A human protein scaffold library based on the n-terminal domain of human fact complex subunit SSRP1

Abstract

The present invention provides a method of constructing a library of binder scaffolds (library of protein scaffolds) comprising the steps a) providing an initial polypeptide, wherein said initial polypeptide comprises or consists of a polypeptide having at least 90% identity to SEQ ID NO: 1, and b) introducing diversity into copies of said initial polypeptide to form the binder scaffold library (protein scaffold library).

Description

[0001] Title

[0002] A human protein scaffold library based on the N-terminal domain of human FACT complex subunit SSRP1

[0003] Field of the invention

[0004] The present invention generally relates to protein scaffolds having the ability to bind to targets such as antigens, in particular the present invention is directed to a human binder scaffold (protein scaffold) based on the N-terminal domain of human FACT (facilitates chromatin transcription) complex subunit SSRP1 (structure-specific recognition protein-1).

[0005] Background of the invention

[0006] Monoclonal antibodies or antigen binding fragments thereof are the most widely used class of therapeutic proteins when high affinity and specificity for a target molecule are desired. However, non-antibody proteins that can be engineered to bind such targets are also of high interest in the biopharmaceutical industry. These "alternative scaffold" proteins may have advantages over traditional antibodies due to their small size, lack of disulfide bonds, high stability, and ability to be expressed in prokaryotic hosts.

[0007] Chimeric antigen receptor (CAR) T-cells are being used clinically in order to treat patients e.g. suffering from cancer. Within this treatment the patients' autologous T cells are obtained using leukapheresis. Subsequently, the patients' T cells are genetically transduced with a CAR molecule, which will recognize e.g. a tumor associated antigen (TAA) on the patients' cancer cells (Majzner and Mackall, 2019). A CAR molecule is usually made up of an antigen-binding domain, a hinge and transmembrane region, a co-stimulatory and stimulatory domain (Majzner and Mackall, 2019). All of these domains are crucial in order to achieve a functional CAR T- cell and cytotoxic activity. Clearly, it is of uttermost importance for the functionality of this treatment that the CAR T-cell will remain functional inside the body of a patient. Unfortunately, this is not always the case and the genetically modified T-cells can become „exhausted“ and will thus show reduced proliferation rates as well as performance (Long et al., 2015, Nat Med 21, 581-590; Majzner and Mackall, 2019, Nat Med 25, 1341-1355). One phenomenon, which is known to be associated with T cell exhaustion, is tonic signaling. Tonic signaling describes T-cell signaling in the absence of an actual TAA and might lead to nonfunctional, exhausted T cells (Ajina and Maher, 2018, Mol Cancer Ther 17, 1795-1815; Long et al., 2015, Nat Med 21, 581-590). Antigen-binding domains within CAR molecules are usually made up of single chain variable fragments (scFvs) (Ajina and Maher, 2018, Mol Cancer Ther 17, 1795-1815; Labanieh et al., 2018, Nat Biomed Eng 2, 377-391; Long et al., 2015, Nat Med 21, 581-590; Majzner and Mackall, 2019, Nat Med 25, 1341-1355), which are prone to dimerize or even oligomerize (Ajina and Maher, 2018, Mol Cancer Ther 17, 1795- 1815; Arndt et al., 1998, Biochemistry 37, 12918-12926; Salzer et al., 2020, Nat Commun 11, 4166; Todorovska et al., 2001, J Immunol Methods 248, 47-66; Worn and Pluckthun, 2001, J Mol Biol 305, 989-1010) and thus might be more likely to trigger CAR clustering and - as a consequence - tonic signaling and T-cell exhaustion (Ajina and Maher, 2018, Mol Cancer Ther 17, 1795-1815; Long et al., 2015, Nat Med 21, 581-590; Salzer et al., 2020, Nat Commun 11, 4166).

[0008] There is a range of „ alternative binder scaffolds“ (or “engineered binder scaffolds”) available that are usable as binders (antigen binding domains) of a CAR, e.g., DARPins, monobodies, affibodies or nanobodies (reviewed by Zajc et al., 2021, FEBS J 288, 2103-2118). Though, only monobodies-from these prominent examples mentioned-are in fact of human origin.

[0009] As CAR molecules will be used clinically, it is highly beneficial if all protein components used are of human origin to reduce the risk of immunogenicity. In fact, it has been demonstrated that recognition of non-human CAR components such as murine scFvs by the host immune system can lead to immune-mediated rejection of CAR T cells and therefore poor anti -turn or activity, as well as limited persistence and expansion of the CAR T cells in the patients (Turtle et al., J Clin Invest. 2016 Jun 1; 126(6):2123-38).

[0010] There is a need in the art for an improved or alternative human binder scaffold library (protein scaffold library) that may provide binders usable e.g. as antigen binding domains in CARs.

[0011] Brief description of the invention

[0012] The inventors aimed at the development of a novel binder scaffold (protein scaffold) that may be engineered to yield a binder e.g. usable as antigen binding domain (component) of chimeric antigen receptors that bind an antigen.

[0013] The novel binder scaffold (protein scaffold) should fulfill most or all of the following properties, especially with regard to be usable for CARs:

[0014] (i) The novel binder scaffold (protein scaffold) should be a human protein domain to reduce the risk of immunogenicity;

[0015] (ii) the binder scaffold (protein scaffold) should show high expression levels on primary human T cells when fused to a CAR backbone;

[0016] (iii) the binder scaffold (protein scaffold) should be a single domain protein to prevent domain mispairing as is observed with scFvs; (iv) it should show favorable biochemical properties, such as a high midpoint of thermal denaturation (Zm) above 70 °C (since these proteins will be further engineered for antigen recognition, their thermostability will most probably be reduced (Teufl et al., ACS Synth Biol. 2022 Mar 18; 11(3): 1030-1039), thus high initial thermostability was a crucial requirement here);

[0017] (v) the novel binder scaffold (protein scaffold) should lack disulfide bonds as well as free cysteines in order to enable intra- and extracellular applications and

[0018] (vi) it should lack N-glycosylation sites to enable efficient translation from one expression host to another without altering protein characteristics;

[0019] (vii) it should be derived from an intracellular protein, because engineered proteins used therapeutically in humans sometimes induce antibody-mediated immune responses, which sometimes cross-react with the non-mutated, endogenous protein (Saxton et al., Nat Rev Drug Discov. 2023 Jan;22(l):21-37); thus, to avoid this cross-reactivity of potential antibody responses with endogenous wild type counterparts, the novel binder scaffold (protein scaffold) should be derived from an intracellular protein; in that case, potential cross-reactive antibodies cannot bind to the endogenous, non-mutated protein, because it is located intracellularly and therefore not accessible to antibodies.

[0020] The inventors surprisingly found that the N-terminal domain of human FACT complex subunit SSRP1 matches above listed criteria.

[0021] Therefore, the present invention provides a method of constructing a library of binder scaffolds (library of protein scaffolds) based on a part of the amino acid sequence of the human N- terminal domain of human SSRP1. PDB entry “5UMR” represents the crystal structure of the N-terminal domain of SSRP1, but additionally contains the sequence "GH" at the N-terminal end of the 5UMR sequence, which is not part of SSRP1 (obviously derived from a linker or cloning site). The 2 additional amino acids at the beginning of the 5UMR sequence may be potentially immunogenic in humans because they are not part of a human sequence, so the inventors dismiss these 2 amino acids from the relevant sequence set forth in SEQ ID NO: 1. The 5UMR sequence without the N-terminal amino acids GH represent the first 100 amino acids of human SSRP1 as annotated in UniProt (ID: Q08945). An initial polypeptide that comprises or consists of a polypeptide having at least 90% identity to said SEQ ID NO: 1 is the starting point for the method of constructing a library of binder scaffolds (library of protein scaffolds) as disclosed herein. The present invention also provides the library of binder scaffolds (library of protein scaffolds) obtained by the methods as disclosed herein. A method of identifying and / or isolating a binder (antigen binding domain) from the library as disclosed herein binding to a target such as an antigen is also disclosed herein.

[0022] In addition, herein it is disclosed an isolated binder (antigen binding domain) having at least 75% identity to SEQ ID NO: 1 but with substitutions at least at some defined positions, a fusion protein comprising said isolated binder, and a CAR comprising an antigen binding domain that is or comprises a binder obtained by the method as disclosed herein. The use of the N-terminal domain of human FACT complex subunit SSRP1 (SEQ ID NO: 1) for the generation (construction) of a library of binder scaffolds (library of protein scaffolds) is disclosed as well. Exemplary binders (antigen binding domains) were generated by using the method as disclosed herein that are specific for CD22 and CD276. But no limitation to these exemplary generated binders (antigen binding domains) is intended.

[0023] Brief description of the drawings

[0024] Figure 1 : (A) Expression of scaffold-CARs on primary human T cells; (B) Thermostability of soluble scaffolds

[0025] Figure 2: SEC-HPLC analysis of proteins 3 SHU, 2P9R, 5UMR and FN3

[0026] Figure 3: (A): Protein structure including to-be-randomized positions in 5UMR libraries highlighted, (B) Full-length expression of NNK -randomized 5UMR libraries on yeast cells.

[0027] Figure 4: Intended amino acid frequencies in randomly mutated binding surface

[0028] Figure 5: C-myc expression of NNK- vs. improved yeast libraries

[0029] Figure 6: (A) AA frequency from library sequences vs. initial design for randomized 5UMR positions except Q28 and Q44; (B) AA frequency from library sequences vs. initial design for randomized 5UMR positions Q28 and Q44

[0030] Figure 7: (A) Affinities of CD22 binders [nM]; (B) Titration of CD22 binders on targetexpressing NALM6 cells

[0031] Figure 8: (A) Thermostability of CD22 binders; (B) SEC-HPLC analysis of CD22 binders Figure 9: (A) SEC-HPLC analysis of CD276 binders; (B) Thermostability of CD276 binders Figure 10: (A) Titration curves of CD276 binders on A-431 cells; (B) Titration curves of CD276 binders on A549 cells; (C) Titration curves of CD276 binders on SK-BR-3 cells; (D) Titration curves of CD276 binders on Caco-2 cells Figure 11 : (A) Affinities of CD276 binders [pM]; (B) Residual binding [%] of CD276 binders C152, 152 and C146 upon blocking with MGA271 and Omburtamab; (C) Specificity of CD276 binders assessed by binding to Jurkat and Caco-2 cells;

[0032] Figure 12: (A): Expression of aCD22-CARs on Nur77 Jurkat reporter cells; (B) Activation of aCD22-CARs on Nur77 Jurkat reporter cells with and without NALM6 cells s

[0033] Figure 13: (A) Cytotoxicity assay of aCD22-CAR T cells with NALM6 target cells; (B) Cytotoxicity assay of aCD22-CAR T cells with Raji target cells

[0034] Figure 14: (A) Release of IFN-y by aCD22-CAR T cells upon co-incubation with NALM6 cells;

[0035] (B) Release of IL-2 by aCD22-CAR T cells upon co-incubation with NALM6 cells

[0036] Figure 15: (A) Release of IFN-y by aCD22-CAR T cells upon co-incubation with Raji cells;

[0037] (B) Release of IL-2 by aCD22-CAR T cells upon co-incubation with Raji cells

[0038] Figure 16: (A) Expression of aCD276-CARs on Nur77 Jurkat reporter cells; (B) Activation of aCD276-CARs on Nur77 Jurkat reporter cells with and without A549 cells

[0039] Figure 17: (A) Cytotoxicity of aCD276-CAR T cells with target cells A431, A549 or Caco-2 after six hours of co-incubation; (B) Cytotoxicity of aCD276-CAR T cells with target cells A431, A549 or Caco-2 after 24 hours of co-incubation

[0040] Figure 18: (A) Release of IFN-y by aCD276-CAR T cells upon six hours of co-incubation with A431, A549 or Caco-2 cells; (B) Release of IL-2 by aCD276-CAR T cells upon six hours of co-incubation with A431, A549 or Caco-2 cells;

[0041] Figure 19: (A) Release of IFN-y by aCD276-CAR T cells upon 24 hours of co-incubation with A431, A549 or Caco-2 cells; (B) Release of IL-2 by aCD276-CAR T cells upon 24 hours of co-incubation with A431, A549 or Caco-2 cells;

[0042] Figure 20: Specificity of CD22 binders assessed by binding to Jurkat, NALM6 and Raji cells;

[0043] Detailed description of the invention

[0044] In a first aspect the present invention provides a method of constructing a library of binder scaffolds (protein scaffolds) comprising the steps a) providing an initial polypeptide, wherein said initial polypeptide comprises or consists of a polypeptide having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO:1, b) introducing diversity into copies of said initial polypeptide to form the binder scaffold library (protein scaffold library).

[0045] In one embodiment of the invention said initial polypeptide comprises or is SEQ ID NO:1. Said method, wherein the introducing diversity step comprises mutating one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO:1.

[0046] Said method, wherein the introducing diversity step comprises mutating one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO:1 but does not comprise the introduction of any cysteine at said one or more amino acid positions.

[0047] Said method, wherein the introducing diversity step comprises mutating at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.

[0048] Said method, wherein the introducing diversity step comprises mutating at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, but does not comprise the introduction of any cysteine at said least 5, at least 6, at least 7, at least 8, at least 9 or 10 of said amino acid positions.

[0049] Said method as disclosed herein, wherein other positions than R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 may also be modified.

[0050] Said method as disclosed herein, wherein other positions than R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 may not be modified.

[0051] Said method, wherein not more than 10, not more than 9, not more than 8, not more than 7, not more than 6, not more than 5, not more than 4, not more than 3, not more than 2 other positions than R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are modified. Said method, wherein said binder (protein) scaffolds (said copies of said initial polypeptide) comprise an amino acid sequence that have at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1, respectively.

[0052] Said method, wherein said introducing diversity step is performed by using random mutagenesis on the level of the nucleic acid sequence encoding said polypeptide having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1.

[0053] Random mutagenesis describes the process of altering the nucleic acid sequence in an undirected manner. Random mutagenesis of a gene can be achieved by a range of different technologies, including (a) employing oligonucleotides containing randomly mutated codon positions encoded by, e.g., NNN, NNK, NNS and / or NNB, (b) employing oligonucleotides containing randomly mutated codon positions synthesized by trimer synthesis, (c) error prone PCR, among other methods well known to a person skilled in the art.

[0054] Random mutagenesis with NNN, NNK, NNS and / or NNB encoding oligonucleotides are a standard technique. Respective oligonucleotides can be purchased through several vendors, e.g. Sigma-Aldrich, and used for a PCR. Within the IUPAC nucleotide code “N” encodes for any of the bases adenine, cytosine, guanine or thymine. Thus, NNN randomized codons encode for all 64 possible codons in the genetic code, including three stop codons. “K” encodes for bases guanine or thymine, “S” encodes for bases guanine or cytosine and “B” encodes for bases cytosine, guanine or thymine, thus NNK, NNS and NNB encode for fewer codons compared to NNN. NNN, NNK, NNS and / or NNB can be used for saturation mutagenesis, wherein an amino acid position of interest will be randomized to encode for all 20 proteinogenic amino acids. NNN, NNK, NNS and / or NNB randomization are methods which have been used heavily in the field of protein engineering (Koide et al., J Mol Biol. 1998 Dec 11 ;284(4): 1141-51 ; Hasenhindl et al., Protein Eng Des Sei. 2013 Oct;26(10):675-82).

[0055] Random mutagenesis with primers synthesized by trimer synthesis describes a more precise approach compared to NNN, NNK, NNS and / or NNB randomization. Codons are synthesized as trinucleotides, i.e. “trimers”. Each trinucleotide can be used as a codon or “building block” during oligonucleotide synthesis, thus the ratio of the respective codons can be controlled. This technology enables the synthesis of oligonucleotides with a specific codon, thus amino acid, ratio in each amino acid position. Trimer-synthesized oligonucleotides (Kayushin et al., Nucleic Acids Res. 1996 Oct l;24(19):3748-55; Mauriala et al., J Pharm Biomed Anal. 2004 Jan 27;34(1): 199-206; Yagodkin et al., Nucleosides Nucleotides Nucleic Acids. 2007;26(5):473- 97) can be commercially purchased through vendors, e.g. Ella Biotech (https: / / www.ellabiotech.com / ) or Metkinen Chemistry

[0056] (https: / / www.metkinenchemistry.com / ), to be used in a PCR and have been used as a more precise alternative to NNN, NNK, NNS and / or NNB randomization in the field of protein engineering (Traxlmayr et al., J Biol Chem. 2016 Oct 21;291(43):22496-22508; Kelly et al., J Mol Biol. 2018 Jan 5;430(l): 119-130).

[0057] Random mutagenesis employing error prone PCR is a heavily used technology in the field of protein engineering (Stern et al., ACS Comb Sci. 2019 Mar 11;21(3):207-222; Wagner et al., J Mol Biol. 2021 Nov 5;433(22): 167210; Wagner et al., Biochemistry. 2022 Oct 4;61(19):2049- 2062) and results in the insertion of random mutations at random sites within the gene in an undirected fashion. Increased MgCh concentration, the addition of MnCE, a non-proofreading polymerase or dNTP analogs can lead to the introduction of “errors”, i.e. mutations.

[0058] Said method, wherein said introducing diversity step is performed by using random mutagenesis on the level of the nucleic acid sequence encoding said polypeptide having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1, and wherein the introducing diversity step comprises mutating at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, but does not comprise the introduction of any cysteine at said one or more amino acid positions.

[0059] Said method, wherein said introducing diversity step is performed by using random mutagenesis on the level of the nucleic acid sequence encoding said polypeptide having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1 with degenerate oligonucleotides containing randomly mutated codons at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, a) wherein said randomly mutated codon positions within the oligonucleotides are encoded by NNN, NNK, NNS and / or NNB codons, and / or b) wherein said randomly mutated codon positions within the oligonucleotides are synthesized by trimer synthesis.

[0060] Said method, wherein said randomly mutated codon positions within the oligonucleotides are synthesized by trimer synthesis, and said oligonucleotides do not code for any cysteine at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.

[0061] Said method, wherein said randomly mutated codon positions within the oligonucleotides are synthesized by trimer synthesis, and said oligonucleotides code for a reduced percentage of prolines and a higher percentage of tyrosines as compared to the codon degeneracy at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.

[0062] Said method, wherein said reduced percentage of prolines is defined by about 4% or less, about 3% or less, about 2% or less or about 1% or less prolines that are incorporated into each individual amino acid position at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, and wherein said higher percentage of tyrosines is defined by about 10% or more, 15% or more or 20% or more tyrosines that are incorporated into each individual amino acid position of at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO:1.

[0063] Said method, wherein said reduced percentage of prolines is defined by about 4% to about 1% or by about 3% to about 2%, or by about 2% to about 1% prolines that are incorporated into each individual amino acid position at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, and wherein said higher percentage of tyrosines is defined by about 10% to about 20% or about 15% to about 20% or about 10% to about 15% tyrosines that are incorporated into each individual amino acid position of at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.

[0064] Said method, wherein said method comprises step c: displaying the diversified copies of said initial polypeptide in a yeast surface display, phage display, mRNA display, DNA display, bacterial display, mammalian cell display or ribosome display format.

[0065] In all of those display formats, the basic principle is the coupling of genotype and phenotype. That is, a randomly mutated polypeptide library is screened for a certain property (e.g. antigen binding) and after successfully selecting or screening, the coupled nucleic acid (DNA or RNA), which encodes for the displayed protein, is amplified and sequenced in order to get information about the sequence of the displayed protein or polypeptide that bound to the antigen.

[0066] The most widely used methods include phage display, yeast surface display and ribosome display.

[0067] In the case of phage display, randomly mutated versions of the protein of interest (i.e. of the protein that shall be engineered) are fused to a phage surface protein, usually the pill phagecoat protein, in order to display them on the surface of the phage particle. After selection for antigen binding, the selected phages are amplified and the plasmids encoding for the enriched protein variants are sequenced (Ponsel et al., Molecules 2011, 16, 3675-3700).

[0068] In the case of yeast surface display (also known as yeast display), randomly mutated versions of the protein of interest (i.e. of the protein that shall be engineered) are fused to a yeast surface protein, usually the Aga2p protein, in order to display them on the surface of the yeast cell. Yeast surface display libraries can then be selected or screened for desired properties such as antigen binding by using e.g. antigen-coated magnetic beads or flow cytometric sorting. After amplification of enriched yeast cells, plasmids encoding for the enriched protein variants are sequenced (Chao et al., Nat Protoc. 2006;l(2):755-68; Angelini et al., Methods Mol Biol. 2015;1319:3-36).

[0069] In the case of ribosome display, DNA encoding randomly mutated versions of the protein of interest (i.e. of the protein that shall be engineered) is transcribed to mRNA in vitro and subsequently translated to the corresponding polypeptides in vitro. The resulting polypeptide molecules are held at the ribosomes together with the mRNA molecules, thus forming complexes comprising mRNA, ribosome and polypeptide. These complexes are selected for desired functions (e.g. antigen binding), followed by reverse transcription of the mRNA into DNA, DNA amplification and DNA sequencing in order to obtain the sequence of the encoded polypeptide that bound to the antigen (Amstutz et al., Curr Opin Biotechnol. 2001 Aug;12(4):400-5).

[0070] In another aspect the present invention provides a library of binder (protein) scaffolds comprising diversified copies of an initial polypeptide, wherein said initial polypeptide comprises or consists of a polypeptide having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1.

[0071] In one embodiment of the invention said library of binder (protein) scaffolds may be obtained by the methods of constructing a library of binder (protein) scaffolds as disclosed herein.

[0072] In a further aspect the present invention provides a method of isolating a binder (antigen binding domain) that binds to a target (e.g. an antigen), comprising contacting the library as disclosed herein with said target and isolating a binder (antigen binding domain) that binds to said target. Said method of isolating a binder (antigen binding domain), wherein said binder (antigen binding domain) may be identified after said isolation.

[0073] Said method of isolating a binder (antigen binding domain), wherein a random mutagenesis step is performed after said isolation of said binder (non-modified binder; low affinity binder) (antigen binding domain), and subsequently isolating and identifying a binder (antigen binding domain) that binds to said target with a higher affinity to said target (high affinity binder) compared to the non-modified binder (low affinity binder) (non-modified antigen binding domain).

[0074] Said random mutagenesis may be performed by error prone PCR. The procedure of achieving said high affinity binder (high affinity antigen binding domain) may comprise first PCR, followed by an error prone PCR, followed by a final PCR; or may comprise error a prone PCR, followed by a final PCR.

[0075] Said method of isolating a binder (antigen binding domain) that binds to a target, wherein said method comprises contacting the library as disclosed herein with said target and isolating a binder (antigen binding domain) that binds to said target within a predefined affinity range.

[0076] Said method of isolating a binder (antigen binding domain), wherein the isolating step comprises isolating binder molecules (antigen binding domain molecules) binding to said target and testing the isolated binder molecules (antigen binding domain molecules) for binding affinity to said target, wherein optionally:

[0077] (i) the isolating step comprises incubating the library with said target, identifying binder molecules (antigen binding domain molecules) binding to said target, and identifying the sequences encoding for the binder molecules (antigen binding domain molecules) binding to said target; and / or

[0078] (ii) the affinity is defined by a KD less than 10'5M.

[0079] Said method of isolating an antigen binding domain, that binds to an antigen, comprising contacting the library as disclosed herein with said antigen and isolating an antigen binding domain that binds to said antigen, and wherein said isolated antigen binding domain does not comprise any cysteine.

[0080] Said method of isolating an antigen binding domain, that binds to an antigen, comprising contacting the library as disclosed herein with said antigen and isolating an antigen binding domain that binds to said antigen, and discarding isolated antigen binding domains that comprise any cysteine.

[0081] In another aspect the present invention provides a fusion protein comprising a) a binder (antigen binding domain) as a first polypeptide, wherein said binder (antigen binding domain) may be obtained by the method of isolating a binder (antigen binding domain) as disclosed herein, and b) a second polypeptide.

[0082] Said fusion protein, wherein said second polypeptide is an effector molecule.

[0083] Said effector molecule may be an intracellular portion of a receptor, a transmembrane receptor that comprises one or several extracellular antigen binding domain(s), a transmembrane protein that comprises one or several intracellular signaling domain(s), a transmembrane receptor that comprises one or several extracellular antigen binding domain(s) and one or several intracellular signaling domain(s), an intracellular receptor, an extracellular protein, a caspase, a kinase, an enzyme or a protease.

[0084] In one embodiment of the invention the fusion protein is a fusion protein comprising a) an antigen binding domain that binds to an antigen as a first polypeptide comprising an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to SEQ ID NO: 1, but in which at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target, wherein said at least 5 of said amino acid positions that are altered do not comprise any cysteine, and b) a second polypeptide, wherein said second polypeptide is an effector molecule, wherein said antigen binding domain of a) is obtained by the method of constructing a library of protein scaffolds as disclosed herein, followed by the method of isolating an antigen binding domain, that binds to an antigen, as disclosed herein, (and wherein the isolated antigen binding domain is the antigen binding domain of a)).

[0085] In a further aspect the present invention provides an isolated binder (antigen binding domain) comprising an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1, but in which at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target.

[0086] Said isolated binder (antigen binding domain), wherein said at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of said amino acid positions that are altered do not comprise any cysteine.

[0087] In one embodiment of the invention the isolated antigen binding domain that binds to an antigen is an isolated antigen binding domain that binds to an antigen, wherein said isolated antigen binding domain is obtained by the steps of the method of constructing a library of protein scaffolds as disclosed herein, followed by the steps of method of isolating an antigen binding domain, that binds to an antigen, as disclosed herein, wherein said isolated antigen binding domain comprises an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to SEQ ID NO: 1, but in which at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind said antigen, and wherein said at least 5 of said amino acid positions that are altered do not comprise any cysteine.

[0088] In another aspect the present invention provides the use of the N-terminal domain of human FACT complex subunit SSRP1 for the generation (construction) of a library of binder scaffolds (protein scaffolds).

[0089] Said use, wherein said N-terminal domain of human FACT complex subunit SSRP1 comprises or consists of a polypeptide (the initial polypeptide) having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1.

[0090] Said use, wherein said N-terminal domain of human FACT complex subunit SSRP1 comprises or consists of a polypeptide (the initial polypeptide) having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1, and wherein said use comprises introducing diversity into copies of said polypeptide (the initial polypeptide) having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1.

[0091] Said use, wherein said N-terminal domain of human FACT complex subunit SSRP1 comprises or consists of a polypeptide (the initial polypeptide) having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1, and wherein said use comprises introducing diversity into copies of said polypeptide (the initial polypeptide) having at least 90%, at least 95%, at least 98%, at least 99% or 100% identity to SEQ ID NO: 1, and wherein said binder (protein) scaffolds (said copies of said initial polypeptide) comprise an amino acid sequence that have at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO:1 (the initial polypeptide), respectively.

[0092] Said use, wherein said binder (protein) scaffolds (said copies of said initial polypeptide) comprise an amino acid sequence that have at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1 (the initial polypeptide), respectively, and wherein said binder (protein) scaffolds (said copies of said initial polypeptide) do not comprise any cysteine.

[0093] In another aspect, the present invention provides an isolated nucleic acid sequence encoding the isolated binder (antigen binding domain) as disclosed herein.

[0094] In another aspect, the present invention provides an isolated nucleic acid vector comprising the isolated nucleic acid sequence as disclosed herein. Said isolated nucleic acid vector may be e.g. a viral vector. The vector may be a DNA vector, an RNA vector, a plasmid vector, a cosmid vector, a herpes virus vector, a measles virus vector, a lentivirus vector, adenoviral vector, or a retrovirus vector, or a combination thereof. In a preferred embodiment, said nucleic acid sequence encoding an isolated binder (antigen binding domain) as disclosed herein may be contained in a lentiviral vector.

[0095] In some embodiments of the invention, the vector further comprises a promoter wherein the promoter is an inducible promoter, a tissue specific promoter, a constitutive promoter, a suicide promoter or any combination thereof.

[0096] In another aspect the present invention provides a prokaryotic or eukaryotic host cell comprising the isolated nucleic acid sequence as disclosed herein, wherein said host cell optionally is at least one selected from E. coli BL21 Star(DE3), other E. coli cell, yeast, COS-1, COS-7, HEK293, BHK21, CHO, BSC-1, Hep G2, 653, SP2 / 0, 293, HeLa, myeloma, or lymphoma cells, an immune cell such as a T cell or a natural killer (NK) cell, or any derivative, immortalized or transformed cell thereof.

[0097] In another aspect the present invention provides a fusion protein comprising a) a binder (antigen binding domain) as a first polypeptide comprising an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1, but in which at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target, and b) a second polypeptide.

[0098] Said fusion protein, wherein said second polypeptide is an effector molecule.

[0099] Said effector molecule may be an intracellular portion of a receptor, a transmembrane receptor that comprises one or several extracellular antigen binding domain(s), a transmembrane protein that comprises one or several intracellular signaling domain(s), a transmembrane receptor that comprises one or several extracellular antigen binding domain(s) and one or several intracellular signaling domain(s), an intracellular receptor, an extracellular protein, a caspase, a kinase, an enzyme or a protease.

[0100] Said fusion protein, wherein said at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of said amino acid positions that are altered do not comprise any cysteine.

[0101] In one embodiment of the invention the fusion protein is a fusion protein comprising a) an antigen binding domain that binds to an antigen as a first polypeptide comprising an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to SEQ ID NO: 1, but in which at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target, wherein said at least 5 of said amino acid positions that are altered do not comprise any cysteine, and b) a second polypeptide, wherein said second polypeptide is an effector molecule, wherein said antigen binding domain of a) is obtained by the method of constructing a library of protein scaffolds as disclosed herein, followed by the method of isolating an antigen binding domain, that binds to an antigen, as disclosed herein, (and wherein the isolated antigen binding domain is the antigen binding domain of a)).

[0102] In a further aspect, the present invention provides a chimeric antigen receptor (CAR) comprising a) an antigen binding domain, wherein said antigen binding domain binds to a target, b) a transmembrane domain, and c) an intracellular signaling domain, wherein said antigen binding domain is or comprises a binder obtained by the method of isolating a binder (antigen binding domain) as disclosed herein.

[0103] Said CAR, wherein said binder (antigen binding domain) binds (specifically) an antigen expressed on a target cell such as a cancer cell or a cell associated with an autoimmune disease or an infected cell.

[0104] Said target may be an antigen expressed on a target cell such as a cancer cell or a cell associated with an autoimmune disease or a cell infected with a virus or a pathogen such as a virus, or said target may be a soluble antigen, e.g. expressed in a tumor microenvironment of a subject.

[0105] Said CAR, wherein said binder (antigen binding domain) binds (specifically) a tag of a tagged polypeptide, wherein said tagged polypeptide may bind specifically to an antigen expressed on a target cell such as a cancer cell or a cell associated with an autoimmune disease or an infected cell.

[0106] Said tag of the tagged polypeptide may be said target.

[0107] The polypeptide of said tagged polypeptide may be an antibody or an antigen binding fragment thereof.

[0108] In another aspect the present invention provides a chimeric antigen receptor (CAR) comprising a) an antigen binding domain, wherein said antigen binding domain binds to a target, b) a transmembrane domain, and c) an intracellular signaling domain, wherein said antigen binding domain is or comprises an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1, but in which at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target. Said CAR, wherein said at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of said amino acid positions that are altered do not comprise any cysteine.

[0109] In one embodiment of the invention the chimeric antigen receptor (CAR) is a CAR comprising a) an antigen binding domain, wherein said antigen binding domain binds to an antigen, b) a transmembrane domain, and c) an intracellular signaling domain, wherein said antigen binding domain is or comprises an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1, but in which at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target, wherein said at least 5 of said amino acid positions that are altered do not comprise any cysteine, wherein said antigen binding domain of said CAR is obtained by the method of constructing a library of protein scaffolds as disclosed herein, followed by the method of isolating an antigen binding domain, that binds to an antigen, as disclosed herein, (and wherein the isolated antigen binding domain is the antigen binding domain of the CAR).

[0110] The CARs as disclosed herein may be expressed in immune cells such as T cells, NK cells, or gamma delta T cells.

[0111] The immune cells expressing a CAR as disclosed herein may be used for immunotherapy.

[0112] In another aspect the present invention provides a combination of compositions for use in immunotherapy or a kit comprising

[0113] I) an immune cell expressing a CAR comprising a) an antigen binding domain, wherein said antigen binding domain binds to a tag of a tagged polypeptide, and wherein said antigen binding domain is or comprises an amino acid sequence that has at least 75%, at least 80%, at least 85%, or at least 90% identity to Seq ID NO: 1, but in which at least 5, at least 6, at least 7, at least 8, at least 9 or 10 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind said tag, b) a transmembrane domain, and c) an intracellular signaling domain, and

[0114] II) said tagged polypeptide, wherein said polypeptide of said tagged polypeptide binds an antigen expressed on a target cell such as a cancer cell or a cell associated with an autoimmune disease or an infected cell.

[0115] In one embodiment of the invention the immune cells expressing the CAR having a binder (antigen binding domain) based on the N-terminal domain of human FACT complex subunit SSRP1 as disclosed herein as an antigen binding domain are for use in treatment of a disease associated with a target cell of a subject suffering from said disease, the disease may be e.g. cancer and the target cell a cancerous cell. Immune cells, e.g. T cells or NK cells of a subject may be isolated by methods known in the art (e.g. enrichment of CD4+ and / or CD8+ T cells). The subject may e.g. suffer from said cancer or may be a healthy subject. These cells are genetically modified in vitro to express the CAR as disclosed herein. These engineered cells may be activated and expanded in vitro or in-vivo. In a cellular therapy these engineered cells are infused to a recipient in need thereof. These cells may be a pharmaceutical composition (said cell plus pharmaceutical acceptable carrier). The infused cells may be e.g. able to kill (or at least stop growth of) cancerous cells in the recipient. The recipient may be the same subject from which the cells were obtained (autologous cell therapy) or may be from another subject of the same species (allogeneic cell therapy).

[0116] The immune cells, preferentially T cells or NK cells engineered to express the CAR as disclosed herein may be administered either alone, or as a pharmaceutical composition in combination with diluents and / or with other components such as IL-2 or other cytokines or cell populations. Briefly, pharmaceutical compositions of the present invention may comprise a cell population of genetically modified cells (a plurality of immune cells) as described herein, in combination with one or more pharmaceutically or physiologically acceptable carriers, diluents or excipients. Such compositions may comprise buffers such as neutral buffered saline, phosphate buffered saline and the like; carbohydrates such as glucose, mannose, sucrose or dextrans, mannitol; proteins; polypeptides or amino acids such as glycine; antioxidants; chelating agents such as EDTA or glutathione; adjuvants (e.g., aluminum hydroxide); and preservatives. Preferentially, the compositions of the present invention are formulated for intravenous administration. The administration of cell compositions to the subject may be carried out in any convenient manner known in the art.

[0117] Pharmaceutical compositions of the present invention may be administered in a manner appropriate to the disease to be treated. Appropriate dosages may be determined by clinical trials. But the quantity and frequency of administration will also be determined and influenced by such factors as the condition of the patient, and the type and severity of the patient's disease.

[0118] A pharmaceutical composition comprising the immune cells, preferentially T cells or NK cells as disclosed herein may be administered at a dosage of 104to 109cells / kg body weight, preferably 105to 106cells / kg body weight. The cell compositions may also be administered several times at these dosages. The compositions of cells may be injected e.g. directly into a tumor, lymph node, or site of infection.

[0119] The genetically engineered immune cells may be activated and expanded to therapeutic effective amounts using methods known in the art.

[0120] The immune cells of the invention may be used in combination with e.g. chemotherapy, radiation, immunosuppressive agents, antibodies or antibody therapies.

[0121] All definitions, characteristics and embodiments defined herein with regard to the first aspect of the invention as disclosed herein also apply mutatis mutandis in the context of the other aspects of the invention as disclosed herein.

[0122] Definitions

[0123] Unless defined otherwise, technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0124] As used herein the term “comprising” or “comprises” is used in reference to compositions, methods, and respective component s) thereof, that are essential to the method or composition, yet open to the inclusion of unspecified elements, whether essential or not.

[0125] The terms “nucleic acid (molecule)”, “nucleic acid sequence” or “polynucleotide” as used interchangeably herein refer to polymers of nucleotides. Polynucleotides, which can be hydrolyzed into monomeric “nucleotides.” The monomeric nucleotides can be hydrolyzed into nucleosides. As used herein, the term “polynucleotides” encompasses, but is not limited to, all nucleic acid sequences which are obtained by any means available in the art, including, without limitation, recombinant means, i.e., the cloning of nucleic acid sequences from a recombinant library or a cell genome, using ordinary cloning technology and PCR, and the like, and by synthetic means.

[0126] The term “nucleic acid” is the overall name for deoxyribonucleic acid (DNA) and ribonucleic acid (RNA). Nucleic acids (nucleic acid sequences) are linear polymers of nucleotides. Each nucleotide consists of three components: a purine or pyrimidine nucleobase (base), a pentose sugar, and a phosphate group. The substructure consisting of a nucleobase plus sugar is termed a nucleoside. Nucleic acid types differ in the structure of the sugar in their nucleotides - DNA contains 2'-deoxyribose while RNA contains ribose. Also, the nucleobases found in the two nucleic acid types are different: adenine, cytosine, and guanine are found in both RNA and DNA, while thymine occurs in DNA and uracil occurs in RNA

[0127] Degeneracy or redundancy of codons (codon degeneracy) is the redundancy of the genetic code, exhibited as the multiplicity of three-base pair codon combinations that specify an amino acid. The codons encoding one amino acid may differ in any of their three positions; however, more often than not, this difference is in the second or third position. For instance, the amino acid glutamic acid is specified by GAA and GAG codons (difference in the third position); the amino acid leucine is specified by UUA, UUG, CUU, CUC, CUA, CUG codons (difference in the first or third position); and the amino acid serine is specified by UCA, UCG, UCC, UCU, AGU, AGC (difference in the first, second, or third position).

[0128] Degeneracy results because there are more codons than encodable amino acids. For example, if there were two bases per codon, then only 16 amino acids could be coded for (42=16). Because at least 21 codes are required (20 amino acids plus stop) and the next largest number of bases is three, then 43gives 64 possible codons, meaning that some degeneracy must exist.

[0129] Since proline is encoded by 4 codons, and since there are 64 codons in total, the codon degeneracy would yield 6.25% (4 / 64) prolines.

[0130] Since tyrosine is encoded by 2 codons, and since there are 64 codons in total, the codon degeneracy would yield 3.125% (2 / 64) tyrosines.

[0131] The term “NNN” defines a codon, wherein at all three nucleotide positions all four bases (A, C, G and T) are allowed. In other words, “N” defines a nucleotide position that is randomly mutated and contains a mix of all four DNA bases.

[0132] The term “NNK” defines a codon, wherein the first two positions defined by “N” contain a mix of all four bases (A, C, G and T) and the third position “K” defines a mix of G and T.

[0133] The term “NNS” defines a codon, wherein the first two positions defined by “N” contain a mix of all four bases (A, C, G and T) and the third position “S” defines a mix of C and G. The term “NNB” defines a codon, wherein the first two positions defined by “N” contain a mix of all four bases (A, C, G and T) and the third position “B” defines a mix of C, G and T.

[0134] The terms “library of binder scaffolds”, “binder scaffold library” and “library of protein scaffolds” as used herein may be used interchangeably. The term “binder scaffold” as used herein refers to an initial starting protein (a protein scaffold) that can be engineered for antigen binding. The terms “binder” or “engineered binder scaffold” as used herein refer to a protein or polypeptide that is able to bind to a target such as an antigen. The binder may be an antigen binding domain, preferentially an engineered binder scaffold identified by the methods as disclosed herein.

[0135] As used herein, the terms “peptide,” “polypeptide,” and “protein” are used interchangeably, and refer to a compound comprised of amino acids covalently linked by peptide bonds. A protein or peptide must contain at least two amino acids, and no limitation is placed on the maximum number of amino acids that can comprise a protein’s or peptide’s sequence. Polypeptides include any peptide or protein comprising two or more amino acids joined to each other by peptide bonds. As used herein, the term refers to both short chains, which also commonly are referred to in the art as peptides, oligopeptides and oligomers, for example, and to longer chains, which generally are referred to in the art as proteins, of which there are many types. “Polypeptides” include, for example, biologically active fragments, substantially homologous polypeptides, oligopeptides, homodimers, heterodimers, variants of polypeptides, modified polypeptides, derivatives, analogs, fusion proteins, among others. The polypeptides include natural peptides, recombinant peptides, synthetic peptides, or a combination thereof.

[0136] The amino acid residues as used herein normally refer to the 20 naturally occurring amino acids that make up proteins (complete name - three-letter code - one-letter code): alanine - ala - A; arginine - arg - R; asparagine - asn - N; aspartic acid - asp - D; cysteine - cys

[0137] - C; glutamine - gin - Q; glutamic acid - glu - E; glycine - gly - G; histidine - his - H; isoleucine

[0138] - ile - I; leucine - leu - L; lysine - lys - K; methionine - met - M; phenylalanine - phe - F; proline - pro - P; serine - ser - S; threonine - thr - T: tryptophan - trp - W; tyrosine - tyr - Y; valine - val - V.

[0139] The term “does not comprise the introduction of any cysteine” as used herein refers to the intentional avoidance of the introduction of cysteines into the polypeptide as disclosed herein at the best technical possibilities using the methods as disclosed herein. In practice, a 100% avoidance of introduction of cysteines into the polypeptide as disclosed herein is not always achievable. A maximal percentage of 0.2% cysteines of a given randomly mutated amino acid position might be incorporated into the polypeptide as disclosed herein by the methods used to generate the polypeptide (copies of the initial polypeptide) as disclosed herein. This would mean that 0.2% of a given randomly mutated amino acid position would contain a cysteine. So if the copies of said initial polypeptide contain 10 randomly mutated amino acid positions (each of them containing 0.2%), this would yield 2 polypeptide variants in 100 which contain a cysteine. Therefore such a maximal percentage of 0.2% cysteines of a given randomly mutated amino acid position incorporated into the copy of the initial polypeptide as disclosed herein by the methods as disclosed herein may be tolerated. Generated copies of the initial polypeptide containing a cysteine may be sorted out before the generated copies may be further processed for further applications such as integration of a copy of the initial polypeptide as a binder (antigen binding domain) into a CAR construct.

[0140] Structure-specific recognition protein-1 (SSRP1) is part of the FACT complex (facilitates chromatin transcription), which is involved in processes such as DNA repair, replication and transcription (Marciano et al., J Biol Chem. 2018 Jun 29;293(26): 10071-10083). The FACT complex itself consists of two subunits: SSRP1 and the suppressor protein of Ty 16 (SPT16), which heterodimerize to form the FACT complex. Human SSRP1 consists of several domains including N-terminal PHI and PH2 domains, a middle domain (consisting of PH3 and PH4) and a high-mobility group (HMG) domain flanked by two intrinsic disordered regions (ID1 and ID2) (Marciano et al., J Biol Chem. 2018 Jun 29;293(26): 10071-10083).

[0141] CD22 is a molecule belonging to the SIGLEC family of lectins. It is found on the surface of mature B cells and to a lesser extent on some immature B cells. Generally speaking, CD22 is a regulatory molecule that prevents the overactivation of the immune system and the development of autoimmune diseases. CD22 is a transmembrane protein with a molecular weight of 140 kDa. The extracellular part of CD22 consists of seven immunoglobulin domains and the intracellular part is formed by 141-amino acid cytoplasmic tail.

[0142] CD276 (B7-H3) is a 316 amino acid-long type I transmembrane protein, existing in two isoforms determined by its extracellular domain. In non-malignant tissues, CD276 has a predominantly inhibitory role in adaptive immunity, suppressing T cell activation and proliferation.

[0143] The term "antibody" as used herein is used in the broadest sense to cover the various forms of antibody structures including but not being limited to monoclonal and polyclonal antibodies (including full length antibodies), multispecific antibodies (e.g. bispecific antibodies), antibody fragments, i.e. antigen binding fragments of an antibody, immunoadhesins and antibody - immunoadhesin chimeras, that specifically recognize (i.e. bind) an antigen. "Antigen binding fragments of an antibody" comprise a portion of a full-length antibody, preferably the variable domain thereof, or at least the antigen binding site thereof (“an antigen binding fragment of an antibody”). Examples of antigen binding fragments include Fab (fragment antigen binding), scFv (single chain fragment variable), single domain antibodies (nanobodies or VHHs, or single VH domains), diabodies, dsFv, Fab’, diabodies, single-chain antibody molecules, and multispecific antibodies formed from antibody fragments.

[0144] The terms “having specificity for”, “specifically binds” or “specific for” with respect to an antigen-binding domain of an antibody, of a fragment thereof or of a CAR, or with respect to a binder (antigen binding domain) generated with the methods disclosed herein, refer to an antigen-binding domain or binder which recognizes and binds to a specific antigen, but does not substantially recognize or bind other molecules in a sample. An antigen-binding domain or binder that binds specifically to an antigen from one species may bind also to that antigen from another species. This cross-species reactivity is not contrary to the definition of that antigenbinding domain or binder is specific. An antigen-binding domain or binder that specifically binds to an antigen may bind also to different allelic forms of the antigen (allelic variants, splice variants, isoforms etc.). This cross reactivity is not contrary to the definition of that antigenbinding domain or binder is specific.

[0145] As used herein, the term “antigen” is intended to include substances that bind to or evoke the production of one or more antibodies and may comprise, but is not limited to, proteins, peptides, polypeptides, oligopeptides, lipids, carbohydrates such as dextran, haptens and combinations thereof, for example a glycosylated protein or a glycolipid. The term “antigen” as used herein refers to a molecular entity that may be expressed e.g. on the surface of a target cell and that can be recognized by means of the adaptive immune system including but not restricted to antibodies or TCRs, or engineered molecules including but not restricted to endogenous or transgenic TCRs, CARs, the binders (antigen binding domains) based on the N-terminal domain of human FACT complex subunit SSRP1 as disclosed herein, or any other molecule that can execute binding to a structure with high affinity.

[0146] The terms “immune cell” or “immune effector cell” may be used interchangeably and refer to a cell that may be part of the immune system and executes a particular effector function such as T cells, alpha-beta T cells, NK cells, NKT cells, B cells, innate lymphoid cells (ILC), cytokine induced killer (CIK) cells, lymphokine activated killer (LAK) cells, gamma-delta T cells, regulatory T cells (Treg), monocytes or macrophages. Preferentially these immune cells are human immune cells. Preferred immune cells are cells with cytotoxic effector function such as alpha-beta T cells, NK cells, NKT cells, ILC, CIK cells, LAK cells or gamma-delta T cells. Most preferred immune effector cells are T cells and NK cells. Tumor infiltrating lymphocytes (TILs) are T cells that have moved from the blood of a subject into a tumor. These TILs may be removed from a patient's tumor by methods well known in the art, e.g. enzymatic and mechanic tumor disruption followed by density centrifugation and / or cell marker specific enrichment. TILs may be genetically engineered as disclosed herein, and then given back to the patient. "Effector function" means a specialized function of a cell, e.g. in a T cell an effector function may be cytolytic activity or helper activity including the secretion of cytokines.

[0147] T cells or T lymphocytes are a type of lymphocyte that play a central role in cell-mediated immunity. They can be distinguished from other lymphocytes, such as B cells and natural killer cells (NK cells), by the presence of a T cell receptor (TCR) on the cell surface. There are several subsets of T cells, each with a distinct function.

[0148] Immunotherapy is a medical term defined as the "treatment of disease by inducing, enhancing, or suppressing an immune response". Immunotherapies designed to elicit or amplify an immune response are classified as activation immunotherapies, while immunotherapies that reduce or suppress are classified as suppression immunotherapies. Cancer immunotherapy as an activating immunotherapy attempts to stimulate the immune system to reject and destroy tumors. Adoptive cell transfer uses cell-based, preferentially T cell-based or NK cell-based cytotoxic responses to attack cancer cells. T cells that have a natural or genetically engineered reactivity to a patient's cancer are generated in-vitro and then transferred back into the cancer patient. If the immune cells have been genetically modified to express a CAR, the immunotherapy is referred to as “CAR cell immunotherapy” or in case of use of T cells only as “CAR T cell therapy” or “CAR T cell immunotherapy”.

[0149] The term “treatment” as used herein means to reduce the frequency or severity of at least one sign or symptom of a disease.

[0150] The terms “therapeutically effective amount” or “therapeutically effective population” mean an amount of a cell population which provides a therapeutic benefit in a subject.

[0151] As used herein, the term “subject” refers to an animal. Preferentially, the subject is a mammal such as mouse, rat, cow, pig, goat, chicken dog, monkey or human. More preferentially, the subject is a human. The subject may be a subject suffering from a disease such as cancer (a patient) or from an autoimmune disease or from an allergic disease or from an infectious disease or from graft rejection.

[0152] The term "expression" as used herein is defined as the transcription and / or translation of a particular nucleotide sequence driven by its promoter in a cell. In general, a CAR as used herein may comprise an extracellular domain (extracellular part) comprising the antigen binding domain, a transmembrane domain and a cytoplasmic signaling domain (intracellular signaling domain). The extracellular domain may be linked to the transmembrane domain by a linker or spacer. The extracellular domain may also comprise a signal peptide. In some embodiments of the invention the antigen binding domain of a CAR binds a tag or hapten that is coupled to a polypeptide (“haptenylated” or “tagged” polypeptide), wherein the polypeptide may bind to a disease-associated antigen such as a tumor associated antigen (TAA) that may be expressed on the surface of a cancer cell. Such a CAR may be referred to as “anti-tag” CAR or “adapterCAR” or “universal CAR” as disclosed e.g. in US9233125B2.

[0153] The haptens or tags may be coupled directly or indirectly to a polypeptide (the tagged polypeptide), wherein the polypeptide may bind to said disease associated antigen expressed on the (cell) surface of a target. The tag may be e.g. dextran or a hapten such as biotin or fluorescein isothiocyanate (FITC) or phycoerythrin (PE) or thiamin, but the tag may also be a peptide sequence e.g. chemically or recombinantly coupled to the polypeptide part of the tagged polypeptide. The tag may also be streptavidin. The tag portion of the tagged polypeptide is only constrained by being a molecular component that can be recognized and specifically bound by the antigen binding domain specific for the tag of the CAR. For example, when the tag is FITC (fluorescein isothiocyanate), the tag-binding domain may constitute an anti-FITC scFv. Alternatively, when the tag is biotin or PE (phycoerythrin), the tag-binding domain may constitute an anti-biotin scFv or an anti-PE scFv, respectively.

[0154] A "signal peptide" refers to a peptide sequence that directs the transport and localization of the protein within a cell, e.g. to a certain cell organelle (such as the endoplasmic reticulum) and / or to the cell surface.

[0155] Generally, an “antigen binding domain” in the context of a CAR refers to the region of the CAR that specifically binds to an antigen, e.g. to a tumor associated antigen (TAA) or tumor specific antigen (TSA). The CARs of the invention may comprise one or more antigen binding domains (e.g. a tandem CAR). Generally, the targeting regions on the CAR are extracellular. Generally, the antigen binding domain may comprise an antibody or an antigen binding fragment thereof. The antigen binding domain may comprise, for example, full length heavy chain, Fab fragments, single chain Fv (scFv) fragments, divalent single chain antibodies, diabodies or single domain antibodies (VHHs or nanobodies). The CAR as disclosed herein may comprise a binder as antigen binding domain based on the N-terminal domain of human FACT complex subunit SSRP1 as disclosed herein.

[0156] Often the antigen binding domain is a scFv or single domain antibody. Normally, in a scFv, the variable regions of an immunoglobulin heavy chain and light chain are fused by a flexible linker to form a scFv. Such a linker may be for example the “(G4 / S)3-linker”.

[0157] “Spacer” or “hinge” as used herein refers to the hydrophilic region which is between the antigen binding domain and the transmembrane domain. The CARs of the invention may comprise an extracellular spacer domain but it is also possible to leave out such a spacer. The spacer may include e.g. Fc fragments of antibodies or fragments thereof, hinge regions of antibodies or fragments thereof, CH2 or CH3 regions of antibodies, accessory proteins, artificial spacer sequences or combinations thereof. A prominent example of a spacer is the CD8alpha hinge.

[0158] The transmembrane domain of the CAR may be derived from any desired natural or synthetic source for such domain. When the source is natural the domain may be derived from any membrane-bound or transmembrane protein. The transmembrane domain may be derived for example from CD8alpha or CD28. When the key signaling and antigen recognition modules (domains) are on two (or even more) polypeptides then the CAR may have two (or more) transmembrane domains. Separating the key signaling and antigen recognition modules on two different polypeptide chains enables small molecule-dependent, titratable and reversible control over CAR signaling (e.g. WO2014127261A1) due to small molecule-dependent heterodimerizing domains in each polypeptide of the CAR.

[0159] The cytoplasmic signaling domain (the intracellular signaling domain or the activating endodomain) of the CAR is responsible for activation of at least one of the normal effector functions of the immune cell in which the CAR is expressed, if the respective CAR is an activating CAR (normally, a CAR as described herein refers to an activating CAR, otherwise it is indicated explicitly as an inhibitory CAR (iCAR)). "Effector function" means a specialized function of a cell, e.g. in a T cell an effector function may be cytolytic activity or helper activity including the secretion of cytokines. The intracellular signaling domain refers to the part of a protein which transduces the effector function signal and directs the cell expressing the CAR to perform a specialized function. The intracellular signaling domain may include any complete, mutated or truncated part of the intracellular signaling domain of a given protein sufficient to transduce a signal which initiates or blocks immune cell effector functions. Prominent examples of intracellular signaling domains for use in the CARs include the cytoplasmic signaling sequences of the T cell receptor (TCR) complex, co-receptors and costimulatory receptors that initiate signal transduction following receptor engagement.

[0160] Generally, T cell activation can be mediated by two distinct classes of cytoplasmic signaling sequences, firstly those that initiate antigen-dependent primary activation through the TCR complex (primary cytoplasmic signaling sequences, primary cytoplasmic signaling domain) and secondly those that usually act in an antigen-independent manner to provide a secondary or co- stimulatory signal (secondary cytoplasmic signaling sequences, co-stimulatory signaling domain). Therefore, an intracellular signaling domain of a CAR may comprise one or more primary cytoplasmic signaling domains and / or one or more secondary cytoplasmic signaling domains.

[0161] Primary cytoplasmic signaling domains that act in a stimulatory manner may contain ITAMs (immunoreceptor tyrosine-based activation motifs).

[0162] Examples of IT AM containing primary cytoplasmic signaling domains often used in CARs are those derived from TCR^ (CD3Q, FcRgamma, FcRbeta, CD3gamma, CD3delta, CD3epsilon, CD5, CD22, CD79a, CD79b, and CD66d. Most prominent is sequence derived from CD3^.

[0163] The cytoplasmic domain of the CAR may be designed to comprise the CD3^ signaling domain by itself or combined with any other desired cytoplasmic domain(s). The cytoplasmic domain of the CAR can comprise a CD3^ chain portion and a co-stimulatory signaling region (domain). The co-stimulatory signaling region refers to a part of the CAR comprising the intracellular domain of a co-stimulatory molecule. A co-stimulatory molecule is a cell surface molecule other than an antigen receptor or their ligands that is required for an efficient response of lymphocytes to an antigen. Examples for a co-stimulatory molecule are CD27, CD28, 4-1BB (CD137), 0X40, CD30, CD40, ICOS, lymphocyte function-associated antigen- 1 (LFA-1), CD2, CD7, LIGHT, NKG2C, B7-H3.

[0164] The cytoplasmic signaling sequences within the cytoplasmic signaling part of the CAR may be linked to each other with or without a linker in a random or specified order. A short oligo- or polypeptide linker, which is preferably between 2 and 10 amino acids in length, may form the linkage. A prominent linker is the glycine-serine doublet.

[0165] As an example, the cytoplasmic domain may comprise the signaling domain of CD3^ and the signaling domain of CD28. In another example, the cytoplasmic domain may comprise the signaling domain of CD3^ and the signaling domain of CD137. In a further example, the cytoplasmic domain may comprise the signaling domain of CD3^, the signaling domain of CD28, and the signaling domain of CD137.

[0166] As aforementioned either the extracellular part or the transmembrane domain or the cytoplasmic domain of a CAR may also comprise a heterodimerizing domain for the aim of separating key signaling and antigen recognition modules of the CAR onto two (or more) separate polypeptide chains.

[0167] The CAR may be further modified to include on the level of the nucleic acid encoding the CAR one or more operative elements to eliminate CAR expressing immune cells by virtue of a suicide switch. The suicide switch can include, for example, an apoptosis inducing signaling cascade or a drug that induces cell death. In one embodiment, the nucleic acid expressing and encoding the CAR can be further modified to express an enzyme such thymidine kinase (TK) or cytosine deaminase (CD). The CAR may also be part of a gene expression system that allows controlled expression of the CAR in the immune cell. Such a gene expression system may be an inducible gene expression system and wherein when an induction agent is administered to a cell being transduced with said inducible gene expression system, the gene expression system is induced and said CAR is expressed on the surface of said transduced cell.

[0168] In some embodiments, the endodomain may contain a primary cytoplasmic signaling domain or a co-stimulatory region, but not both.

[0169] In some embodiments, the CAR may be a “SUPRA” (split, universal, and programmable) CAR, where a “zipCAR” domain may link an intracellular co-stimulatory domain and an extracellular leucine zipper (WO2017 / 091546). This zipper may be targeted with a complementary zipper fused e.g. to an scFv region to render the SUPRA CAR T cell tumor specific. This approach would be particularly useful for generating universal CAR T cells for various tumors; adapter molecules could be designed for tumor specificity and would provide options for altering specificity post-adoptive transfer, key for situations of selection pressure and antigen escape.

[0170] The CARs as described herein may be designed to comprise any portion or part of the above- mentioned domains as described herein in any order and / or combination resulting in a functional CAR, i.e. a CAR that mediated an immune effector response of the immune effector cell that expresses the CAR as disclosed herein.

[0171] The term “tagged polypeptide” as used herein refers to a polypeptide that has bound thereto directly or indirectly at least one additional component, i.e. the tag. The tagged polypeptide as used herein is able to bind an antigen expressed on a target cell. The polypeptide may be an antibody or antigen binding fragment thereof that binds to an antigen expressed on the surface of a target cell such as a tumor associated antigen on a cancer cell. The polypeptide of the tagged polypeptide alternatively may be a cytokine or a growth factor or another soluble polypeptide that is capable of binding to an antigen of a target cell.

[0172] The terms “adapter” or “adapter molecule” or “tagged polypeptide” as used herein may be used interchangeably.

[0173] The tag may be e.g. a hapten or dextran and the hapten or dextran may be bound by the antigen binding domain of the polypeptide, e.g. a CAR, comprising an antigen binding domain specific for the tag.

[0174] Haptens such as e.g. FITC, biotin, or dextran are small molecules that elicit an immune response only when attached to a large carrier such as a protein; the carrier may be one that also does not elicit an immune response by itself. Once the body has generated antibodies to a hapten-carrier adduct, the small-molecule hapten may also be able to bind to the antibody, but it will usually not initiate an immune response; usually only the hapten-carrier adduct can do this.

[0175] However, the tag may also be a peptide sequence e.g. chemically or recombinantly coupled to the polypeptide part of the tagged polypeptide. The peptide may be selected from the group consisting of c-Myc-tag, Strep-Tag, Flag-Tag, and Polyhistidine-tag. The tag may also be streptavidin. The tag portion of the tagged polypeptide is only constrained by being a molecular component that can be recognized and specifically bound by the antigen binding domain specific for the tag of the CAR. For example, when the tag is FITC (Fluorescein isothiocyanate), the tag-binding domain may constitute an anti-FITC scFv. Alternatively, when the tag is biotin or PE (phycoerythrin), the tag-binding domain may constitute an anti -biotin scFv or an anti -PE scFv.

[0176] As used herein, the term “identity” (of proteins and polypeptides) with respect to amino acid sequences is used for a comparison of protein chains. Calculations of "sequence identity" between two sequences may be performed as follows. The sequences are aligned for optimal comparison purposes (e.g., gaps can be introduced in one or both of a first and a second amino acid sequence for optimal alignment and non-homologous sequences may be disregarded for comparison purposes). The optimal alignment is determined as the best score using the “ssearch36” program in the FASTA36 software package (http: / / faculty.virginia.edu / wrpearson / fasta / ) with a Blossum 50 scoring matrix with a gap-open penalty of -10, and a gap-extension penalty of -2. The amino acid residues at corresponding amino acid positions are then compared. When a position in the first sequence is occupied by the same amino acid residue at the corresponding position in the second sequence, then the molecules are identical at that position. The percent identity between the two sequences is a function of the number of identical positions shared by the sequences.

[0177] The terms “engineered cell” and “genetically modified cell” as used herein can be used interchangeably. The terms mean containing and / or expressing a foreign gene or nucleic acid sequence which in turn modifies the genotype and / or phenotype of the cell or its progeny. Especially, the terms refer to the fact that cells, preferentially T cells can be manipulated by recombinant methods well known in the art to express stably or transiently peptides or proteins which are not expressed in these cells in the natural state. For example, T cells, preferentially human T cells are engineered to express an artificial construct such as a chimeric antigen receptor (CAR) on their cell surface.

[0178] The term “cancer” is known medically as a malignant neoplasm. Cancer is a broad group of diseases involving unregulated cell growth and includes all kinds of leukemia, among many others. In cancer, cells (cancerous cells) divide and grow uncontrollably, forming malignant tumors, and invading nearby parts of the body. The cancer may also spread to more distant parts of the body through the lymphatic system or bloodstream. There are over 200 different known cancers that affect humans.

[0179] The cancer to be treated as disclosed herein, may be a solid cancer or may be a lymphoma or a hematological malignancy.

[0180] Said solid cancer (tumor) may be adrenal cancer, anal cancer, bile duct cancer, bladder cancer, bone cancer, brain / CNS tumors in children or adults, breast cancer, cervical cancer, colon / rectum cancer, endometrial cancer, esophagus cancer, ewing family of tumors, eye cancer, gallbladder cancer, gastrointestinal carcinoid tumors, gastrointestinal stromal tumor (GIST), gestation trophoblastic disease, hodgkin disease, kaposi sarcoma, kidney cancer, laryngeal and hypopharyngeal cancer, leukemia, acute lymphocytic leukemia, acute myeloid leukemia, chronic lymphocytic leukemia, chronic myeloid leukemia, chronic myelomonocytic leukemia, liver cancer, lung cancer, non-small cell lung cancer, small cell lung cancer, lung carcinoid tumor, lymphoma, malignant mesothelioma, multiple myeloma, myelodysplastic syndrome, nasal cavity and paranasal sinum cancer, nasopharyngeal cancer, neuroblastoma, non-hodgkin lymphoma, oral cavity or oropharyngeal cancer, osteosarcoma, ovarian cancer, pancreatic cancer, penile cancer, prostate cancer, rhabdomyosarcoma, skin cancer, melanoma, merkel cell skin cancer, small intestine cancer, stomach cancer, testicular cancer, thyroid cancer, uterine sarcoma, vaginal cancer, vulvar cancer, or nephroblastoma. Autoimmune diseases are a condition arising from autoimmunity or disbalance in the immune homeostasis resulting in pathologies that can affect multiple different organ systems. Examples include Behcet’s disease, Juvenile idiopathic arthritis, Type 1 diabetes, Rheumatoid arthritis, Wegener Granulomatosis, Systemic lupus erythematosus, Systemic sclerosis, Crohn's disease, Graves' disease, Hashimoto thyroiditis, Goodpasture syndrome, Primary biliary cholangitis, Myasthenia gravis, Dermato polymyositis, Vasculitis, Mixed connective tissue disease, Scleroderma, Multiple sclerosis, Psoriasis, Ulcerative colitis and Uveitis.

[0181] Infection (infectious disease) is the invasion of an organism's body tissues by disease-causing agents, their multiplication, and the reaction of host tissues to the infectious agents and the toxins they produce. Infections are caused by infectious agents (pathogens) including: viruses, bacteria, fungi and parasites. Said infection may be an acute or a chronic infection.

[0182] A fusion protein may be a protein created through the joining of two or more genes or parts of genes that originally coded for separate proteins. Translation of this fusion gene may result in a single polypeptide with functional properties derived from each of the original proteins or polypeptides.

[0183] Examples

[0184] The following examples are intended for a more detailed explanation of the invention but without restricting the invention to these examples.

[0185] It should be noted that all constructs termed “5UMR” in the examples of this invention do not contain the sequence "GH" at the N-terminal end of the 5UMR-sequence defined in the Protein Data Bank (PDB). This “GH” sequence is not part of the human protein SSRP1 and obviously derived from a linker or cloning site and was therefore omitted in all constructs.

[0186] Example 1 :

[0187] Identification of the the N-terminal domain of human FACT complex subunit SSRP1 as a promising binder scaffold with highly beneficial properties

[0188] To identify potential binder scaffolds for diverse applications, but especially for applications in CAR T cells, the inventors screened the Protein Data Bank (PDB; https: / / www.rcsb.org / ) for protein domains with the following characteristics: (i) a P-sheet content of >40%, since binding surfaces based on flat, rigid P-sheets have been shown previously to yield stable binders with low tendency to aggregate (Traxlmayr et al., J Biol Chem. 2016 Oct 21;291(43):22496-22508; Paumann-Page et al., Redox Biol. 2021 Oct;46: 102090), (ii) a molecular weight between 4400 and 15400 Daltons, (iii) being of human origin to reduce the risk of immunogenicity when being used in human therapeutics such as CAR T cells (Turtle et al., J Clin Invest. 2016 Jun l;126(6):2123-38), (iv) lack of disulfide bonds and lack of free cysteines in order to enable intra- and extracellular applications and (v) lack of N-glycosylation motifs to enable efficient translation from one expression host to another without altering protein characteristics. Based on these characteristics, the inventors screened the PDB. After additional manual screening, 15 potential binder scaffolds were selected:

[0189] • 5UMR (SEQ ID NO: 1)

[0190] • 1KIK (SEQ ID NO:2)

[0191] • 2CQA (SEQ ID NO:3)

[0192] • 2DMH (SEQ ID NO:4)

[0193] • 2E7C (SEQ ID NO: 5)

[0194] • 2K1M (SEQ ID NO:6)

[0195] • 2K45 (SEQ ID NO: 7)

[0196] • 2P9R (SEQ ID NO: 8)

[0197] • 2V37 (SEQ ID NO: 9)

[0198] • 2 VH7 (SEQ ID NO: 10)

[0199] • 3135 (SEQ ID NO: 11)

[0200] • 3 MX7 (SEQ ID NO: 12)

[0201] • 3SHU (SEQ ID NO: 13)

[0202] • 5NOO (SEQ ID NO: 14)

[0203] • 6MYF (SEQ ID NO: 15)

[0204] • FMC63 (SEQ ID NO: 16)

[0205] • FN3 (SEQ ID NO: 17)

[0206] For several binder scaffold candidates, a few amino acids were deleted at the N- and / or C-terminus, resulting in slightly shorter sequences when compared with the PDB entries, such as the deleted “GH" sequence at the N-terminal end of the 5UMR sequence, as was already discussed. The exact sequences that were used in this example (and in the other examples) are defined by SEQ ID NO: 1 to SEQ ID NO: 15. Nevertheless, despite the deletion of some amino acids at the N- and / or C-terminus of some protein domains, their PDB-IDs are used as the nomenclature for these protein domains in this invention. These 15 binder scaffold candidates, as well as the CD19-specific FMC63 scFv (used as the antigen binding domain in clinically approved CAR products) and FN3 (tenth type III domain of human fibronectin) as controls, were fused to CAR molecules and expressed in primary human T cells. As a representative example the nt sequence of 1KIK in the respective CAR backbone (GM-CSF-Ra-SP - GGSGGSGG linker - 1KIK - 2x(G4S) - FLAG tag- CD8a hinge - 4-1BB - CD3z) is defined in SEQ ID NO: 18. Bases 91-261 encode for the 1KIK amino acid sequence (SEQ ID NO:2) and were exchanged to the nucleotide sequences of SEQ ID NO: 1 or any of SEQ ID NO:3 to SEQ ID NO: 17 in the respective CARs. In addition, a CAR based on the CD19-specific FMC63 scFv, which is used as the antigen binding domain in clinically approved CAR products, as well as a CAR based on the tenth type III domain of human fibronectin (FN3), which is a frequently used binder scaffold (Hackel et al., J Mol Biol. 2008 Sep 19;381(5): 1238-52; Hackel et al., J Mol Biol. 2010 Aug 6;401(l):84-96; Chen et al., Methods Enzymol. 2013;523:303-26) were included as control CARs (Figure 1A). The CARs based on the 15 binder scaffold candidates are labeled with the PDB-IDs of the respective binder scaffold candidate.

[0207] DNA encoding the transgene, T7 RNA polymerase promoter site and Kozak sequence was PCR amplified and subsequently used as a template for in vitro transcription using the mMESSAGE mMACHINE® T7 Ultra Kit (Thermo Fisher Scientific), followed by subsequent mRNA purification using the RNeasy Kit (Qiagen).

[0208] Buffy coats from de-identified healthy donors were purchased from the Austrian Red Cross, Vienna, Austria. Primary human T cells were isolated by negative selection using the RosetteSep Human T Cell Enrichment Cocktail (STEMCELL Technologies), followed by gradient density centrifugation using Ficoll-Paque™ PLUS (VWR) and finally cryopreserved in RPMI-1640 GlutaMAX™ medium supplemented with 20 % (v / v) FCS and 10 % DMSO until further usage. Upon thawing T cells were immediately activated using Dynabeads™ Human T-Activator CD3 / CD28 (Thermo Fisher Scientfic) at 25 pL for 3 million T cells. T cells were expanded until sufficient cell numbers were reached for at least 7-10 days until further usage in RPMI-1640 GlutaMAX™ supplemented with 10 % FCS (v / v), 100 U / mL penicillin, 100 pg / mL streptomycin and 200 U / mL of IL-2 (PreproTech).

[0209] 2 million primary human T cells from 2 different donors were electroporated with the purified mRNA samples using the Gene Pulser Xcell Electroporation system (Bio-Rad) and 4 mm electroporation cuvettes with 5 pg of scaffold-CAR mRNA and 1 pg of mRNA encoding for GFP. Beforehand cells were washed with RPMI-1640 (without phenol red) and Opti-MEM (300 - 500 g, 5 min, RT) and set to a final concentration of 2xl06cells / 100 pL Opti-MEM. Electroporation was performed using the square wave protocol (single pulse, 500 V, 4 mm, 5 ms pulse length) and cells were rescued directly after electroporation in 2 mL of pre-warmed full-growth medium (RPMI-1640 GlutaMAX™, 10 % FCS, 100 U / mL penicillin, 100 pg / mL streptomycin and 200 U / mL IL-2). After 16-18 hours of incubation at 37 °C (5 % CO2, 97 % humidity) 100,000 cells were used for subsequent FACS analysis. After a washing step (515 g, 5 min, 4 °C) with 1 mL ice-cold FACS buffer (lx PBS, 0.2 % human albumin, 0.02 % sodium azide) cells were blocked with 50 pL FACS buffer containing 10 % human serum for 10 min, 4 °C. 1.2 pg / mL anti-FLAG-PE (BioLegend, clone L5) antibody were added to label the CAR molecules containing a FLAG-tag, incubated for 25 minutes (in the dark, 4 °C) and subsequently washed twice with 1 mL ice-cold FACS buffer each. Samples were analyzed with a LSR Fortessa instrument.

[0210] Figure 1A shows the expression levels of CAR molecules which all contain identical CAR backbone sequences (i.e. identical signal peptide, as well as hinge, transmembrane and intracellular domains), but which contain different binder scaffold candidates in their extracellular domains or the FMC63 scFv or FN3 as controls. The binder scaffold candidates were termed according to the PDB-ID they are derived from. As shown in Figure 1 A, most of these scaffold-CARs were expressed at higher levels compared with the control CAR based on the FMC63 scFv. Only 2V37-based and 2E7C-based CARs were expressed at levels that were lower or comparable to that of the FMC63-scFv-based CAR.

[0211] Next, all binder scaffold candidates, as well as FN3 as a control, were expressed as soluble proteins in E. coli. All proteins were expressed in pET-21a(+) vectors with an N- terminal linker and a C-terminal (2xG4S)-G linker followed by a hexahistidine (6xHis) tag. As a representative example the amino acid sequence of 1KIK with the respective linkers and tags is shown in SEQ ID NO: 19. Amino acids 11-67 show protein 1KIK followed by a 2xG4S linker and a 6xHis tag and were exchanged to SEQ ID NO: 1 or any of SEQ ID NO: 3 to SEQ ID NO:15 or SEQ ID NO: 17 in the respective samples. Plasmids were transformed into E. coli Tuner cells via heat-shock transformation and grown in lysogeny broth (LB) with 100 pg / mL ampicillin. At stationary phase the culture was diluted to an ODeoo of 0.1 - 0.2 in terrific broth (TB). At ODeoo ~ 0.8-1 the culture was induced with 1 mM isopropyl-beta-D- thiogalactopyranoside (IPTG) and the culture was shaken at 20 °C for roughly 16-20 hours. Cells were harvested (5,000 g, 20 min, 4 °C) and the supernatant was discarded. The pellets were resuspended thoroughly in 30 mL sonication buffer (50 mM sodium phosphate, 300 mM NaCl, 3 % glycerol, 1 % Triton-X 100, pH 8), then sonicated (2x 3 min, pulse 1.0 - 1.0 s, 97 % amplitude) on ice. Afterwards the cell lysate was centrifuged (20,000 g, 30 min, 4 °C) in order to separate the soluble proteins from the rigid cell matter.

[0212] All binder scaffold candidates and FN3 were purified via metal affinity chromatography using TALON metal affinity resin (Takara Bio). Supernatants from crude cell lysates with 10 mM imidazole were applied to the washed and equilibrated TALON matrix twice. The protein holding matrix was washed several times with equilibration buffer (50 mM sodium phosphate, 300 mMNaCl, pH 8) containing increasing amounts of imidazole (5 mM, 15 mM). The proteins were eluted with equilibration buffer containing 250 mM imidazole.

[0213] Buffer exchange of the protein solution to PBS were either performed using Amicon tubes or an overnight dialysis using SnakeSkin Dialysis Tubing at 4 °C. Protein concentration was determined using A280 and the protein was aliquoted to small aliquots needed per experiment and frozen at -80 °C. The resin was regenerated with equilibration buffer containing 250 mM imidazole, regeneration buffer (300 mM NaCl, 20 mM MES, pH 5) and H2O and stored in 20 % ethanol at 4 °C.

[0214] To analyze their thermostability, all binder scaffold candidates and the control protein FN3 were analyzed by differential scanning calorimetry (DSC) using a MicroCai PEAQ-DSC instrument (Malvern Panalytical). All proteins were measured at 50 pM in PBS from 20 - 120 °C at a scan rate of 1 °C / min. Data analysis was performed with the MicroCai PEAQ-DSC Software by Malvern Panalytical performing a buffer baseline subtraction followed by normalization for protein concentration and fitting to a non-two state unfolding model.

[0215] As shown in Figure IB, three binder scaffold candidates showed a midpoint of thermal denaturation (Zm) of >70 °C: 2P9R, 3 SHU and 5UMR (i.e. the N-terminal domain of human SSRP1). In addition to the proteins 2P9R, 3SHU and 5UMR, also the FN3 control showed a Tmabove >70 °C. Since the engineering process (required to generate an antigen binding site) usually results in a reduction of stability (Teufl et al., ACS Synth Biol. 2022 Mar 18; 11(3): 1030- 1039), a high thermostability of the initial (parental) binder scaffold was a crucial design criterion. Therefore, we focused our attention on the proteins 2P9R, 3 SHU and 5UMR.

[0216] Next, the aggregation tendencies of these proteins were analyzed by size exclusion chromatography (SEC). Soluble proteins were analyzed using HPLC (Shimadzu prominence LC20, Komeuburg, Austria), equipped with MALS (WYATT Heleos Dawn8 + plus QELS, soft- ware astra 6, Dembach, Germany), refractive index detector (RID- 10 A, Shimadzu), and a diode array detector (SPD-M20A, Shimadzu). The particle size of the Superdex 75 10 / 300 column was 13 pm and the run was conducted with PBS additionally containing 200 mM NaCl at a flow rate of 0.75 mL / min. 50 pg of the proteins were loaded onto the column. UV absorption data at A280 was analyzed and is represented in Figure 2, demonstrating that no aggregation (which would be indicated by a peak eluting earlier than the main peak) could be detected. The same was observed for the control protein FN3 (Figure 2). However, for 2P9R, peak tailing could be observed. That is, the peak was not symmetric, but it showed an asymmetric appearance with tailing on the right side. This indicates non-specific interactions with the column matrix and / or a monomer / dimer equilibrium, both of which are undesired for potential binder scaffolds. Therefore, 2P9R was excluded as a binder scaffold candidate.

[0217] Based on these experiments shown in Figures 1 and 2, the protein domain based on the PDB-ID 5UMR (without "GH" at the N-terminal end of the 5UMR sequence, which is not part of SSRP1 and obviously derived from a linker or cloning site), which represents the N-terminal domain of human SSRP1 and is defined by SEQ ID NO: 1 in the present invention, was identified as a promising binder scaffold with highly beneficial properties:

[0218] (i) it is a part of a human protein, thus reducing the risk of immunogenicity;

[0219] (ii) the N-terminal domain of human SSRP1 showed very high expression levels on primary human T cells when fused to a CAR backbone (Figure 1 A); surprisingly, this binder scaffold- CAR (labeled with “5UMR” in Figure 1 A) was expressed at much higher levels compared with a CAR based on the clinically used FMC63 scFv, but which was otherwise (apart from the presence of the FMC63 scFv instead of SEQ ID NO: 1) identical (Figure 1 A);

[0220] (iii) the N-terminal domain of human SSRP1 is a single domain protein, thus preventing domain mispairing as is observed with scFvs;

[0221] (iv) the N-terminal domain of human SSRP1 shows favorable biochemical properties, such as a high midpoint of thermal denaturation (Zm) of 71.1 °C (Figure IB, labeled with “5UMR”); since this binder scaffold will be further engineered for antigen recognition, the thermostability will most probably be reduced (Teufl et al., ACS Synth Biol. 2022 Mar 18; 11(3): 1030-1039), thus high initial thermostability was a crucial requirement here;

[0222] (v) the N-terminal domain of human SSRP1 is a monomeric protein without any detectable aggregation (Figure 2, labeled with “5UMR”);

[0223] (vi) the N-terminal domain of human SSRP1 neither contains disulfide bonds, nor free cysteines, thus enabling intra- and extracellular applications;

[0224] (vi) the N-terminal domain of human SSRP1 lacks N-glycosylation sites, thus enabling efficient translation from one expression host to another without altering protein characteristics;

[0225] (vii) the N-terminal domain of human SSRP1 is derived from an intracellular protein; this is another crucial advantage, because engineered proteins which are derived from human proteins and which are used therapeutically in humans sometimes induce antibody-mediated immune responses, which sometimes cross-react with the non-mutated, endogenous protein (Saxton et al., Nat Rev Drug Discov. 2023 Jan;22(l):21-37); however, the N-terminal domain of human SSRP1 is expressed intracellularly as an endogenous protein and therefore potential cross- reactive antibodies cannot bind to the endogenous, non-mutated protein, because it is located intracellularly and therefore not accessible to antibodies; importantly, this is also a critical advantage of the N-terminal domain of human SSRP1 compared with the FN3 domain, which is frequently used as a binder scaffold, but expressed extracellularly as an endogenous protein.

[0226] Construction of a library of binder scaffolds based on SEQ ID NO: 1 with NNK-randomized

[0227] To test whether the N-terminal domain of human SSRP1 can be used for the construction of a library for binder engineering, the investigators constructed randomly mutated libraries based on the N-terminal domain of human SSRP1 (i.e. based on SEQ ID NO: 1).

[0228] First, four different randomly mutated libraries based on the N-terminal domain of human SSRP1 were constructed and displayed on the surface of yeast cells.

[0229] In order to test the effect of the different mutated positions, several NNK randomized libraries were established. In a first PCR, random mutations were inserted. The NNK-codon-containing primers 5UMR_c_NNKl (SEQ ID NO:20), 5UMR_cp_NNKl (SEQ ID NO:21), 5UMR_cy_NNKl (SEQ ID NO:22) and 5UMR_PCRl_rev (SEQ ID NO:23) were used for the amplification with Q5® HiFi DNA Polymerase.

[0230] 5UMR_wt was amplified using primers 5UMR_PCR2_fwd (SEQ ID NO:24) and 5UMR_PCR2_rev (SEQ ID NO:25).

[0231] Gene fragments of the correct size were cut from a preparative DNA gel and gel purified. All libraries were further amplified in a large reaction volume (200 pL); libraries 5UMR_c, 5UMR_cp and 5UMR_cy were amplified using primers 5UMR_PCR2_fwd (SEQ ID NO: 24) and 5UMR_PCR2_rev (SEQ ID NO:25). For the amplification of library 5UMR_cb PCR product “5UMR_c” was used as a template and primers 5UMR_cb_NNK2 (SEQ ID NO:26) and 5UMR_PCR2_rev (SEQ ID NO:25) were used. The final library constructs encoded for Aga2p-HA-tag-(Gly4Ser)3 linker-NNK randomized SEQ ID NO: 1 gene-c-myc tag. All primers were ordered at Sigma-Aldrich. Amplified PCR2 products (200 pL) were used for ethanol purification and electroporation of Saccharomyces cerevisiae strain EBY100 together with BamHI / Nhel digested pCTCON2V vector. Diversities of the NNK randomized yeast libraries averaged around ~107individual clones after electroporation.

[0232] Amplified genes for 5UMR_wt (i.e. SEQ-ID NO: 1) were assembled into a BamHI / Nhel digested pCTCON2V (SEQ ID NO:27) vector using HiFi DNA Assembly (New England Biolabs GmbH) and subsequently amplified in E. coli XL- 10 and sequence verified. The verified pCTCON2V plasmid was either electroporated into Saccharomyces cerevisiae strain EBY100 or transformed into EBY100 using Frozen-EZ Yeast Transformation II kit (Zymo Research).

[0233] For electroporation of yeast libraries, 20 mL of freshly prepared YPD medium was inoculated with a single colony of S. cerevisiae strain EBY100 and shaken overnight at 30 °C. The following day, the culture was diluted to an ODeoo of 0.2. 50 mL of culture volume was prepared for 2 electroporations. The culture was shaken at 30 °C for 4-6 hours until an ODeoo of 1.3-1.5 was reached. The yeast cells were centrifuged (2000 g, 3 min) and the supernatant was discarded. The cell pellet was resuspended in 25 mL of 100 mM lithium acetate and freshly prepared dithiothreitol was added to reach a final concentration of 10 mM. The cultures were incubated shaking at 30 °C (in 50 mL tubes) for 10 minutes and subsequently pelleted to discard the supernatant. All following steps were performed on ice with chilled reagents and cuvettes. The cell pellet was resuspended in 25 mL MQ-H2O, pelleted and the supernatant was removed. Next, the pellet was resuspended in 250 pL of MQ-H2O. 250 pL of cells were mixed with 4 pg of BamHI / Nhel digested pCTCON2V vector and the ethanol precipitate of 200 pL PCR product dissolved in 4-6 pL of MQ-H2O or cell culture grade H2O, and transferred to pre-chilled 2 mm electroporation cuvettes. The electroporation was conducted with the Gene Pulser Xcell Electroporation system (Bio-Rad) using a square wave protocol (single pulse, 500 V, 15 ms pulse length). Immediately after the electroporation, cells were rescued with 1 mL pre-warmed (30 °C) YPD medium and subsequently incubated at 30 °C without shaking in 15 mL tubes. Different dilutions of the cell suspension were plated onto SD-CAA plates (in order to assess the diversity) and the cell suspension was inoculated in SD-CAA and incubated at 30 °C.

[0234] To analyze the quality of the randomly mutated yeast libraries, yeast library cultures were diluted to an ODeoo of 0.2 in SD-CAA and grown at 30 °C for approximately 4-6 hours. At ODeoo 0.8- 1.2, cultures were centrifuged (2000 g, 3 min) and the supernatant was discarded. The cell pellet was taken up and induced in SG-CAA and shaken overnight at 20 °C or at 37 °C. 1 or 2 million yeast cells were used per measurement. The required total amount of cells in SG- CAA was harvested, centrifuged and washed with ice-cold PBSA (PBS with 0.1% bovine serum albumin). The cell pellet was resuspended in ice-cold PBSA in order to reach the desired cell concentration and subsequently aliquoted to 96-well V-bottom plates. A master mix of anti- c-myc-Alexa Fluor 488 and anti-HA-Alexa Fluor 647 was prepared and added to reach a final concentration of 5 pg / mL anti-c-myc-Alexa Fluor 488 (Thermo Fisher, clone 9E10) and 1 pg / mL anti-HA-Alexa Fluor 647 (BioLegend, clone 16B12) in a final staining volume of 50 pL. The samples were incubated for 30 minutes (4 °C, shaking, in the dark) and subsequently washed twice with 200 pL ice-cold PBSA. Finally, the samples were analyzed with a Cytoflex S instrument (Beckman Coulter).

[0235] As shown in Figure 3 A, four different libraries were constructed based on SEQ ID NO: 1, in which different surface positions were randomly mutated with NNK-randomized primers as described above. 14 amino acid residues within the 5UMR protein were hypothesized to facilitate antigen binding, thus those residues were combined in various ways to establish the different libraries. Libraries 5UMR_cp, 5UMR_cy and 5UMR_cb hold ten randomized residues, while library 5UMR_c holds eight randomized positions. In library 5UMR_c positions R22, R24, S26, G29, 131, K33, V40 and N42 were randomized, in library 5UMR_cb positions E6, N8, R22, R24, S26, G29, 131, K33, V40 and N42 were randomized, in library 5UMR_cp R22, R24, S26, Q28, G29, 131, K33, V40, N42 and Q44 were randomized and library 5UMR_cy contains randomized positions R22, R24, S26, G29, 131, K33, S35, G38, V40 and N42. As shown in Figure 3B, in all four libraries, a certain fraction of the randomly mutated proteins could be expressed as full-length proteins, as demonstrated by detection of the C- terminal c-myc-tag (geometric mean fluorescence intensities of the anti-c-myc binding signal of the HA-positive population, background corrected, are visualized). Of note, full-length expression of proteins on yeast cells is known to correlate with protein folding and protein stability and can therefore be used as a measure for the overall biochemical fitness (i.e. folding properties and stabilities) of a yeast display library (Hackel et al., J Mol Biol. 2010 Aug 6;401(l):84-96; Traxlmayr and Obinger, Arch Biochem Biophys. 2012 Oct 15;526(2): 174-80). It is known that expression at 37 °C has a stronger negative effect on the expression of poorly folded proteins compared with expression at 20 °C (Shusta et al., Nat Biotechnol. 2000 Jul;18(7):754-9). In other words, the difference in expression levels between well-folded and poorly folded (i.e. more stable vs. less stable) proteins is more pronounced after induction at 37 °C, which was also observed in the experiments shown in Figure 3B, where all data were normalized to the non-mutated (and therefore highly stable) N-terminal domain of human SSRP1 (SEQ ID NO: 1, termed 5UMR WT in Figure 3B). Together, it can be concluded from these experiments that a certain fraction of the randomly mutated library members is properly folded and showed a reasonably high stability, enabling full-length expression on the surface of yeast cells, even after expression at 37 °C. In other words, it is possible to introduce random mutations into the N-terminal domain of human SSRP1, while preserving its overall fold, which is an important prerequisite for the selection of binders from a randomly mutated library. Moreover, these data clearly demonstrate the feasibility of using NNK-randomized primers for the construction of libraries based on the N-terminal domain of human SSRP1. Example 3 :

[0236] Construction of a library of binder scaffolds based on SEP ID NO: 1 with oligonucleotides that were synthesized by trimer synthesis

[0237] To be able to precisely control the amino acid distribution at the randomized positions, the investigators constructed an additional library based on SEQ ID NO: 1 by using oligonucleotides (primers) that were synthesized by trimer synthesis. Trimer synthesis allows for precise control of amino acid distributions in the randomly mutated binding surface (Traxlmayr et al., J Biol Chem. 2016 Oct 21;291(43):22496-22508). Since the NNK- randomized library 5UMR_cp performed well in the experiments shown in Figure 3B of Example 2 this library design (i.e. the same amino acid positions) was also used for the new library that was constructed with oligonucleotides synthesized by trimer synthesis.

[0238] Figure 4 shows the intended amino acid distributions in the randomly mutated binding surface. Since Cys-residues are undesired, no Cys-encoding codons were included. Furthermore, to reduce the hydrophobicity of the engineered binding surfaces, codons encoding for hydrophobic amino acids (F, I, L, M, V) were kept at a lower percentage of -3.0%. The same was done for arginine (i.e. also reduced to -3%), because this amino acid is known to promote unspecific binding and aggregation (Perchiacca et al., Protein Eng Des Sei. 2014 Feb;27(2):29- 39; Starr et al., Curr Opin Biotechnol. 2019 Dec:60: 119-127). Since Pro-residues introduce kinks in the amino acid backbone, the frequency of Pro was reduced to -2%. Since Tyr-residues are known to be important for antigen binding (Zemlin et al., J Mol Biol. 2003 Dec 5;334(4):733-49; Bogan and Thom, J Mol Biol. 1998 Jul 3;280(l): 1-9; Birtalan et al., J Mol Biol. 2008 Apr 11;377(5): 1518-28), the frequency of Tyr was elevated to -15%. All other amino acids were encoded with a frequency of -5.9% (Figure 4). Positions Q28 and Q44 were designed to encode for a different amino acid distribution, including 30% of wild-type glutamine (Q). The remaining 70% were divided to proportionately encode for the same amino acid distribution as explained before.

[0239] The library with these amino acid distributions in the randomly mutated binding surface was constructed as follows: the gene encoding the N-terminal domain of human SSRP1 was randomized and amplified in an initial PCR using primers 5UMR_cp_PCRl_fwd (SEQ ID NO:28) and 5UMR_PCRl_rev (SEQ ID NO:23) (5UMR_cp_PCRl_fwd synthesized by Ella Biotech, 5UMR_PCRl_rev synthesized by Sigma- Aldrich) using Q5® HiFi DNA Polymerase. X01 refers to codons encoding for a defined amino acid (AA) frequency in randomly mutated 5UMR positions except for Q28 and Q44 (i.e. positions R22, R24, S26, G29, 131, K33, V40 and N42), while X02 refers to codons encoding for an AA frequency for positions Q28 and Q44 in 5UMR, as specified in Figure 4. Gel-purified gene fragments of the correct amplicon size were amplified in large volume (200 pL per electroporation) with primers 5UMR_PCR2_fwd (SEQ ID NO:24) and 5UMR_PCR2_rev (SEQ ID NO:25) using Q5® HiFi DNA Polymerase; 10 ng of PCR1 product were used as template DNA (per 200 pL reaction volume). 20 electroporations of EBY100 cells with ethanol -purified DNA were performed (same electroporation procedure and preparation of EBY100 cells for electroporation as described in Example 2), reaching a diversity of 5.3xl08. The library was frozen in SD-CAA with 15 % glycerol and stored at -80 °C for subsequent usage.

[0240] To assess the overall quality of the library, the new yeast display library that was constructed with oligonucleotides synthesized by trimer synthesis (termed “5UMR_final”) was tested for full length expression of the yeast displayed constructs, as determined by measuring the C-terminal c-myc-tag, and also compared with the NNK-randomized library 5UMR_cp of Example 2. Yeast cultivation, induction of surface expression and analysis of c-myc-expression levels (i.e. full length protein display on yeast) were conducted as described in Example 2. In both of these libraries, the same amino acid positions were randomly mutated, but in 5UMR_cp NNK-randomized oligonucleotides were used (not allowing for precise control of amino acid distributions and also incorporating stop codons), whereas the library 5UMR_final was constructed with the oligonucleotides synthesized by trimer synthesis, allowing the inventors to control the amino acid distribution (Figure 4) and to exclude stop codons. As shown in Figure 5, the 5UMR_final library performed much better than the 5UMR_cp library after expression at either 20 °C or 37 °C, demonstrating that a larger fraction of the library 5UMR_final was expressed as properly folded full-length proteins.

[0241] To analyze whether the obtained amino acid distribution matches the initially intended one (shown in Figure 4), plasmids were isolated from the yeast library 5UMR_final and analyzed with respect to the amino acid distribution at the randomly mutated positions. Plasmid DNA was isolated using the Zymoprep Yeast Plasmid Miniprep Kit II (Zymo Research) and used for the electroporation of NEB® 10-beta Electrocompetent E. coli (New England Biolabs). Electroporation cuvettes (1 mm) and microcentrifuge tubes were placed on ice, while E. coli cells were thawed slowly (10 minutes) on ice. Subsequently, E. coli cells were mixed by flicking and 25 pL cells were transferred to each chilled microcentrifuge tube before 1 pL of the Zymoprep isolate (yeast DNA) was added. The cell suspension was carefully transferred to the bottom of the cuvette and electroporated using the Gene Pulser Xcell Electroporation system (Bio-Rad) with the following conditions (2.0 kV, 200 Omega, 25 pF). Immediately after the electroporation 975 pL of pre-warmed NEB outgrowth medium was added to the cuvette, gently resuspended and transferred to a microcentrifuge tube. Cells in NEB outgrowth medium were cultivated shaking at 37 °C, 300 rpm for 1 hour, before plating several volumes of the cell suspension on pre-warmed selective LB agar plates and incubated at 37 °C overnight. E. coli clones were used for the inoculation of selective LB medium and plasmids were subsequently isolated and sequenced in a 96-well plate format by Microsynth AG. The resulting nucleotide sequence were translated into all three reading frames using EMBOSS Transeq (https: / / www.ebi.ac.uk / Tools / st / emboss_transeq / ) and the correct AA translation was used for subsequent Multiple Sequence Alignment (MSA) using Clustal Omega (https: / / www.ebi.ac.uk / Tools / msa / clustalo / ). Finally, the amino acid distributions were analyzed at the randomly mutated amino acid positions in the 5UMR_final library based on SEQ ID NO: 1. 81 gene sequences (i.e. 81 library members) were analyzed.

[0242] As shown in Figure 6, the amino acid distributions in the library (analyzed separately for positions Q28 and Q44 or all the remaining randomized positions within the N-terminal domain of human SSRP1) closely matched the intended amino acid distribution. For example, Tyr was the most abundant amino acid, Cys codons were not detected at all and Pro was detected only at very low frequencies (Figure 6), as intended in the library design.

[0243] Overall, the data presented in Example 3 clearly demonstrate the feasibility of constructing a library based on N-terminal domain of human SSRP1 (i.e. based on SEQ ID NO: 1) by using oligonucleotides that were synthesized by trimer synthesis and that the obtained amino acid distributions indeed closely match the intended amino acid frequencies.

[0244] Example 4:

[0245] Isolation of high-affinity binders with beneficial biophysical properties from a library based on the N-terminal domain of human SSRP1

[0246] To test whether antigen-specific binders can be isolated from a library based on the N-terminal domain of human SSRP1, the investigators applied the yeast surface display technology using standard yeast display magnetic bead selections, as well as flow cytometric sorting that are well-known to a person skilled in the art (Angelini et al., Methods Mol Biol. 2015;1319:3-36).

[0247] Yeast display selections were carried out with CD22 and CD276 proteins. As the starting protein library holds roughly 5.3xl08different protein variants (=diversity), IO10cells were used for the initial rounds of selections in order to “oversample” the library diversity. The selection campaign started with magnetic bead selections using Dynabeads™ Biotin Binder (Thermo Fisher Scientific) as described in Chen et al., Methods Enzymol. 2013;523:303-26. Yeast display selections for binders against CD22 were based on a soluble, biotinylated CD22 protein by AcroBiosystems (SI2-H82E3). Initially, 2 rounds of bead selections were performed with the naive library, followed by a round of affinity maturation (i.e. error prone PCR) and several rounds of FACS selections, including a round of negative sorting using a CD276 protein, which carried the same tags (Biotinylated Human B7-H3 (41g), His, Avitag; B7B-H82E8).

[0248] Yeast display selections for binders against CD276 incorporated the use of two recombinantly expressed proteins by AcroBiosystems (Biotinylated Human B7-H3 (41g), B7B- H82E8 and Human B7-H3, B73-H52E2), blocking strategies with well-known anti-CD276 antibody Omburtamab and cell based pannings with Caco-2 cells. Initially, all sorting strategies started with a naive 5UMR library and two rounds of bead selections with CD27641g, followed by a round of affinity maturation (i.e. error prone PCR) and one FACS selection with CD276 41g. Subsequently, the yeast library was split up to five different sorting campaigns including selections against CD276 41g only (with and without blocking sorts with Omburtamab), selections incorporating both 21g and 41g isoforoms (with and without blocking sorts with Omburtamab) and selections incorporating cell pannings on Caco-2 cells (according to Panton and Stern, Methods Mol Biol. 2022;2491:217-233) as well as FACS selections with CD27641g.

[0249] In order to increase the affinity of the selected binders, rounds of affinity maturations were incorporated into the selection campaigns. For the introduction of random mutations, the GeneMorph II Random Mutagenesis Kit (Agilent) was used. Yeast DNA was isolated using the Zymoprep Yeast Plasmid Miniprep Kit II (Zymo Research) and subsequently amplified with Q5® HiFi DNA Polymerase (NEB) using primers 917 (SEQ ID NO: 29) and 918 (SEQ ID NO:30). An error prone PCR using the GeneMorph II Kit was conducted employing 500 ng of template DNA from the previous PCR, primers 917 and 918 and 20 cycles, as the number of mutations per gene can be varied by template amount and cycle number. The gel-purified amplicon of the correct size was amplified in a subsequent PCR reaction of large volume (200 pL) with primers 917 and 918 for a subsequent yeast electroporation.

[0250] For cell sorting either a FACS Aria™ Fusion cell sorter (BD Biosciences) or SH800S cell sorter (Sony Biotechnology) were used. Either 3xl07or 5xl06yeast cells from an induced library were used for antigen staining and subsequent sorting. The relevant number of yeast cells were harvested and washed twice with 1 mL ice-cold PBSA. Yeast cells were stained with the relevant antigen concentration in either 500 or 100 pL staining volume for one hour (4 °C, shaking / rotating). Subsequently cells were washed twice with 1 mL ice-cold PBSA and stained with a secondary staining reagent (100 pL staining volume, 30 minutes, 4 °C, rotating, in the dark) with either 5 pg / mL anti-penta-His-AF647 (Qiagen), 5 pg / mL anti-penta-His-AF488 (Qiagen), 5 pg / mL Streptavidin-AF67 (Invitrogen) or 5 pg / mL Streptavidin-AF488 (Invitrogen) and washed again twice with 1 mL ice-cold lx PBSA. In order not to carry over free antigen from the first staining step, washing and removal of supernatant was carried out with extra care prior to the secondary staining step. For expression normalization the yeast library sample was always additionally stained with an anti-HA or anti-c-myc antibody coupled to a matching fluorophore (AF488 or AF647, depending on which channel was already occupied by the reagent used for detection of antigen binding) during one of the two staining steps. Prior to sorting the yeast cells were resuspended in ice-cold PBSA and finally incubated in SD-CAA with 100 U / mL penicillin and 100 pg / mL streptomycin at 30 °C after sorting.

[0251] After yeast display selection, plasmid DNA was isolated from the enriched yeast cells using the Zymoprep Yeast Plasmid Miniprep Kit II (Zymo Research) and used for the electroporation of NEB® 10-beta Electrocompetent E. coli (New England Biolabs) as described in Example 3. Next, E. coli clones were used for the inoculation of selective LB medium and plasmids were subsequently isolated and sequenced in a 96-well plate format by Microsynth AG as described in Example 3. After sequence analysis, pCTCON2V plasmids encoding enriched clones were transformed again into Saccharomyces cerevisiae strain EBY100 using Frozen-EZ Yeast Transformation II kit (Zymo Research). Yeast cultures displaying individual clones were cultivated and surface expression was induced in order to measure antigen binding affinities of yeast displayed binders.

[0252] To additionally analyze the biophysical properties of the isolated binders, as well as their binding properties to human cells expressing their target antigens, their encoding genes were sub-cloned into pET-21a(+) vectors and expressed as soluble proteins with an N-terminal linker and a C-terminal (2xG4S)-G linker followed by a hexahistidine (6xHis) tag as described in Example 1 and SEQ ID NO: 19. Also the protein purification was performed as described in Example 1.

[0253] Binding studies of CD276 binders were performed with the following cell lines: A431, A549, SK-BR-3, Caco-2 and Jurkat. Binding experiments with CD22 binders were performed with Jurkat, NALM6 and Raji cells.

[0254] Binding experiments with CD22 binders: Cells were harvested and washed with ice- cold PBSA (200 g for Raji, 300 g for NALM6 and Jurkat cells, 5 min, 4 °C). For affinity determination, CD22 binders were used for titration experiments on NALM6 cells. 100,000 cells were stained with concentrations of 0 - 250 or 1000 nM CD22 binders in 200 pL staining volume for 4 hours (4 °C, shaking, 96-well V-bottom plates). Subsequently, the samples were washed twice with 200 pL ice-cold PBSA and stained with 5 pg / mL anti-penta-His-AF488 antibody (Qiagen) for 20 - 30 minutes (4 °C, shaking, dark) in 25 pL staining volume to detect the His-tag on the binders. After two more washing steps the cells were pelleted and resuspended just before measurement with a Cytoflex S instrument (Beckman Coulter). Data were fitted to a 1 : 1 binding model to calculate the KD values. Analysis of unspecific binding (with NALM6, Raji and Jurkat cells) was performed likewise, but staining was only performed with 250 nM CD22 binder in 50 pL staining volume for 1 hour (4 °C, shaking). The data shown in Fig. 20 clearly demonstrate that the engineered binders only bind to cells that express the target antigen CD22 (NALM6 and Raji), but not to CD22-negative cells (Jurkat), demonstrating that they are highly specific for their antigen and do not show detectable unspecific binding.

[0255] Binding experiments with CD276 binders: Titrations with CD276 binders were performed likewise as described for the analysis of CD22 binders, but 100,000 cells were stained with 0 - 4000 nM of CD276 binders in 50 pL staining volume for 1 hour (4 °C, shaking, 96-well V-bottom plates). Analysis of unspecific binding (with Caco-2 and Jurkat cells) was performed likewise, but staining was performed with 1000 nM CD276 binder in 50 pL staining volume for 1 hour (4 °C, shaking). All centrifugation steps were carried out at 200 g, 5 minutes, 4 °C.

[0256] Data were analyzed as described previously (Zajc et al., Methods Mol Biol. 2022;2491 : 155-173). The obtained titration curves for binders 22 1611 (SEQ ID NO:31), 22_154 (SEQ ID NO:32), 22_164 (SEQ ID NO:33), 22_1415 (SEQ ID NO:34), 22_163 (SEQ ID NO: 35), 22 1410 (SEQ ID NO: 36), 22 1316 (SEQ ID NO: 37), 22 1317 (SEQ ID NO: 38) and 22 1333 (SEQ ID NO:39) on NALM6 cells are shown in Figure 7B and the corresponding K values (i.e. affinities) that were obtained from the fitted curves are shown in Figure 7A. Binders are visualized in either light or dark grey according to the selection campaign they originate from. As demonstrated by the values shown in Figure 7A, binders with affinities in the sub-nanomolar (i.e. picomolar) to double digit nanomolar range were obtained. Thus, binders with antibody-like affinities can be isolated from libraries based on the N-terminal domain of human SSRP1.

[0257] To analyze their thermostability, all binders were analyzed by differential scanning calorimetry (DSC) using a MicroCai PEAQ-DSC instrument (Malvern Panalytical; binder concentration of 50 pM in PBS; scan from 20 - 100 or 120 °C). Data analysis was performed with the MicroCai PEAQ-DSC Software by Malvern Panalytical performing a buffer baseline subtraction followed by normalization for protein concentration and fitting to a non-two state unfolding model. As shown in Figure 8A, all binders were stable with Tmvalues of 51.0-68.3 °C. Binders are visualized in either light or dark grey according to the selection campaign they originate from. Although this represents some loss in stability when compared with the parental protein (the N-terminal domain of human SSRP1, labeled with 5UMR in Figure 8A), this is expected and usually observed during protein engineering processes (Teufl et al., ACS Synth Biol. 2022 Mar 18; 11(3): 1030-1039). Nevertheless, despite the slight loss in stability, the obtained binders are still stable proteins. In fact, three of the binders (22 161 Im 22 163 and 22 1410) showed a remarkable small reduction in melting temperature, accounting for only two to three °C. This neglectable loss in thermostability illustrates the great tolerance for mutations of the established protein library.

[0258] To analyze the aggregation tendencies of the engineered binders, they were analyzed by size exclusion chromatography (SEC) as described in Example 1. As shown in Figure 8B, all binders were mainly monomeric, as indicated by one main elution peak. While most binders showed comparable elution time to the parental protein (i.e. 5UMR in this Figure), some binders showed a slight or significant increase in the protein elution time (22 1333).

[0259] Binders against CD276 were solubly expressed and characterized as described above. As shown in Figure 9B, the Tmof the binders was reduced to 56.5 to 63.7 °C. Despite the loss in thermostability the binders were still stable as represented by their low aggregation tendencies analyzed using SEC (Figure 9A). Binders 276_C152 (SEQ ID NO:40) and 276_152 (SEQ ID NO:41) eluted as monomeric proteins, as indicated by the main peak at an elution time of 19 min, while binder 276 C146 (SEQ ID NO:42) showed an increase in the elution time to 22 min.

[0260] The obtained binding curves for titration of CD276 binders on a panel of CD276- positive cells lines (A-431, A549, SK-BR-3 and Caco-2 cell lines) are shown in Figure 10 A-D and the corresponding KD values are shown in Figure 11 A. Analysis of unspecific binding (data shown in Figure 11C) revealed, that the engineered binders only bind to cells that express the target antigen CD276 (Caco-2), but not to CD276-negative cells (Jurkat), demonstrating that they are highly specific for their antigen and do not show detectable unspecific binding.

[0261] In order to assess the potential binding epitope of CD276 binders, blocking experiments with well-known anti-CD276 antibodies MGA271 and Omburtamab were carried out. 100,000 Caco-2 cells were washed once with ice-cold PBSA and subsequently blocked with 666 nM of Omburtamab (Thermo Fisher Scientific) or MGA271 (Abeomics) for 1 hour (4°C, shaking, 96- well V-bottom plates) in 18.75 pL staining volume. Subsequently, CD276 binder was added to reach a final concentration of 500 nM binder and 500 nM antibody in a staining volume of 25 pL. Samples were incubated for 1 hour (4 °C, shaking, 96-well V-bottom plates) and subsequently washed twice with 200 pL ice-cold PBSA (200 g, 5 minutes, 4 °C). All subsequent washing steps were conducted likewise. After a staining step with 5 pg / mL anti-penta-His- AF488 antibody (Qiagen) for 30 minutes (4 °C, shaking, dark) to detect the His-tagged binders, two more washing steps were conducted prior to analysis with a Cytoflex S instrument (Beckman Coulter). As illustrated in Figure 1 IB, by assessing the amount of residual binding it seems that binders 276_C152, 276_152 and 276_C146 do not bind to the same binding epitope as MGA271, but indeed seem to be partially “blocked” (i.e. prevented from binding) by antibody Omburtamab. This result suggests, that binders 276_ Cl 52, 276_152 and 276 C146 might bind to an overlapping epitope with Omburtamab.

[0262] Together, the data presented in this example demonstrate that it is possible to engineer binders based on the N-terminal domain of human SSRP1 for different protein antigens and that these engineered binders show high affinities to their antigen (Figures 7A, 7B, 10A-D and 11 A), that they are stable (Figure 8A and 9B) and that they show only minimal or no detectable aggregation (Figure 8B and 9A).

[0263] Example 5:

[0264] Proofing CAR T cell functionality of obtained anti-CD22 and anti-CD276 binders isolated from a library based on the N-terminal domain of human SSRP1

[0265] In order to assess whether novel anti-CD22 and anti-CD276 binders would show CAR T function, a panel of different experiments were conducted. First, transiently transformed Jurkat Nur77 reporter cells (cell line generated by Dobersberger et al. Cell Rep Methods. 2024 Mar 12:100728) were used to assess CAR activation with and without target cells. Subsequently, human primary T cells were transduced with a chosen set of CARs based on engineered binders or benchmark scFvs and used for cytotoxcitiy assays followed by the detection of cytokine release.

[0266] In-vitro transcription of CAR transgenes was performed with Hi Scribe T7 ARC A mRNA Kit (New Eng-land Biolabs), followed by RNA purification with Monarch RNA Cleanup Kit (New England Biolabs). Electroporation was performed as described in Example 1, using 2 x 106Jurkat Nur77 reporter cells, 5 pg mRNA and a square wave protocol (single pulse, 500 V, 4 mm, 3 ms pulse length). Cells were rescued in full-growth medium without IL-2 subsequently. After 2 hours of incubation, cells were counted using a Vi-CELL XR (Beckman Coulter). 50,000 Jurkat Nur77 reporter cells were co-cultured with 100,000 target cells (A549 for aCD276-CARs, NALM6 for aCD22-CARs) or no target cells in 200 pL volume in sterile 96- well U-bottom plates in full-growth medium overnight (15-18 hours of incubation). The samples were transferred to 96-well V-bottom plates and washed once with 200 pL ice-cold PBSA (300 g, 5 min, 4 °C). All subsequent washing steps were conducted likewise. Cells were resuspended in 50 pL PBSA containing 10% of human serum (PAN Biotech) and incubated for 10 minutes at 4 °C. Subsequently, samples were stained with 2.5 pg / mL anti-MAP-AF647 antibody (BioLegend, clone pmab-1) for 30 minutes (4 °C, shaking, dark) and finally washed two times prior to analysis with a Cytoflex S instrument (Beckman Coulter).

[0267] For the transduction of primary human T cells, Lenti-X 293T cells (Takara) were split one day prior to transfection in DMEM (Thermo Scientific) + 10% FCS (Capricorn Scientific) to reach confluency of 70-80% at the time of transfection. Lenti-X 293T cells were co-transfected with a pCDH expression vector (System Biosciences) carrying puromycin resistance genes and second-generation viral packaging plasmids pMD2.G and psPAX2 (Addgene plasmids #12259 and #12260, respectively; gifts from Didier Trono) using the PureFection Transfection Reagent (System Biosciences). The transfection mixture was vortexed for 10 seconds and incubated at RT for 15 minutes, before the dropwise addition to the cells. Supernatants were collected on day two and three after transfection and concentrated 50x using Lenti-X Concentrator (Takara). Lentiviral particles were resuspended gently in AIM V (Thermo Scientific) supplemented with 2% Octaplas (Octapharma) supplemented with 2.5% HEPES (PAN Biotech), 1% L-Glutamine (Thermos Fisher Scientific) and 200 U / ml IL-2 (PeproTech) and stored at -80 °C. One day prior to transduction primary human T cells were activated with T-Activator CD3 / CD28 dynabeads (Thermo Fisher Scientfic) according to the manufacturer’s instructions in AIM V™ medium supplemented with 2% Octaplas, 2.5% HEPES, 1% L-Glutamine and 200 U / ml IL-2. Activated T cells (106cells / mL) were transduced in retronectin (Takara) coated cell culture dishes with thawed lentviral suspension at a final dilution of 1 :2. To ensure growth of transduced cells only, puromycin was added to a final concentration of 1 pg / mL for 48 hours two days post transduction. Transduced T cells were expanded at densities below 1.5 x 106cells / mL to ensure high cell fitness.

[0268] A luciferase-assay was used to determine the cytotoxicity of primary human CAR T cells. For anti-CD22 CAR T cells, GFP & luciferase expressing target cell lines Nalm6 and Raji were cocultured at effector to target (E:T) ratios 2: 1 and 5 : 1 as indicated for either 4 h (Nalm6) or 24 h (Raji) in 100 pL RPMI without phenol red (Thermo Fisher Scientifc, RPMI 1640 Medium, no phenol red) supplemented with 2% Octaplas and 1% penicillin-streptomycin (Thermo Fisher Scientific) in white round-bottom 96-well plates (Coming). For anti-CD276 CAR T cells GFP & luciferase expressing target cells lines A431, A549 and Caco-2 were seeded 16 h prior to the assay in 100 pL RPMI without phenol red supplemented with 2% Octaplas and 1% penicillin- streptomycin in white round-bottom 96-well plates. On the following day 50 pL primary T cells were added to reach E:T ratios of 5: 1 and 10: 1 for either 6 or 24 hr. After co-culture the number of viable cells was determined by their luciferase activity. The plates were equilibrated 10 min at RT, before luciferin (Revvity) was added to reach a final concentration of 150 pg / mL.

[0269] Luminescence was quantified after 20 min at RT using an ENSPIRE Multimode plate reader

[0270] (Perkin Elmer). The lysis of target cells was calculated as following:

[0271] , . , _ _

[0272] % lysis = 100 100

[0273] Cytokine secretion was quantified in the supernatants from cytotoxicity experiments (described above). Supernatants were centrifuged (500 g, 5 min, 4 °C) and subsequently stored at -20 °C. IFN-y and IL-2 were quantified by enzyme-linked immunosorbent assay (ELISA) using ELISA MAX Deluxe Set for human IFN-y and IL-2 (Biozym).

[0274] All in-vitro experiments were carried out with the same CAR architecture, holding a signal peptide (SP), followed by a MAP -tag and G4S linker, the respective binder or scFv, CD28 hinge and transmembrane domain, CD28 co-stimulatory domain and CD3^. As a representative example binder 22 1410 is shown in the respective CAR backbone in SEQ ID NO: 43.

[0275] Expression analysis of aCD22-CARs on Nur77 Jurkat reporter cells revealed, that expression varies greatly between ten CD22 binder-CARs. CARs based on four CD22 binders (22_1410, 22_163, 22 1611 and 22_1333) show comparable expression to benchmark scFv- CARs HA22 (SEQ ID NO:44), HA22p (SEQ ID NO:45), m971-lxG4S (SEQ ID NO:46) and m971-4xG4S (SEQ ID NO:47) (Fig 12A). Two different HA22 sequences were tested due to inconsistent sequences found in literature and two different linker lengths for m971, as the linker length has been shown to have an effect on the function (Singh et al., Nat Med. 2021 May;27(5):842-850.). Activation was assessed via the expression of mKusabira Orange (mK02) under the control of a Nur77 promoter. Nur77 is known to be a gene involved in the early activation of T cell response, thus it is used as a readout for T cell activation. CARs based on ten CD22 binders and four benchmark scFvs were tested for their activation with and without the addition of target cells NALM6 in a two-fold excess (Fig. 12B). All CARs based on CD22 binders showed activation after the addition of target cells, which was comparable to CARs based on the benchmark scFvs. Interestingly, the level of activation without the addition of target cells varied greatly between the ten binder-CARs. While CARs based on all benchmark scFvs and seven out of ten CD22 binders showed antigen-independent activation and signalling, three CD22 binders (22_1410, 22_163 and 22 1611) showed hardly any antigen-independent activation. Interestingly, those binder-CARs also show best expression among the CD22 binder-CARs based on the N-terminal domain of human SSRP1 (Fig. 12A). No or low antigenindependent activation (commonly referred to as “tonic signalling”) could potentially be highly beneficial. Low tonic signalling is often correlated with better T cell function and persistence (Long et al., Nat Med. 2015 Jun;21(6):581-90).

[0276] Luciferase cytotoxicity assays with NALM6 and Raji target cells at various E:T ratios (2: 1 and 5: 1) revealed that CARs comprising CD22 binders based on the N-terminal domain of human SSRP1 recognize surface displayed CD22 antigen und lead to target cell lysis (Fig. 13). In fact, novel CAR T cells based on CD22 binders showed superior lytic activity on target cells NALM6 compared to well established scFv m971 -based CAR T cells, irrespective of the linker length used (Fig. 13A). As a positive control a FMC63-BBz CAR (recognizing a different antigen on the target cells: CD 19) was used and shows cytotoxic activity that is comparable to that of novel CD22 binder-CARs (Fig. 13 A). When Raji cells were used as target cells, the lytic activity of m971 scFv-CARs was increased, but still, novel CD22 binder-CAR T cells showed significant lytic activity compared to mock T cells (Fig. 13B). Cytokines IFN-y and IL-2 were analysed in the supernatants from cytotoxicity assays and the analysis revealed, that CAR T cells with CD22 binders based on the N-terminal domain of human SSRP1 released both cytokines at comparable or superior amounts compared to benchmark scFv m971 -based and positive control FMC63BBz-based CAR T cells, when NALM6 (N6) cells were used as targets (Fig. 14). When Raji cells were used as target cell line, IFN-y was released to comparable levels between novel CD22 binder-CAR T cells and m971 scFv CAR T cells (Fig. 15 A), while the expression of IL-2 was increased for m971-based and FMC63-BBZ-based CAR T cells compared to 22_1410-, 22_163- and 22_1333-based CAR T cells (Fig. 15B). Overall, these data illustrate cytotoxic activity and in-vitro function comparable to clinically used scFv m971- based CAR T cells and demonstrate feasibility and functionality of using binders based on the N-terminal domain of human SSRP1 as an antigen-binding domain in CAR T cells.

[0277] Expression analysis of aCD276-CARs on Nur77 Jurkat reporter cells revealed lower expression of CARs based on 276_C146 and 276_152 compared to CARs based on 276_C152 and benchmark scFvs MGA271 (SEQ ID NO: 48) and 376.96 (SEQ ID NO:49) (Fig. 16A). Nevertheless, aCD276-CARs showed remarkable activation, despite the lower expression level (Fig. 16B). In fact, CARs based on 276_C146, 276_152 and 276_C152 showed activation after the addition of target cells A549 at a two-fold excess that is comparable to the activation of benchmark scFv-CARs based on MGA271 and 376.96 (Fig. 16B). Analysis of cytotoxicity was performed by co-incubation of transduced human primary T cells with three different target cells (A431, A549 and Caco-2) at different E:T ratios (5: 1 and 10: 1) for six or 24 hours. After six hours of co-incubation only target cells A431 and A549 (at 10: 1 E:T) were lysed by CAR T cells based on 276_152, 276_C152, 376.96 and MGA271 (Fig. 17A). CAR T cells based on binder 276_152 showed lower lytic activity, while CAR T cells based on binder 276_C152 showed lytic activity comparable to benchmark CAR T cells based on scFvs 376.96 and MGA271. After 24 hours of co-incubation, lytic activity could be shown for all CARs and cell lines (Fig. 17B). CAR T cells based on binder 276 152 showed lower lytic activity compared to those based on binder 276 C152 and benchmark scFvs, irrespective of the target cell line and E:T ratio (Fig. 17B). Cytokines IFN-y and IL-2 were analysed in the supernatants of cytotoxicity assays and the analysis revealed that after six hour of co-incubation CAR T cells based on binder 276 C152 showed IFN-y release comparable to those based on benchmark scFvs MGA721 and 376.96 (Fig. 18A), while CAR T cells based on scFv MGA271 showed a higher release of IL-2. CAR T cells based on binder 276 C152 showed the second highest release of IL-2 (Fig. 18B). After 24 hours of co-incubation, IFN-y release of CAR T cells based on binder 276_C 152 was comparable to CAR T cells based on 376.96 and MGA271 (Fig. 19A), while release of IL-2 was measured to be highest for CAR T cells based on scFv MGA271 and second-highest for those based on binder 276 C152 (Fig. 19B).

[0278] Overall, these data illustrate cytotoxic activity and in-vitro function and demonstrate feasibility and functionality of using binders based on the N-terminal domain of human SSRP1 as an antigen-binding domain in CAR T cells.

[0279] Description of the sequences of the sequence protocol

[0280] SEQ ID NO: 1 : 5UMR wild-type sequence

[0281] SEQ ID NO:2: 1KIK

[0282] SEQ ID NO:3: 2CQA

[0283] SEQ ID NO:4: 2DMH

[0284] SEQ ID NO: 5: 2E7C

[0285] SEQ ID NO:6: 2K1M

[0286] SEQ ID NO:7: 2K45

[0287] SEQ ID NO: 8: 2P9R

[0288] SEQ ID NO:9: 2V37

[0289] SEQ ID NO: 10: 2VH7

[0290] SEQ ID NO: 11 : 3135

[0291] SEQ ID NO: 12: 3MX7

[0292] SEQ ID NO: 13: 3SHU

[0293] SEQ ID NO: 14: 5NOO

[0294] SEQ ID NO: 15: 6MYF

[0295] SEQ ID NO: 16: FMC63

[0296] SEQ ID NO: 17: FN3

[0297] SEQ ID NO: 18: 1KIK in CAR backbone SEQ ID NO: 19 Soluble 1KIK protein with linkers and tags SEQ ID NO:20 Primer 5UMR_c_NNKl SEQ ID NO:21 Primer 5UMR_cp_NNKl SEQ ID NO:22 Primer 5UMR_cy_NNKl SEQ ID NO:23 Primer 5UMR_PCRl_rev SEQ ID NO:24 Primer 5UMR_PCR2_fwd SEQ ID NO:25 Primer 5UMR_PCR2_rev SEQ ID NO:26 Primer 5UMR_cb_NNK2 SEQ ID NO:27 Primer pCTCON2V sequence including insert 5UMR wild type SEQ ID NO:28 Primer 5UMR_cp_PCRl_fwd

[0298] The first, second, third, fifth, sixth, seventh, eighth and nineth NNN-triplet from 5' to 3' end within the sequence of SEQ ID NO:28 are termed X01, and the fourth and tenth NNN-triplet from 5' to 3' end within the sequence of SEQ ID NO:28 are termed X02. X01 refers to codons encoding for a defined amino acid (AA) frequency in randomly mutated 5UMR positions except for Q28 and Q44 (i.e. positions R22, R24, S26, G29, 131, K33, V40 and N42), while X02 refers to codons encoding for an AA frequency for positions Q28 and Q44 in 5UMR, as specified in Figure 4.

[0299] SEQ ID NO:29 Primer 917 SEQ ID NO:30 Primer 918 SEQ ID NO:31 CD22 binder 22 1611 SEQ ID NO:32 CD22 binder 22_154 SEQ ID NO:33 CD22 binder 22_164 SEQ ID NO:34 CD22 binder 22_1415 SEQ ID NO:35 CD22 binder 22_163 SEQ ID NO:36 CD22 binder 22_1410 SEQ ID NO:37 CD22 binder 22 1316 SEQ ID NO:38 CD22 binder 22 1317 SEQ ID NO:39 CD22 binder 22_1333 SEQ ID NO:40 CD276 binder 276_C152 SEQ ID NO:41 CD276 binder 276_152 SEQ ID NO:42 CD276 binder 276_C146 SEQ ID NO:43 Binder 22 1410 in a CAR backbone SEQ ID NO:44 scFv HA22 SEQ ID NO:45 scFv HA22_p SEQ ID NO:46 scFv m971-lxG4S SEQ ID NO:47 scFv m971-4xG4S SEQ ID NO:48 scFv mGA271 SEQ ID NO:49 scFv 376.96

Claims

Claims1) A method of constructing a library of protein scaffolds comprising the steps a) providing an initial polypeptide, wherein said initial polypeptide comprises a polypeptide having at least 90% identity to SEQ ID NO: 1, b) introducing diversity into copies of said initial polypeptide to form the protein scaffold library, wherein the introducing diversity step comprises mutating one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, wherein said protein scaffolds comprise amino acid sequences that have at least 80% identity to Seq ID NO: 1, respectively.2) The method of claim 1, wherein the introducing diversity step comprises mutating at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.3) The method of claim 1 or 2, wherein said introducing diversity step is performed by using random mutagenesis on the level of the nucleic acid sequence encoding said polypeptide having at least 90% identity to SEQ ID NO: 1.4) The method of any one of claims 1 to 3, wherein the introducing diversity step comprises mutating one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 but does not comprise the introduction of any cysteine at said one or more amino acid positions.5) The method of any one of claims 1 to 3, wherein said introducing diversity step is performed by using random mutagenesis on the level of the nucleic acid sequence encoding said polypeptide having at least 90% identity to SEQ ID NO: 1 with degenerate oligonucleotides containing randomly mutated codons at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, a) wherein said randomly mutated codon positions within the oligonucleotides are encoded by NNN, NNK, NNS and / or NNB codons, wherein according to the IUPAC nucleotide code “N” encodes for any of the bases adenine, cytosine, guanine or thymine, “K” encodes for bases guanine or thymine, “S” encodes for bases guanine or cytosine and “B” encodes for bases cytosine, guanine or thymine, and / orb) wherein said randomly mutated codon positions within the oligonucleotides are synthesized by trimer synthesis.6) The method of claim 5, wherein said randomly mutated codon positions within the oligonucleotides are synthesized by trimer synthesis, and said oligonucleotides do not code for any cysteine at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.7) The method of claim 5 or 6, wherein said randomly mutated codon positions within the oligonucleotides are synthesized by trimer synthesis, and said oligonucleotides code for a reduced percentage of prolines and a higher percentage of tyrosines as compared to the codon degeneracy at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, wherein said reduced percentage of prolines is defined by about 3% or less prolines that are incorporated into each individual amino acid position at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1, and wherein said higher percentage of tyrosines is defined by about 10% or more tyrosines that are incorporated into each individual amino acid position at said one or more of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1.8) The method of any one of claims 1 to 7, wherein said method comprises step c: displaying the diversified copies of said initial polypeptide in a yeast surface display, phage display, mRNA display, DNA display, bacterial display, mammalian cell display or ribosome display format.9) A library of protein scaffolds comprising diversified copies of an initial polypeptide, wherein said initial polypeptide comprises or consists of a polypeptide having at least 90% identity to SEQ ID NO: 1, and wherein said library of protein scaffolds is obtained by the method of any one of claims 1 to 8.10) A method of isolating an antigen binding domain, that binds to an antigen, comprising contacting the library of claim 9 with said antigen and isolating an antigen binding domain that binds to said antigen.11) The method of isolating an antigen binding domain of claim 10, comprising contacting the library of claim 9 with said antigen and isolating an antigen binding domain that binds to said antigen, and wherein said isolated antigen binding domain does not comprise any cysteine.12) The method of claim 10 or 11, wherein the isolating step comprises isolating antigen binding domains binding to said antigen and testing the isolated antigen binding domains for binding affinity to said antigen, wherein optionally:(i) the isolating step comprises incubating the library with said antigen, identifying antigen binding domains binding to said antigen, and identifying the sequences encoding for the antigen binding domains binding to said antigen; and / or(ii) the affinity is defined by a KD less than 10'5M.13) A fusion protein comprising a) an antigen binding domain that binds to an antigen as a first polypeptide comprising an amino acid sequence that has at least 80% identity to SEQ ID NO: 1, but in which at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO:1 are altered in order to bind a target, wherein said at least 5 of said amino acid positions that are altered do not comprise any cysteine, and b) a second polypeptide, wherein said second polypeptide is an effector molecule.14) A chimeric antigen receptor (CAR) comprising a) an antigen binding domain, wherein said antigen binding domain binds to an antigen, b) a transmembrane domain, and c) an intracellular signaling domain, wherein said antigen binding domain is or comprises an amino acid sequence that has at least 80% identity to SEQ ID NO: 1, but in which at least 5 of the amino acid positions R22, R24, S26, Q28, G29, 131, K33, V40, N42 and / or Q44 of SEQ ID NO: 1 are altered in order to bind a target, wherein said at least 5 of said amino acid positions that are altered do not comprise any cysteine.15) The use of the N-terminal domain of human FACT complex subunit SSRP1 for the generation of a library of protein scaffolds, wherein said N-terminal domain of human FACT complex subunit SSRP1 comprises or consists of an initial polypeptide having SEQ ID NO: 1, and wherein said use comprises introducing diversity into copies of said initial polypeptide, andwherein said copies of said initial polypeptide comprise an amino acid sequence that have at least 75% identity to said initial polypeptide, respectively.

Images