on a Biacore T200 instrument (GE Healthcare).

BiotinCAPture reagent (GE Healthcare) was immobilized on the CAP sensor chip at a concentration of 50 μg/ml and a flow rate of 2 μl/min for 300 s. The reference channel was processed in a similar way. Biotinylated huCD137-Fc (R&D systems) was captured on the chip surface for 300 s at a concentration of 1 μg/mL and at a flow rate of 5 μL/min through the other channel.

To analyze whether the fusion proteins tested compete (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 90 and 91, SEQ ID NO: 88, and SEQ ID NO: 89) with CD137L for CD137 binding, running buffer (HBS-EP+ buffer) or 500 nM huCD137L-His was applied to the chip surface for 180 s at a flow rate of 30 μl/min. The test fusion proteins were then applied to the prepared chip surface in HBS-EP+ buffer at a fixed concentration of 1 μM. The binding assay was performed at a contact time of 180 s, a dissociation time of 15 s, and a flow rate of 30 μL/min. As a control, a buffer was introduced with the same parameters. Chip surface regeneration was achieved by injecting 6M guanidine-HCl, 0.25M NaOH for 120 sec at a flow rate of 10 µL/min, followed by an additional _H2O washing step (120 sec, 10 µL/min).

Illustrative examples for the relevant segment of the obtained sensorgrams are presented in Fig. 6 giving fusion proteins SEQ ID NO: 90 and 87, SEQ ID NO: 86 and 91, SEQ ID NO: 92 and 87, SEQ ID NO: 86 and 93, SEQ ID NO: 90 and 91, SEQ ID NO: 88 and SEQ ID NO: 89. The SPR curve for binding of the corresponding fusion protein to huCD137-Fc alone is indicated by a solid arrow. The SPR curve for binding of the huCD137-Fc fusion protein saturated with huCD137L-His is indicated by a dotted arrow. The data show that all fusion proteins bind to huCD137 in the presence of huCD137L, but with slightly reduced signals compared to their binding to CD137 in the absence of CD137L. This suggests that the fusion proteins tested may be somewhat sterically hindered by the binding of CD137L to CD137. This binding behavior of fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 90 and 91, SEQ ID NO: 88 and SEQ ID NO: 89) behaves similarly to the anti-CD137 antibody with SEQ ID NOs: 28 and 29.

Example 9. Competition of PD-L1 fusion proteins for binding to PD-1 determined by ELISA.

To demonstrate the ability of the fusion proteins to inhibit the interaction between PD-1 and PD-L1, a competitive ELISA format was used.

Recombinant huPD-1-His (Acrobiosystems) in PBS (1 μg/ml) was applied to microtiter plates overnight at 4°C. The plates were washed five times after each incubation step with 100 μl PBS0.05%T (PBS supplemented with 0.05% (v/v) Tween 20). Plates were blocked with 2% BSA (w/v) in PBS0.1%T (PBS supplemented with 0.1% (v/v) Tween 20) for 1 h at room temperature and then again

- 41 046634 washed. Various concentrations of fusion proteins were mixed with 15 nM recombinant huPD-L1-Fc (R&D systems) as an indicator and incubated for 1 h at room temperature. Fusion protein and indicator mixtures were added to the plates and incubated for 20 min at room temperature, followed by five washing steps with 100 μl PBS-0.05%T. A 1:5000 dilution of goat anti-human Fc-IgG HRP antibody (Jackson) was then added to the wells and incubated for 1 h. After an additional washing step, the fluorogenic HRP substrate (QuantaBlu, Thermo) was added to each well and the fluorescence intensity was determined using a fluorescence reader. devices for microplates.

Competition data for exemplary fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, and SEQ ID NOs: 90 and 91) are shown in FIG. . 7 along with fitting curves obtained from a 1:1 sigmoidal binding curve fit where the IC ₅₀ value and maximum signal were free parameters and the slope was set to unity. Summary IC ₅₀ values are given in table. 9. All bispecific fusion proteins demonstrated clear inhibition of PD-1/PD-L1 interaction with IC ₅₀ values comparable to the building block antibody (SEQ ID NOs: 86 and 87) and the reference PD-L1 antibody (SEQ ID NO: 26 and 27).

Table 8

Competition of PD-L1 fusion proteins for binding to PD-1

Example 10: PD-L1-Dependent T Cell Costimulation Using a CD137 Bioassay.

The potential of the selected fusion proteins to induce activation of the CD137 signaling pathway in the presence of PD-L1 was assessed using a commercially available doubly stably transfected Jurkat cell line expressing CD137 and the luc2 gene (a humanized version of firefly luciferase), where luc2 expression is regulated by the NFKB response element. In this bioassay, CD137 binding results in intracellular CD137 signaling, resulting in NFκB-mediated luminescence.

The PD-L1-expressing colorectal cancer cell line RKO was cultured as described in Example 7. The day before analysis, RKO cells were seeded at 1.25 x 10 ⁴ cells per well and allowed to attach overnight at 37°C in a humidified 5% atmosphere. _CO2 .

The next day, 3.75 x 10 ⁴ NF-kB-Luc2/CD137 Jurkat cells were added to each well, followed by the addition of various concentrations, typically ranging from 0.001 to 5 nM, of fusion proteins or CD137 reference antibody (SEQ ID NOs: 28 and 29 ). The plates were covered with gas-permeable film and incubated at 37 °C in a humidified atmosphere with 5% CO ₂ . After 4 hours, 30 μl of Bio-Glo™ reagent was added to each well and the bioluminescent signal was quantified using a luminometer (PHERAstar). Four-parameter logistic curve analysis was performed using GraphPad Prism® to calculate EC ₅₀ (common lower asymptote, fixed slope) values, which are summarized in Table. 9. To demonstrate the PD-L1 dependence of CD137 fusion protein binding, the same experiment was performed in parallel in the absence of RKO cells. The analysis was carried out in triplicate.

The results of an illustrative experiment are shown in FIG. 8A-8D.

Data show that all fusion proteins tested (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, and SEQ ID NOs: 86 and 93) induced strong CD137-mediated T cell co-stimulation . In fig. 8B and 8D show that activation of CD137 by fusion proteins is dependent on PDL1, since activation of NF-κB-Luc2/CD137 Jurkat cells was not detected in the absence of tumor cells expressing PD-L1. In contrast, the reference mAb to CD137 (SEQ ID NOs: 28 and 29) demonstrated CD137-mediated costimulation of T cells regardless of the presence or absence of target cells.

- 42 046634

Table 9

Assessing T cell activation using CD137 bioassay

Example 11: Assessment of T cell activation using human peripheral blood mononuclear cells (PBMCs).

The T cell assay was used to evaluate the ability of selected fusion proteins to co-stimulate T cell responses as well as prevent co-inhibition mediated by PDL1 binding to PD-1. For this purpose, fusion proteins at various concentrations were added to staphylococcal enterotoxin B (SEB)-stimulated human peripheral blood mononuclear cells (PBMCs) and incubated for 4 days at 37°C. IL-2 secretion levels were measured in the supernatants.

PBMCs from healthy volunteer donors were isolated from the buffy platelet layer by polysucrose density gradient centrifugation (Biocoll, 1.077 g/ml, Biochrom) following Biochrom protocols. Purified PBMCs were resuspended in a buffer consisting of 90% FCS and 10% DMSO, immediately frozen, and stored in liquid nitrogen until further use. For analysis, PBMCs were thawed and placed in culture medium (RPM11640, Life Technologies) supplemented with 10% FCS and 1% penicillin-streptomycin (Life Technologies) for 16 h at 37°C in a humidified atmosphere with 5% CO2.

The following procedure was performed in triplicate for each experimental condition: 2.5 x 104 PBMCs were incubated in each well of 384-well flat-bottom tissue culture plates in culture medium. Serial dilutions of fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NO: 90 and 91), anti-PD-L1 building block antibody (SEQ ID NO: 86 and 87), reference anti-PD-L1 antibody (SEQ ID NO: 26 and 27), reference anti-CD137 antibody (SEQ ID NO: 28 and 29 ), a cocktail of reference anti-CD137 antibody (SEQ ID NO: 28 and 29) and anti-PD-L1 antibody (SEQ ID NO: 86 and 87 or SEQ ID NO: 26 and 27) or isotype control (SEQ ID NO: 24 and 25), typically ranging from 10 to 0.002 nM, and SEB at a concentration of 0.1 ng/ml was added to the appropriate wells. The plates were covered with gas-permeable film (4titude) and incubated at 37°C in a humidified atmosphere with 5% CO ₂ for four days. IL-2 levels in the supernatant were then assessed using the Human IL-2 DuoSet kit (R&D Systems) as described in the following procedures.

384-well plates were coated for 2 hours at room temperature with 1 μg/ml anti-human IL-2 capture antibody in PBS. The wells were then washed 5 times with 80 μl PBS supplemented with 0.05% Tween (PBS-T). After 1 h of blocking in PBS-0.05%T containing 1% casein (w/w), assay supernatants and a series of concentrations of IL-2 standard diluted in culture medium were transferred to appropriate wells and incubated overnight at 4°C. The next day, a mixture of 100 ng/ml goat anti-hIL-2-Bio detection antibody (R&D Systems) and 1 μg/ml Sulfotag-labeled streptavidin (Mesoscale Discovery) was added in PBS-T containing 0.5% casein and incubated at room temperature for 1 hour. After washing, 25 μl of reading buffer (Mesoscale Discovery) was added to each well and the resulting electrochemiluminescence (ECL) signal was detected by a Mesoscale Discovery reader. Analysis and quantification were performed using Mesoscale Discovery software.

The result of an illustrative experiment is shown in FIG. 9.

Bispecific Fusion Proteins SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NOs: 90 and 91 capable of inducing T cell activation, as evidenced by increased levels of IL-2 secretion compared to isotype control (hlgG4, Sigma). The largest increase in IL-2 secretion is observed for CD137 tetravalent fusion proteins (SEQ ID NOs: 94 and 87 and SEQ ID NOs: 90 and 91), followed by CD137 bivalent fusion proteins where the lipocalin mutein is fused to C- end of a PD-L1-specific antibody (SEQ ID NOs: 90 and 87 and SEQ ID NOs: 86 and 91). The smallest increase is observed for CD137 bivalent fusion proteins, where a lipocalin mutein is fused to the N-terminus of a PD-L1-specific antibody (SEQ ID NOs: 92 and 87 and SEQ ID NOs: 86 and 93), but is still comparable to a cocktail of a reference antibody to CD137 (SEQ ID NOs: 28 and 29) and a reference antibody to PD-L1 (SEQ ID NOs: 26 and 27). All fusion proteins exhibit higher levels of IL-2 secretion than the individual building blocks, i.e. CD137 specific lipocalin-Fc mutein (SEQ ID NO: 88 or SEQ ID NO: 89) or anti-PD-L1 mAb

- 43 046634 building block (SEQ ID NO: 86 and 87).

Example 12: Evaluation of T cell activation in the presence of tumor cells expressing varying levels of PD-L1.

An additional T cell assay was used to evaluate the ability of the fusion proteins to co-stimulate T cell activation in a PD-L1 target-dependent manner. The fusion proteins were applied at varying concentrations to anti-CD3-stimulated T cells in the presence of tumor cell lines with varying levels of PD-L1 expression. Tumor cell lines tested include RKO (high PD-L1), HCC827 (moderate PD-L1), and Hep-G2 (PD-L1 negative). IL-2 secretion levels were measured in the supernatants.

PBMCs from healthy volunteer donors were isolated from the buffy platelet layer as described in Example 11. T lymphocytes were further isolated from PBMCs by magnetically activated cell sorting using a pan-T cell purification kit (Miltenyi Biotec GmbH) according to the instructions manufacturer. Purified pan T cells were resuspended in a buffer consisting of 90% FCS and 10% DMSO, immediately frozen, and stored in liquid nitrogen until further use.

For analysis, T cells were thawed and placed in culture medium (RPMI 1640, Life Technologies) supplemented with 10% FCS and 1% penicillin-streptomycin (Life Technologies) for 16 h at 37°C in a humidified atmosphere with 5% CO ₂ .

The following procedure was performed in triplicate for each experimental condition: Flat-bottomed tissue culture plates were precoated with 0.25 μg/ml anti-CD3 antibody for 1 h at 37°C and then washed twice with PBS. RKO, HCC827, or Hep-G2 tumor cell lines were treated for 30 min with 30 μg/ml mitomycin C (Sigma Aldrich) to block proliferation. Mitomycin-treated tumor cells were then washed twice with PBS and seeded at 2.5 x 104 cells per well in culture medium to allow attachment overnight at 37°C in a humidified atmosphere of 5% CO2. Target cells previously grown under standard conditions were separated using Accutase (PAA Laboratories) and resuspended in culture medium.

On the following days, after washing the plates twice with PBS, 1.25 x 104 T cells per well were added to the tumor cells. Serial dilutions of fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, and SEQ ID NOs: 86 and 93), reference anti-PD-L1 antibody (SEQ ID NO: 26 and 27) and reference anti-CD137 antibody (SEQ ID NOs: 28 and 29), used alone or in combination, or isotype control (SEQ ID NOs: 24 and 25), typically ranging from 0.005 nM to 10 nM, was added to corresponding holes. The plates were covered with gas-permeable film and incubated at 37°C in a humidified atmosphere with 5% CO ₂ for 3 days.

After 3 days of co-culture, the level of IL-2 in the supernatant was assessed as described in Example 11.

Exemplary data is shown in FIG. 10. Coculture of pan-T cells with RKO (high PD-L1) or HCC827 (moderate PD-L1) cells in the presence of fusion proteins where the lipocalin mutein is fused to the C terminus of a PD-L1-specific antibody (SEQ ID NOs: 90 and 87 and SEQ ID NOs: 86 and 91) resulted in a clear increase in IL-2 secretion compared to the hIgG4 isotype control. The increase in IL-2 secretion caused by fusion proteins where a lipocalin mutein is fused to the N-terminus of a PD-L1-specific antibody (SEQ ID NOs: 92 and 87 and SEQ ID NOs: 86 and 93) was weaker, but still higher than with a cocktail of the reference antibody to CD137 (SEQ ID NO: 28 and 29) and the reference antibody to PD-L1 (SEQ ID NO: 26 and 27). No increase in IL-2 secretion was observed with the building block cocktail of CD137-specific lipocalin mutein (Fc fusion, SEQ ID NO: 89) and anti-PD-L1 antibody (SEQ ID NO: 86 and 87). In addition, co-culture with Hep-G2 (PD-L1 negative) did not increase levels of IL-2 secretion with any fusion proteins, unlike a cocktail of reference anti-CD137 antibody (SEQ ID NOs: 28 and 29) and reference antibodies to PD-L1 (SEQ ID NO: 26 and 27).

Data show that the functional activity of the fusion proteins, as measured by their ability to activate T cells or increase IL-2 secretion, is dependent on PD-L1. In contrast, T cell activation or IL-2 secretion induced by the reference antibody to CD137 (SEQ ID NOs: 28 and 29) when used in combination with the reference antibody to PD-L1 (SEQ ID NOs: 26 and 27) is not necessarily dependent on PD -L1 and is difficult to predict. Additionally, the data show that the bispecific format of targeting PD-L1 and CD137 is superior to a cocktail of two separate molecules targeting CD137 and PD-L1 in the presence of PD-L1-expressing target cells.

Example 13. Evaluation of storage stability of fusion proteins.

For evaluation of storage stability, exemplary fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, SEQ ID NO: 88 and SEQ ID NO: 89) were incubated for 1 week at 37°C at a concentration of 1 mg/ml in PBS. Monomeric fusion proteins were then determined using analytical size exclusion chromatography by applying 20 μg of sample to a Superdex 200, 3.2/300 Increase column (GE Healthcare) at flow rate

- 44 046634

0.15 ml/min and PBS as running buffer. All fusion peptides tested were stable after incubation for 1 week in PBS at 37°C. Exemplary results are shown in FIG. 11A.

Further storage stability evaluations were performed on the selected fusion protein with SEQ ID NOs: 90 and 87. The fusion protein at a concentration of 20 mg/ml was incubated for four weeks at 40°C in 25 mM histidine, 60 mM NaCl, 200 mM arginine pH 6. Functional fusion protein content was measured by quantitative ELISA using a co-binding assay as described in Example 5. Exemplary results are shown in FIG. 11B.

Example 14. Evaluation of mixed lymphocyte culture reaction (MLR) with CD4+ T cells.

A mixed lymphocyte response (MLR) assay was used to evaluate the ability of an exemplary fusion protein to induce CD4+ T cell activation in the presence of antigen presenting cells. The fusion protein (SEQ ID NOs: 90 and 87) at various concentrations was tested in the presence of monocytic dendritic cells (moDC) and CD4+ T cells from matched healthy donors. After 6 days of culture in the presence of the test molecules, the secretion of IL-2 and IFN-gamma in the supernatants was quantified.

PBMCs were purified from the apheresis platelet concentrate packet using Lymphoprep solution following the manufacturer's instructions (StemCell). Total CD4+ T cells were purified from PBMCs using a kit from Miltenyi and frozen in 90% FBS 10% DMSO. CD14' monocytes were purified using a CD14+ microsphere kit (Miltenyi) and used fresh.

MoDC were generated by culturing CD14+ monocytes in RPMI1640 with 10% FBS and penicillin/streptomycin (LifeTech) in the presence of 50 ng/ml IL-4 and 100 ng/ml GM-CSF (Miltenyi) for 6 days at 2x10 ⁶ cells/ml . On day 3, 10 ml of fresh medium containing cytokines was added. Phenotype (CD14, CD1a, HLADR, PD-L1) was assessed at day 7 of differentiation using FACS.

10,000 moDCs were cultured in the presence of 50,000 CD4+ T cells in 96-well U-bottom plates in complete RPMI medium, in the presence of test molecules, for 6 days in RPMI in three parallel wells. At the end of culture, supernatants were immediately frozen and stored for cytokine quantification.

IL-2 levels were measured in the supernatants using Luminex technology and exemplary data are shown in FIG. 12. In FIG. 12A shows that the fusion protein (SEQ ID NO: 90 and 87) demonstrated significantly better IL-2 induction at 10 and 0.1 μg/ml compared to the corresponding building blocks (SEQ ID NO: 89 and SEQ ID NO: 86 and 87) alone or with the reference anti-CD137 or anti-PD-L1 antibody (SEQ ID NOs: 28 and 29 or SEQ ID NOs: 26 and 27) in several series of MLR experiments (N=8). In fig. 12B shows that the fusion protein (SEQ ID NOs: 90 and 87) induced dose-dependent secretion of IL-2 compared to the isotype control antibody. IL-2 levels induced by the fusion protein SEQ ID NOs: 90 and 87 were higher compared to equimolar concentrations of a cocktail of the reference anti-PD-L1 antibody (SEQ ID NOs: 26 and 27) and the reference anti-CD137 antibody (SEQ ID NO: 28 and 29) in the concentration range from 0.001 to 20 µg/ml.

Example 15. Evaluation of the reaction of a mixed culture of lymphocytes with CD8+ T cells.

Given the literature on CD137 expression and activity in human CD8+ T cells, the ability of an exemplary fusion protein to induce CD8+ T cell activation in the presence of antigen presenting cells was assessed in an MLR assay. The fusion protein (SEQ ID NOs: 90 and 87) was tested in the presence of moDC and total CD8+ T cells from matched healthy donors. After 6 days of culture in the presence of the test molecules, the secretion of IL-2 and CD8 effector molecules (perforin, granzyme B and granzyme A) in the supernatants was quantified.

PBMCs were purified from the apheresis platelet concentrate packet using Lymphoprep solution following the manufacturer's instructions (StemCell). Total CD8+ T cells were purified from PBMCs using a kit from Miltenyi and used fresh. CD14-positive monocytes were purified using a CD14+ microsphere kit (Miltenyi) and used fresh.

MoDC were generated by culturing CD14+ monocytes in RPMI1640 with 10% FBS and penicillin/streptomycin (LifeTech) in the presence of 50 ng/ml IL-4 and 100 ng/ml GM-CSF (Miltenyi) for 6 days at 2x10 ⁶ cells/ml . On day 3, 10 ml of fresh medium containing cytokines was added. Phenotype (CD14, CD1a, HLADR, PD-L1) was assessed at day 7 of differentiation using FACS.

10,000 moDC were cultured with 50,000 CD8+ T cells in the presence of test molecules in 96-well U-bottom plates (in three parallel wells) in complete RPMI medium for 6 days. At the end of culture, supernatants were immediately frozen and stored for quantification of secreted factors.

IL-2, perforin, granzyme A, and granzyme B were quantified in the supernatants using Luminex technology. Exemplary data is shown in FIG. 13.

In fig. 13 shows that the fusion protein (SEQ ID NOs: 90 and 87) demonstrated increased secretion of IL-2, perforin, granzyme B and granzyme A compared to an equimolar concentration of the reference anti-PD-L1 antibody (SEQ ID NOs: 26 and 27), reference antibody to CD137 (SEQ ID NO: 28 and 29) or

- 45 046634 with a cocktail of these two antibodies at a concentration of 10 μg/ml (N=4). The data also indicate that the fusion protein (SEQ ID NOs: 90 and 87) has activity against cytotoxic CD8 T cells.

Example 16. Evaluation of functional activity in vivo on a mouse xenograft model with grafted human PBMCs.

To examine the in vivo activity of the proposed fusion proteins, a cell line-derived xenograft mouse model will be used. Accordingly, a human cancer cell line will be subcutaneously implanted into immunodeficient female NOG mice delivered at 4-6 weeks of age with at least 1 week of quarantine. Once the tumors have reached a volume of approximately 80-100 mm ³ , human PBMCs will be transferred to the mice. Test compounds will be injected at least three times, and tumor growth and activity will be continuously measured. Upon reaching the end of the study, mice will be sacrificed. Intratumoral infiltration of CD3-, CD4-, and CD8-positive cells will be assessed using immunohistochemistry. As an additional indicator, IFN-gamma analysis will be carried out using the RNAscope method.

Example 17 Epitope Analysis of Fusion Proteins.

To evaluate the epitopes that the fusion proteins recognize and whether they are clinically significant, a competitive ELISA format was used to determine the competition between the fusion proteins and the reference CD137 antibody.

Microtiter plates were coated with the reference anti-CD137 antibody of SEQ ID NOs: 28 and 29 in PBS (4 μg/ml) at 4°C overnight. The plates were washed five times after each incubation step with 100 μl PBS-0.05%T (PBS supplemented with 0.05% (v/v) Tween 20). Plates were blocked with 2% BSA (w/v) in PBS-0.1%T (PBS supplemented with 0.1% (v/v) Tween 20) for 1 h at room temperature and then washed again . Fusion proteins (SEQ ID NOs: 90 and 87 and SEQ ID NOs: 86 and 91), CD137-specific lipocalin mutein (SEQ ID NO: 42), CD137 reference antibody (SEQ ID NOs: 28 and 29) and control antibody (SEQ ID NO: 86 and 87) at various concentrations were mixed with 1 nM biotinylated human CD137 fused to Fc (huCD137-Fc-bio) as an indicator and incubated for 1 h at room temperature. Mixtures of test molecules and indicator were added to the plates and incubated for 20 min at room temperature, followed by five washing steps with 100 μl PBS-0.05%T. A 1:5000 dilution of ExtrAvidin-HRP (Sigma-Aldrich) in PBS-0.1%T-2%BSA was then added to the wells and incubated for 1 hour. After an additional washing step, fluorogenic HRP substrate (QuantaBlu, Thermo) was added to each well ) and fluorescence intensity was determined using a fluorescence microplate reader.

Competition data for an illustrative experiment is shown in FIG. 14, where the x-axis represents the concentration of the test molecule and the y-axis represents the measured concentration of the indicator molecule. The data were fitted to a 1:1 sigmoidal curve, where the IC ₅₀ value and maximum signal were free parameters and the slope was set to unity. The results demonstrate that the exemplary fusion proteins (SEQ ID NOs: 90 and 87 and SEQ ID NOs: 86 and 91) compete with the anti-CD137 antibody of SEQ ID NOs: 28 and 29 for binding to CD137, suggesting that the fusion proteins bind epitopes that overlap with the antibody.

Table 10

Example 18: Assessing T Cell Activation Using a PD-1/PD-L1 Blockade Bioassay.

The potential of selected fusion proteins to block PD-1/PD-L1-mediated suppression was assessed using PD-1-NFAT-luc Jurkat T cells (a Jurkat cell line genetically modified to express PD-1 and the luc gene) NFAT response element (NFAT-RE) driven) co-cultured with PD-L1 aAPC/CHO-K cells (CHOK1 cells expressing human PD-L1 and a genetically modified cell surface protein designed to activate TCR-related in an antigen-independent manner) . In this bioassay, when PD-1-NFAT-luc Jurkat T cells and PD-L1 aAPC/CHO-K1 cells are co-cultured, the PD-1/PD-L1 interaction inhibits TCR signaling and NFAT-RE-mediated luminescence. Addition of a PD-1/PD-L1 blocking agent, such as CD137-PD-L1 specific fusion proteins as described herein, releases the inhibitory signal and leads to TCR activation and NFAT-RE-mediated luminescence.

- 46 046634

PD-L1 aAPC/CHO-K1 cells were grown in Ham's F12 medium supplemented with 10% FCS and seeded at 8.00 x 10 ³ cells per well and left overnight for attachment at 37°C in a humidified atmosphere with 5% CO2. The next day, the culture media were discarded. 1.00 x 10 ⁴ PD-1-NFAT-luc Jurkat T cells were added to each well, followed by the addition of various concentrations, typically ranging from 0.005 nM to 50 nM, of fusion protein (SEQ ID NOs: 90 and 87) or anti-PD antibody -L1 (SEQ ID NO: 86 and 87 or SEQ ID NO: 26 and 27). The plates were covered with a gas-permeable film and incubated at 37°C in a humidified atmosphere with 5% CO2. After 6 h, 30 μl of BioGlo™ reagent was added to each well and the bioluminescence signal was quantified using a luminometer. Four-parameter logistic curve analysis was performed using GraphPad Prism® to calculate EC50 values, which are summarized in Table. 11. The analysis was carried out in triplicate.

The results of an illustrative experiment are shown in FIG. 15.

The data demonstrate that the test fusion protein inhibits PD-1/PD-L1 blockade and activates T cells in a dose-dependent manner with _an EC ₅₀ value comparable to that of the anti-PD-L1 antibody (SEQ ID NOs: 86 and 87 or SEQ ID NO: 26 and 27). The CD137 reference antibody (SEQ ID NOs: 28 and 29) and/or the isotype control antibody (SEQ ID NOs: 24 and 25) used as negative controls did not result in an increase in the luminescence signal.

Table 11

Assessing T Cell Activation Using a PD-1/PD-L1 Blockade Bioassay

SEQ ID NO: EC ₅ o [nM] 90 and 87 0.49 86 and 91 0.58 28 and 29 0.46

Example 19: Assessment of T cell activation using human PBMCs.

To evaluate the ability of selected fusion proteins to co-stimulate T cell responses, an additional T cell assay was used in which various concentrations of fusion proteins were added to SEB-stimulated human PBMCs and incubated for 3 days at 37°C. IL-2 secretion levels were measured in the supernatants.

PBMCs from healthy volunteer donors were isolated and stored as described in Example 11. For analysis, PBMCs were thawed and placed in culture medium (RPMI 1640, Life Technologies) supplemented with 10% FCS and 1% penicillin-streptomycin (Life Technologies) for 24 hours at 37°C in a humidified atmosphere with 5% CO ₂ .

The following procedure was performed in triplicate for each experimental condition: 2.5 x ¹⁰⁴ PBMCs were incubated in each well of 384-well flat-bottom tissue culture plates in culture medium. Serial dilutions of selected fusion protein (SEQ ID NOs: 90 and 87), anti-PD-L1 building block antibody (SEQ ID NOs: 86 and 87), reference anti-CD137 antibody (SEQ ID NOs: 28 and 29) used separately or in combination with a reference anti-PD-L1 antibody (SEQ ID NOs: 26 and 27), or isotype control (SEQ ID NOs: 24 and 25), typically in the range of 0.0002 to 10 nM, and 0.1 ng/ ml of SEB was added to the appropriate wells. The plates were covered with gas-permeable film (4titude) and incubated at 37°C in a humidified atmosphere with 5% CO ₂ for three days. The level of IL-2 in the supernatant was then assessed as described in Example 11.

The results of an illustrative experiment are shown in FIG. 16.

Summary EC50 values of the tested molecules in relation to the induction of IL-2 secretion are given in table. 12. The bispecific fusion protein of SEQ ID NOs: 90 and 87 induces a strong dose-dependent increase in IL-2 secretion to higher levels compared to an anti-PDL1 building block antibody, a reference anti-CD137 antibody, and a cocktail of reference anti-PDL1 and anti-CD137 antibodies. , and also reduces the EC50 efficacy value compared to anti-PD-L1 and anti-CD137 antibodies used alone or in combination.

Table 12

Assessment of T cell activation using human PBMCs

Donor EC ₅₀ [nM] SEQ ID NO: 90 and 87 SEQ ID NO: 86 and 87 SEQ ID NO: 26 and 27 SEQ ID NO: 28 and 29 + SEQ ID NO: 86 and 87 No. 1 0.019 0.061 0.250 0.279 No. 2 0.026 0.057 0.134 0.089

- 47 046634

Example 20: Evaluation of PD-Ll-dependent T cell activation induced by fusion proteins.

PD-L1 target-dependent costimulation of T cells by fusion proteins was further analyzed using a T cell activation assay. The fusion proteins were applied at various concentrations to T cells stimulated with anti-CD3 antibody cocultured with Flp-InCHO cells transfected with human PD-L1 or mock transfected. IL-2 secretion levels were measured in the supernatants.

PBMCs from healthy volunteer donors were isolated from the buffy platelet layer as described in Example 11. T lymphocytes were further purified and stored as described in Example 12.

For analysis, T cells were thawed and placed in culture medium (RPMI 1640, Life Technologies) supplemented with 10% FCS and 1% penicillin-streptomycin (Life Technologies) overnight at 37°C in a humidified atmosphere with 5% CO ₂ .

The following procedure was performed in triplicate for each experimental condition: Flat-bottom tissue culture plates were precoated with 0.25 μg/ml anti-CD3 antibody for 2 h at 37°C and then washed twice with PBS. CHO cells transfected with human PD-L1 or mock-transfected were treated for 30 min with 30 μg/ml mitomycin C (Sigma Aldrich) to block proliferation. Mitomycin-treated cells were then washed twice with PBS and seeded at 1.0 x 107 cells per well in culture medium to allow attachment overnight at 37°C in a humidified 5% CO ₂ atmosphere. CHO cells previously grown under standard conditions were separated using Accutase (PAA Laboratories) and resuspended in culture medium.

On the following days, 8.33 x 103 T cells per well were added to the CHO cells.

Serial dilutions of an exemplary fusion protein of SEQ ID NOs: 90 and 87, an anti-PD-L1 building block antibody (SEQ ID NOs: 86 and 87), a reference anti-CD137 antibody (SEQ ID NOs: 28 and 29), and a cocktail of reference anti-PD-L1 antibody (SEQ ID NO: 26 and 27) and reference anti-CD137 antibody (SEQ ID NO: 28 and 29), or isotype control (SEQ ID NO: 24 and 25), typically in the range of 0.003 nM to 50 nM were added to the appropriate wells. The plates were covered with gas-permeable film and incubated at 37°C in a humidified atmosphere with 5% CO ₂ for 2 days.

After 2 days of co-culture, IL-2 levels in the supernatant were assessed as described in Example 11.

Exemplary data is shown in FIG. 17. Co-culture of pan-T cells with CHO cells transfected with human PD-L1 in the presence of the fusion protein (SEQ ID NOs: 90 and 87 and SEQ ID NOs: 86 and 91) resulted in strong dose-dependent secretion of IL-2 compared with isotype control hIgG4 and much stronger than that of the reference antibodies, where only a small increase in IL-2 secretion was observed for the reference CD137 antibody or a cocktail of the reference CD137 antibody and the reference PD-L1 antibody. When co-cultured with mock-transfected CHO cells (PD-L1 negative), only the CD137 reference antibody and a cocktail of CD137 reference antibody and PD-L1 reference antibody showed a small dose-dependent increase in IL-2 secretion. The results indicate that T cell activation by the fusion proteins is dependent on PD-L1, while the reference CD137 antibody (SEQ ID NOs: 28 and 29) demonstrated CD137-mediated T cell co-stimulation regardless of the presence or absence of target cells.

Example 21: Pharmacokinetics of Fusion Proteins in Mice.

Pharmacokinetics analyzes of the illustrative fusion proteins (SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87, and SEQ ID NOs) were performed : 90 and 91) in mice. Male CD-1 mice approximately 5 weeks old (3 mice per time point; Charles River Laboratories, Research Models and Services, Germany GmbH) were injected into the tail vein with 10 mg/kg fusion protein. The tested drugs were administered as a bolus of 5 ml/kg. Mouse plasma samples were obtained at time points of 5 min, 1 hour, 4 hours, 8 hours, 24 hours, 48 hours, 4 days, 8 days, 14 days, 21 days, and 28 days. Sufficient whole blood was collected—blood drawn under isoflurane anesthesia—to obtain at least 100 μl of Li-heparin plasma per animal and time point. Drug levels were determined using a sandwich ELISA, detecting the entire bispecific construct through the targets PD-L1 and CD137. Data were fitted using a two-compartment model using Prism GraphPad 5 software.

In fig. 18 shows graphs of plasma concentration versus time for the fusion proteins SEQ ID NOs: 90 and 87, SEQ ID NOs: 86 and 91, SEQ ID NOs: 92 and 87, SEQ ID NOs: 86 and 93, SEQ ID NOs: 94 and 87 and SEQ ID NOs: 90 and 91, plotted along with the values obtained for the anti-PD-L1 building block antibody (SEQ ID NOs: 86 and 87) as a reference. Pharmacokinetics were similar in all cases. Starting with a plasma concentration of about 200 μg/ml, plasma levels fell to about 50 μg/ml within 48 hours and then decreased much more slowly to a level of about 10 μg/ml at the end of the experiment after 28 days. Non-compartmental analysis was applied to these data. Summary values for terminal half-lives are given in Table. 13.

Data demonstrate that the fusion proteins have long, antibody-like terminal periods.

- 48 046634 half-life in mice. Because the assay used to determine plasma concentrations of the fusion proteins requires persistence of activity against both PD-L1 and CD137, the result also demonstrates that the bispecific molecules remain unchanged over 28 days.

Table 13

Terminal half-lives in mice determined using non-compartmental analysis

Example 22 Pharmacokinetics of Fusion Proteins in Mice.

The pharmacokinetics of the exemplary fusion protein of SEQ ID NOs: 90 and 87 was analyzed in mice and compared with two previously described CD137- and PD-L1-binding fusion proteins (SEQ ID NO: 147 and SEQ ID NO: 148). Male CD-1 mice approximately 5 weeks old (2 mice per time point; Charles River Laboratories, Research Models and Services, Germany GmbH) were injected into the tail vein with the appropriate molecule at a dose of 2 mg/kg. Mouse plasma samples were obtained at time points of 5 min, 24 h, 168 h, and 336 h. Sufficient whole blood was collected—blood drawn under isoflurane anesthesia—to obtain at least 30–50 μl of Li-heparin plasma per animal and time point.

Plasma drug levels were then analyzed using ELISA. HuCD137-His (human CD137 with a C-terminal polyhistidine tag) was dissolved in PBS (1 μg/ml) and applied to microtiter plates overnight at 4°C. The plate was washed five times after each incubation step with 80 μl PBS supplemented with 0.05% (v/v) Tween 20. The plates were blocked with PBS/BSA/Tween (PBS containing 2% BSA (w/v) and 0 .1% (v/v) Tween 20) for 1 h at room temperature and then washed. Plasma samples were diluted in PBS/BSA/Tween to a plasma concentration of 20%, added to wells, and incubated for 1 h at room temperature. Then another washing step was carried out. After 1 h of incubation, the bound agents of interest were detected with a mixture of biotinylated human PD-L1 and SULFO-TAG streptavidin (Mesoscale Discovery) at a concentration of 1 μg/ml each, diluted in PBS containing 2% BSA (w/v) and 0.1 % (v/v) Tween 20. After an additional washing step, 35 μl of readout buffer was added to each well and the electrochemiluminescence (ECL) signal was read from each well using a Mesoscale Discovery reader. Data were transferred to Excel for analysis and quantification. A calibration curve was constructed with standard protein dilutions.

In fig. 19 shows graphs of plasma concentration versus time for SEQ ID NOs: 90 and 87, SEQ ID NO: 147 and SEQ ID NO: 148. SEQ ID NOs: 90 and 87 demonstrate a favorable pharmacokinetic profile or antibody-like pharmacokinetics, in contrast to SEQ ID NO: 147 and SEQ ID NO: 148. As used herein, a favorable pharmacokinetic profile or antibody-like pharmacokinetics may be considered to have been achieved if the %c _max was greater than 10% after 336 hours.

The embodiments described herein for purposes of illustration may suitably be practiced in the absence of any element or elements, limitation or limitations not specifically disclosed herein. Thus, for example, the terms containing, including, etc. should be interpreted broadly and without limitation. Further, the terms and expressions used herein are used as terms of description and not limitation, and the use of such terms and expressions is not intended to exclude any equivalents of the features shown and described, or portions thereof, but are deemed to be within the scope of the claimed invention. Various modifications are possible. Thus, it should be understood that while the present embodiments have been specifically disclosed by preferred embodiments and optional features, modifications and variations thereof may be made by those skilled in the art, and that such modifications and variations are considered to be within the scope of the present invention. . All patents, patent applications, textbooks, and peer-reviewed publications described herein are hereby incorporated by reference in their entirety. In addition, if the definition or use of a term in a source incorporated herein by reference is inconsistent with or inconsistent with the definition of that term provided herein, the definition of that term provided herein applies and the definition of that term in the source does not apply. Each of the narrower species and subgeneric

- 49 046634 groups falling under the general disclosure also form part of the invention. This includes a general description of the invention with a condition or negative limitation excluding any subject matter from the class, whether or not the deleted material is specifically mentioned herein. Moreover, if features are described in terms of Markush groups, those skilled in the art will understand that the invention is thereby also described in terms of any individual member or subset of members of a Markush group. Additional embodiments will become apparent from the following claims.

Equivalents: Those skilled in the art will understand or be able to determine, using no more than routine experimentation, many equivalents to the specific embodiments described herein. Such equivalents are intended to be covered by the following claims. All publications, patents and patent applications mentioned in this specification are incorporated herein by reference to the same extent as if each individual publication, patent or patent application had been specifically and individually identified as being incorporated herein by reference.

VI. Non-patent sources

1. GATO-CANAS, M., ZUAZO, M., ARASANZ, N., IBANEZ-VEA, M., LORENZO, L., FERNANDEZ-HINOJAL,

G., VERA, R„ SMERDOU, C., MARTISOVA, E., AROZARENA, 1., WELLBROCK, C., LLOPIZ, D., RUIZ, M., SAROBE, P„ BRECKPOT, K., KOCHAN, G. & ESCORS, D. 2017. PDL1 Signals through Conserved Sequence Motifs to Overcome Interferon-Mediated Cytotoxicity. Cell Rep, 20,1818-1829.

2. KARWACZ, K., BRICOGNE, C., MACDONALD, D., ARCE, F., BENNETT, C. L., COLLINS, M. & ESCORS, D. 2011. PD-L1 co-stimulation contributes to ligand-induced T cell receptor down-modulation on CD8+ T cells. EMBO Mol Med, 3, 581-92.

3. ARASANZ, H., GATO-CANAS, M., ZUAZO, M., IBANEZ-VEA, M., BRECKPOT, K., KOCHAN, G. & ESCORS, D. 2017. PD1 signal transduction pathways in T cells. Oncotarget, 8, 51936-51945.

4. PATSOUKIS, N., BARDHAN, K., CHATTERJEE, P„ SARI, D., LIU, B., BELL, L. N., KAROLY, E. D., FREEMAN, G. J., PETKOVA, V., SETH, P„ LI, L . & BOUSSIOTIS, V. A. 2015. PD-1 alters T-cell metabolic reprogramming by inhibiting glycolysis and promoting lipolysis and fatty acid oxidation. Nat Commun, 6, 6692.

5. XU-MONETTE, Z. Y„ ZHANG, M., LI, J. & YOUNG, K. H. 2017. PD-1/PD-L1 Blockade: Have We Found the Key to Unleash the Antitumor Immune Response? Front Immunol, 8.1597.

6. LI, S. Y. & LIU, Y. 2013. Immunotherapy of melanoma with the immune costimulatory monoclonal antibodies targeting CD137. Clin Pharmacol, 5, 47-53.

7. SNELL, L. M., LIN, G. H., MCPHERSON, A. J., MORAES, T. J. & WATTS, T. H. 2011. T-cell intrinsic effects of GITR and 4-1BB during viral infection and cancer immunotherapy. Immunol Rev, 244, 197-217.

8. WYZGOL, A., MULLER, N., FICK, A., MUNKEL, S., GRIGOLEIT, G. U., PFIZENMAIER, K. & WAJANT, H.

2009. Trimer stabilization, oligomerization, and antibody-mediated cell surface immobilization improve the activity of soluble trimers of CD27L, CD40L, 41BBL, and glucocorticoid-induced TNF receptor ligand. J Immunol, 183,1851-61.

9. YAO, S., ZHU, Y. & CHEN, L. 2013. Advances in targeting cell surface signaling molecules for immune modulation. Nat Rev Drug Discov, 12, 130-46.

10. MELERO, 1., BACH, N., HELLSTROM, K. E., ARUFFO, A., MITTLER, R. S. & CHEN, L. 1998.

Amplification of tumor immunity by gene transfer of the co-stimulatory 4-1BB ligand: synergy with the CD28 co-stimulatory pathway. Eur J Immunol, 28, 1116-21.

11. YANG, Y„ YANG, S., YE, Z., JAFFAR, J., ZHOU, Y„ CUTTER, E., LIEBER, A., HELLSTROM, I. & HELLSTROM, K. E. 2007. Tumor cells expressing anti-CD137 scFv induce a tumor-destructive environment. Cancer Res, 67, 2339-44.

12. ZHANG, H., KNUTSON, K. L., HELLSTROM, K. E., DISIS, M. L. & HELLSTROM, I. 2006. Antitumor efficacy of CD137 ligation is maximized by the use of a CD137 single-chain Fv-expressing wholecell tumor vaccine compared with CD137-specific monoclonal antibody infusion. Mol Cancer Ther, 5, 149-55.

13. YE, Z., HELLSTROM, I., HAYDEN-LEDBETTER, M., DAHLIN, A., LEDBETTER, J. A. & HELLSTROM, K. E.

2002. Gene therapy for cancer using single-chain Fv fragments specific for 4-1BB. Nat Med, 8, 343-8.

14. MARTINET, 0., DIVINO, S. M., ZANG, Y„ GAN, Y„ MANDELI, J., THUNG, S., PAN, P. Y. & CHEN, S. H.

2002. T cell activation with systemic agonistic antibody versus local 4-1BB ligand gene delivery combined with interleukin-12 eradicate liver metastases of breast cancer. Gene Ther, 0, 786-92.

15. YE, Q„ SONG, D. G„ POUSSIN, M„ YAMAMOTO, T„ BEST, A., LI, C„ COUKOS, G. & POWELL, D. J., JR. 2014. CD137 accurately identifies and enriches for naturally occurring tumor-reactive T cells in tumor. Clin Cancer Res, 20, 44-55.

16. CHACON, J. A., WU, R. C., SUKHUMALCHANDRA, P„ MOLLDREM, J. J., SARNAIK, A., PILON-

THOMAS, S., WEBER, J., HWU, P. & RADVANYI, L. 2013. Co-stimulation through 4-1BB/CD137 improves the expansion and function of CD8(+) melanoma tumor-infiltrating lymphocytes for adoptive T-cell therapy. PLoS One, 8, e60031.

- 50 046634

17. FISHER, T. S., KAM PERSC H RO ER, C., OLIPHANT, T., LOVE, V. A., LIRA, P. D., DOYONNAS, R„ BERGQVIST, S., BAXI, S. M., ROHNER, A., SHEN, A. C., HUANG, C., SOKOLOWSKI, S. A. & SHARP, L. L. 2012. Targeting of 4-1BB by monoclonal antibody PF-05082566 enhances T-cell function and promotes anti-tumor activity. Cancer Immunol Immunother, 61,1721-33.

18. SKERRA, A. 2000. Lipocalins as a scaffold. Biochim Biophys Acta, 1482, 337-50.

19. FLOWER, D. R., NORTH, A. C. & SANSOM, S. E. 2000. The lipocalin protein family: structural and sequence overview. Biochim Biophys Acta, 1482, 9-24.

20. FLOWER, D. R. 1996. The lipocalin protein family: structure and function. Biochem J, 318 (Pt 1), 1-14.

21. ALTSCHUL, S. F., MADDEN, T. L., SCHAFFER, A. A., ZHANG, J., ZHANG, Z., MILLER, W. & LIPMAN, D. J. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res, 25, 3389-402.

22. ALTSCHUL, S. F., GISH, W., MILLER, W., MYERS, E. W. & LIPMAN, D. J. 1990. Basic local alignment search tool. J Mol Biol, 215, 403-10.

23. SMITH, T. F. & WATERMAN, M. S. 1981. Identification of common molecular subsequences. J Mol Biol, 147, 195-7.

24. WARD, E. S., GUSSOW, D., GRIFFITHS, A. D., JONES, P. T. & WINTER, G. 1989. Binding activities of a repertoire of single immunoglobulin variable domains secreted from Escherichia coli. Nature, 341, 544-6.

25. HOLLIGER, P„ PROSPERO, T. & WINTER, G. 1993. Diabodies: small bivalent and bispecific antibody fragments. Proc Natl Acad Sci USA, 90, 6444-8.

26. JOHNSON, G. & WU, T. T. 2000. Kabat database and its applications: 30 years after the first variability plot. Nucleic Acids Res, 28, 214-8.

27. EHRENMANN, F., KAAS, Q. & LEFRANC, M. P. 2010. IMGT/3Dstructure-DB and IMGT/DomainGapAlign: a database and a tool for immunoglobulins or antibodies, T cell receptors, MHC, IgSF and MhcSF. Nucleic Acids Res, 38, D301-7.

28. BULLIARD, Y„ JOLICOEUR, R„ ZHANG, J., DRANOFF, G., WILSON, N. S. & BROGDON, J. L. 2014. 0X40 engagement depletes intratumoral Tregs via activating FcgammaRs, leading to antitumor efficacy. Immunol Cell Biol, 92, 475-80.

29. BULLIARD, Y„ JOLICOEUR, R„ WINDMAN, M., RUE, S. M., ETTENBERG, S., KNEE, D. A., WILSON, N. S., DRANOFF, G. & BROGDON, J. L. 2013. Activating Fc gamma receptors contribute to the antitumor activities of immunoregulatory receptor-targeting antibodies. J Exp Med, 210.168593.

30. SILVA, J. P„ VETTERLEIN, 0., JOSE, J., PETERS, S. & KIRBY, H. 2015. The S228P mutation prevents in vivo and in vitro lgG4 Fab-arm exchange as demonstrated using a combination of novel quantitative immunoassays and physiological matrix preparation. J Biol Chern, 290, 5462-9.

31. GLAESNER, W., VICK, A. M., MILLICAN, R„ ELLIS, B., TSCHANG, S. H., TIAN, Y„ BOKVIST, K., BRENNER, M., KOESTER, A., PORKSEN, N., ETGEN, G. & BUMOL, T. 2010. Engineering and characterization of the long-acting glucagon-like peptide-1 analogue LY2189265, an Fc fusion protein. Diabetes Metab Res Rev, 26, 287-96.

32. DALL'ACQUA, W. F., KIENER, P. A. & WU, H. 2006. Properties of human IgGls engineered for enhanced binding to the neonatal Fc receptor (FcRn). J Biol Chern, 281, 23514-24.

33. ZALEVSKY, J., CHAMBERLAIN, A. K., HORTON, N. M., KARKI, S., LEUNG, I. W., SPROULE, T. J., LAZAR, G. A., ROOPENIAN, D. C. & DESJARLAIS, J. R. 2010. Enhanced antibody half-life improves in vivo activity. Nat Biotechnol, 28, 157-9.

34. SHIELDS, R. L., NAMENUK, A. K., HONG, K., MENG, Y. G., RAE, J., BRIGGS, J., XIE, D., LAI, J., STADLEN, A., LI, B., FOX , J. A. & PRESTA, L. G. 2001. High resolution mapping of the binding site on human IgGl for Fc gamma RI, Fc gamma RI I, Fc gamma RIH, and FcRn and design of IgGl variants with improved binding to the Fc gamma R. J Biol Chern, 276, 6591-604.

35. ALTSHULER, E. P„ SEREBRYANAYA, D. V. & KATRUKHA, A. G. 2010. Generation of recombinant antibodies and means for increasing their affinity. Biochemistry (Mose), 75,1584-605.

36. HARLOW, E. & LANE, D. 1999. Using antibodies: a laboratory manual, Cold Spring Harbor, N.Y., Cold Spring Harbor Laboratory Press.

-

Claims (19)

37. HARLOW, E. & LANE, D. 1988. Antibodies: a laboratory manual, Cold Spring Harbor, NY, Cold Spring Harbor Laboratory. 38. LI, J., SAI, T., BERGER, M., CHAO, Q., DAVIDSON, D., DESHMUKH, G., DROZDOWSKI, B., EBEL, W., HARLEY, S., HENRY, M. ., JACOB, S., KLINE, B., LAZO, E., ROTELLA, F., ROUTHIER, E., RUDOLPH, K., SAGE, J., SIMON, P„ YAO, J., ZHOU, Y„ KAVURU, M., BONFIELD, T., THOMASSEN, M. J., SASS, P. M., NICOLAIDES, N. C. & GRASSO, L. 2006. Human antibodies for immunotherapy development generated via a human B cell hybridoma technology. Proc Natl Acad Sci USA, 103, 3557-62. 39. KOZBOR, D. & RODER, J. C. 1983. The production of monoclonal antibodies from human lymphocytes. Immunol Today, 4, 72-9. 40. COLE, S. P„ CAMPLING, B. G., LOUWMAN, I. H., KOZBOR, D. & RODER, J. C. 1984. A strategy for the production of human monoclonal antibodies reactive with lung tumor cell lines. Cancer Res, 44, 2750-3. 41. HOLLIGER, P. & HUDSON, P. J. 2005. Engineered antibody fragments and the rise of single domains. Nat Biotechnol, 23,1126-36. 42. PERVAIZ, S. & BREW, K. 1987. Homology and structure-function correlations between alpha 1acid glycoprotein and serum retinol-binding protein and its relatives. FASEBJ, 1, 209-14. 43. SAMBROOK, J. & RUSSELL, D. W. 2001. Molecular cloning: a laboratory manual, Cold Spring Harbor, N.Y., Cold Spring Harbor Laboratory Press. 44. FLOWER, D. R. 2000. Beyond the superfamily: the lipocalin receptors. Biochim Biophys Acta, 1482, 327-36. 45. BREUSTEDT, D. A., KORNDORFER, I. P„ REDL, B. & SKERRA, A. 2005. The 1.8-A crystal structure of human tear lipocalin reveals an extended branched cavity with capacity for multiple ligands. J Biol Chern, 280, 484-93. 46. SCHMIDT, T. G., KOEPKE, J., FRANK, R. & SKERRA, A. 1996. Molecular interaction between the Strep-tag affinity peptide and its cognate target, streptavidin. J Mol Biol, 255, 753-66. 47. VAJO, Z. & DUCKWORTH, W. C. 2000. Genetically engineered insulin analogs: diabetes in the new millennium. Pharmacol Rev, 52,1-9. 48. FUERTGES, F. & ABUCHOWSKI, A. 1990. The clinical efficacy of polyethylene glycol)-modified proteins. Journal of Controlled Release, 11,139-148. 49. DENNIS, M. S., ZHANG, M., MENG, Y. G., KADKHODAYAN, M., KIRCHHOFER, D., COMBS, D. & DAMICO, L. A. 2002. Albumin binding as a general strategy for improving the pharmacokinetics of proteins. J Biol Chern, Τ1Ί, 35035-43. 50. KONIG, T. & SKERRA, A. 1998. Use of an albumin-binding domain for the selective immobilization of recombinant capture antibody fragments on ELISA plates. J Immunol Methods, 218, 73-83. 51. OSBORN, B. L., OLSEN, H. S., NARDELLI, B., MURRAY, J. H., ZHOU, J. X., GARCIA, A., MOODY, G., ZARITSKAYA, L. S. & SUNG, C. 2002. Pharmacokinetic and pharmacodynamic studies of a human serum albumin-interferon-alpha fusion protein in cynomolgus monkeys. J Pharmacol Exp Ther, 303, 540-8. 52. LOWMAN, Η. B. 1997. Bacteriophage display and discovery of peptide leads for drug development. Annu Rev Biophys Biomol Struct, 26, 401-24. 53. RODI, D. J. & MAKOWSKI, L. 1999. Phage-display technology-finding a needle in a vast molecular haystack. Curr Opin Biotechnol, 10, 87-93. 54. VENTURI, M., SEIFERT, C. & HUNTE, C. 2002. High level production of functional antibody Fab fragments in an oxidizing bacterial cytoplasm. J Mol Biol, 315,1-8. 55. BRUCKDORFER, T., MARDER, O. & ALBERICIO, F. 2004. From production of peptides in milligram amounts for research to multi-tons quantities for drugs of the future. Curr Pharm Biotechnol, 5, 29-43. CLAIM 1. A fusion protein capable of binding both CD137 and PD-L1, where the fusion protein contains at least two subunits, where the first subunit contains a full-length immunoglobulin that is specific for PD-L1, and where the second subunit contains a lipocalin mutein, which is specific for CD137, wherein the fusion protein contains amino acid sequences having at least 90% sequence identity with the amino acid sequences set forth in SEQ ID NOs: 90 and 87, wherein the immunoglobulin heavy chain contains the following set of CDR sequences: GFSLSNYD (HCDR1, SEQ ID NO: 60), IWTGGAT (HCDR2, SEQ ID NO: 61) and VRDSNYRYDEPFTY (HCDR3; SEQ ID NO: 62), and the immunoglobulin light chain contains the following set of CDR sequences: QSIGTN (LCDR1, SEQ ID NO: 63), YAS (LCDR2) and QQSNSWPYT (LCDR3; SEQ ID NO: 64), and wherein the amino acid sequence of the lipocalin mutein has at least 90% sequence identity with the amino acid sequence of SEQ ID NO: 42. 2. The fusion protein according to claim 1, characterized in that (i) said fusion protein is capable of binding PD-L1 with a K _D value of no more than approximately - 52 046634 specifically 2 nM, or a value comparable to or lower than the K _D value of the immunoglobulin or its antigen-binding domain included in the specified first subunit, by itself; (ii) said fusion protein is capable of binding CD137 with a K _D value of no more than about 7 nM, or a value comparable to or lower than the K _D value of the CD137-specific lipocalin mutein included in said second subunit by itself; (iii) said fusion protein is capable of binding PD-L1 with an EC50 value of no more than about 0.5 nM, or a value comparable to or lower than the EC50 value of the immunoglobulin or its antigen binding domain included in the specified first subunit, by itself; (iv) said fusion protein is capable of binding CD137 with an EC ₅₀ value of no more than about 0.6 nM, or a value comparable to or lower than the EC ₅₀ value of the CD137-specific lipocalin mutein included in said second subunit alone; (v) said fusion protein is cross-reactive with cynomolgus PD-L1; (vi) said fusion protein is cross-reactive with cynomolgus CD137; (vii) said fusion protein is capable of simultaneously binding CD137 and PD-L1 with EC ₅₀ values of no more than about 10 nM when said fusion protein is assessed in an ELISA assay; (viii) said fusion protein is capable of binding CD137 expressed on the cell with EC ₅₀ values of no more than about 60 nM; (ix) said fusion protein is capable of binding PD-L1 expressed on a cell with EC ₅₀ values of no more than about 10 nM when said fusion protein is assessed in a flow cytometry assay; (x) said fusion protein is capable of binding PD-L1 expressing tumor cells; (xi) said fusion protein is capable of binding CD137 in the presence of a CD137 ligand; (xii) said fusion protein is capable of competing with PD-1 for binding to PD-L1; (xiii) said fusion protein is capable of competing with the antibody shown in SEQ ID NOs: 28 and 29 for binding to CD137; (xiv) said fusion protein has an overlapping CD137-binding epitope with the antibody set forth in SEQ ID NOs: 28 and 29; (xv) said fusion protein is capable of stimulating T cell proliferation and/or responses; (xvi) said fusion protein is capable of stimulating the proliferation of CD4+ and/or CD8+ T cells; (xvii) said fusion protein is capable of inducing increased secretion of IL-2 and/or IFNgamma; (xviii) said fusion protein is capable of inducing increased secretion of cytotoxic factors; (xix) said fusion protein is capable of co-stimulating T cell responses in a PD-L1 dependent manner; (xx) said fusion protein is capable of co-stimulating T cell responses in the tumor microenvironment; (xxi) said fusion protein is unable to co-stimulate T cell responses in the absence of PD-L1; or (xxii) said fusion protein is capable of blocking a PD-1 inhibitory signal. 3. The fusion protein according to claim 1 or 2, in which the amino acid sequence of the lipocalin mutein is contained at positions corresponding to positions 28, 36, 40-41, 49, 52, 65, 68, 70, 72-73, 77, 79, 81 , 83, 87, 94, 96, 100, 103, 106, 125, 127, 132 and 134 linear polypeptide sequence of mature human neutrophil gelatinase-associated lipocalin (hNGAL) shown in SEQ ID NO: 2, at least ten of the following mutated amino acid residues: Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg or Lys; Gln 49>Val, He, His, Ser or Asn; Tyr 52>Met; Asn 65>Asp; Ser 68>Met, Ala or Gly; Leu 70>Ala, Lys, Ser or Thr; Arg 72>Asp; Lys 73>Asp; Asp 77>Met, Arg, Thr or Asn; Trp 79>Ala or Asp; Arg 81>Met, Trp or Ser; Phe 83>Leu; Cys 87>Ser; Leu 94>Phe; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu and Lys 134>Tyr. 4. The fusion protein according to any one of claims 1 to 3, wherein (i) the amino acid sequence of the lipocalin mutein contains the following set of mutated amino acid residues compared to the linear polypeptide sequence of mature hNGAL presented in SEQ ID NO: 2: Gln 28>His; Leu 36>Gln; Ala 40>Ile; Ile 41>Arg; Gln 49>Ile; Tyr 52>Met; Asn 65>Asp; Ser 68>Met; Leu 70>Lys; Arg 72>Asp; Lys 73>Asp; Asp 77>Met; Trp 79>Asp; Arg 81>Trp; Cys 87>Ser; Asn 96>Lys; Tyr 100>Phe; Leu 103>His; Tyr 106>Ser; Lys 125>Phe; Ser 127>Phe; Tyr 132>Glu; and Lys 134 >Tyr; (ii) the amino acid sequence of the lipocalin mutein has at least 95% sequence identity with the amino acid sequence of SEQ ID NO: 42; or (iii) the amino acid sequence of the lipocalin mutein comprises the amino acid sequence of SEQ ID NO: 42. 5. The fusion protein according to any one of claims 1 to 4, wherein the second subunit is linked at the N-terminus via a linker to the C-terminus of the constant region ka - 53 046634 of each heavy chain (CH) of the first subunit, where the linker is preferably the linker of SEQ ID NO: 13. 6. The fusion protein according to any one of claims 1 to 5, wherein the immunoglobulin comprises a heavy chain variable region and a light chain variable region comprising the amino acid sequences set forth in SEQ ID NOs: 77 and 82. 7. The fusion protein according to any one of claims 1 to 6, wherein the immunoglobulin contains a heavy chain and a light chain containing the amino acid sequences presented in SEQ ID NOs: 86 and 87. 8. The fusion protein according to any one of claims 1 to 6, wherein the immunoglobulin has an IgG4 backbone, wherein said IgG4 backbone preferably has one or more of the following mutations: S228P, N297A, F234A, L235A, M428L, N434S, M252Y, S254T and T256E . 9. A fusion protein according to any one of claims 1 to 8, characterized in that (i) said fusion protein contains amino acid sequences having at least 92%, at least 95%, at least 97%, at least 98 % or greater sequence identity with the amino acid sequences presented in SEQ ID NO: 90 and 87; or (ii) said fusion protein contains the amino acid sequences set forth in SEQ ID NOs: 90 and 87. 10. A nucleic acid molecule comprising a nucleotide sequence encoding a fusion protein according to any one of claims 1 to 9, where preferably (i) said nucleic acid molecule is operably linked to a regulatory sequence to enable expression of said nucleic acid molecule; or (ii) said nucleic acid molecule is contained in a vector or phagemid vector. 11. A host cell containing the nucleic acid molecule according to claim 10. 12. A method for producing a fusion protein according to any one of claims 1 to 9, wherein the fusion protein is produced starting from a nucleic acid encoding the fusion protein. 13. The method of claim 12, wherein the fusion protein is produced in a bacterial or eukaryotic host organism and isolated from the host organism or a culture thereof. 14. The use of a fusion protein according to any one of claims 1 to 9 or a composition containing such a fusion protein for (i) simultaneous activation of CD137 downstream signaling pathways and binding of PD-L1 positive tumor cells; (ii) simultaneous costimulation of T cells and binding of PD-L1 positive tumor cells, comprising applying the fusion protein or a composition containing such a fusion protein to tissue containing the tumor; (iii) simultaneous induction of lymphocyte activity and binding of PD-L1-positive tumor cells, comprising applying the fusion protein or a composition containing such a fusion protein to tissue containing the tumor; (iv) inducing clustering and activation of CD137 on T cells and directing said T cells to PD-L1-positive tumor cells, comprising applying the fusion protein or a composition containing such a fusion protein to tissue containing the tumor; (v) inducing a localized lymphocyte response in the vicinity of PD-L1-positive tumor cells, comprising applying the fusion protein or a composition containing such a fusion protein to tissue containing the tumor; (vi) inducing increased secretion of IL-2 and/or cytotoxic factors by T cells in the vicinity of PDL1-positive tumor cells, comprising applying a fusion protein or a composition containing such a fusion protein to tissue containing a tumor, wherein said cytotoxic factors are preferably selected from the group , consisting of perforin, granzyme B and granzyme A; (vii) inducing increased secretion of cytotoxic factors by T cells in the vicinity of PD-L1 positive tumor cells, comprising applying the fusion protein or a composition containing such a fusion protein to tissue containing the tumor; or (viii) preventing, ameliorating, or treating PD-L1-positive cancer. 15. Pharmaceutical composition containing the fusion protein according to any one of claims 1 to 9. 16. Use of the fusion protein according to any one of claims 1 to 9 for therapy. 17. Use according to claim 16 for the treatment of cancer. 18. Use of the fusion protein according to any one of claims 1 to 9 for the production of a medicinal product. 19. Use according to claim 18, wherein said medicinal product is intended for the treatment of cancer. -