Interleukin-15 variants

Abstract

The present invention provides interleukin-15 (IL-15) variants containing amino acid substitutions for improved homogeneity, as well as conjugates and fusion proteins comprising such IL-15 variants. Also provided are nucleic acids, vectors, and host cells for the expression of such IL-15 variants, and pharmaceutical compositions comprising such IL-15 variants.

Description

[Technical field]

[0001] The present invention relates to interleukin-15 (IL-15) variants containing amino acid substitutions for improved homogeneity, as well as conjugates and fusion proteins comprising such IL-15 variants. [Background technology]

[0002] Interleukin 15 (IL-15) regulates cytotoxic lymphocytes and memory phenotype CD8 + It is a naturally occurring cytokine that induces the generation of T cells and stimulates the proliferation and maintenance of natural killer (NK) cells, but in contrast to interleukin-2, it does not mediate activation-induced cell death, does not consistently activate Tregs, and is less likely to cause capillary leak syndrome (Waldmann et al., 2020). Extensive preclinical and clinical studies have been conducted that demonstrate the efficacy and limitations of IL-15 and the growing number of IL-15 analogues / supergonists, particularly in the treatment of cancer, and are reviewed by Robinson and Schluns (Robinson and Schluns 2017).

[0003] IL-15, like interleukin 2 (IL-2), acts through a heterotrimeric receptor with α, β, and γ subunits, but they share a common γ chain receptor (γ cor gamma) and IL-2 / IL-15Rβ (IL-2Rβ, also known as CD122), with a gamma chain receptor also shared with IL-4, IL-7, IL-9 and IL-21. This heterotrimeric receptor contains specific subunits for IL-2 or IL-15, namely IL-2Rα (CD25) or IL-15Rα (CD215), as a third subunit. The downstream IL-2 and IL-15 heterotrimeric receptors share JAK1 (Janus kinase 1), JAK3, and STAT3 / 5 (signal transducer and activator of transcription 3 and 5) molecules for intracellular signaling, resulting in similar functions, although both cytokines also have distinct roles, as reviewed in Waldmann (2015, see e.g. Table 1) and Conlon (2019).

[0004] Therefore, activation of different heterotrimeric receptors by binding of IL-2, IL-15 or their derivatives may result in differential regulation of the immune system and potential side effects. Recently, NK cells and CD8 + Novel compounds containing IL-15 or IL-15 variants have been designed to specifically target T cell activation. These target the intermediate affinity IL-2 / IL-15Rβγ, i.e., NK cells, CD8 + IL-2 / IL-15Rβ and γ, expressed on T cells, NKT cells, and γδ T cells cThese compounds target a receptor composed of IL-15 subunits. This is important for safe and potent immune stimulation mediated by IL-15 transpresentation, but the designed compounds SO-C101 (RLI-15), ALT-803 and hetIL-15 already contain (part of) the IL-15Rα subunit and therefore mimic the transpresentation of the α subunit by antigen-presenting cells. SO-C101 contains a covalently linked sushi+ domain of IL-15Rα and therefore binds only to the intermediate affinity IL-15Rβγ. For this reason, SO-C101 does not bind to either IL-15Rα or IL-2Rα. Similarly, ALT-803 and hetIL-15 possess the IL-15Rα sushi domain or soluble IL-15Rα, respectively, and therefore bind to the intermediate affinity IL-15Rβγ receptor. Therefore, IL-15 and IL-15 analogs / supergonists are promising clinical stage development candidates for the treatment of cancer and infectious diseases (infections).

[0005] However, IL-15 and IL-15 superagonists are known to accumulate heterogeneity during expression, purification, storage and delivery, which can potentially adversely affect their pharmaceutical efficacy. Examples of such heterogeneity are different levels of glycosylation, deamidation of asparagine or glutamine, or oxidation of histidine, methionine, cysteine, tryptophan or tyrosine, in which the amide nitrogen group is exchanged with oxygen, thereby changing the polar amide to a negatively charged carboxylic acid. Such changes induce heterogeneity of the drug product, exposing the risk of increased immunogenicity, i.e., the generation of anti-drug antibodies that limit the pharmaceutical effect of the drug. Thus, there is a continuing need to provide variants of IL-15 and IL-15 superagonists that have reduced heterogeneity, but which essentially retain their activity and are expressed at similar levels. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Waldmann, TA et al. (2020). Frontiers in Immunology 11(868) [Non-Patent Document 2] Robinson, T.O. and K.S. Schluns (2017). Immunol Lett 190:159-168 [Non-Patent Document 3] Waldmann,TA (2015). Cancer Immunol Res 3(3):219-227 [Non-Patent Document 4] Conlon, KC et al., (2019). J Interferon Cytokine Res 39(1):6-21 Summary of the Invention [Means for solving the problem]

[0007] The inventors have surprisingly identified an IL-15 variant with a specific combination of amino acid substitutions that significantly reduces deamidation of asparagine 77 (N77) and glycosylation of IL-15, which surprisingly shows similar activity, similar expression levels and a longer half-life in vivo in fusion proteins with interleukin-15 receptor alpha compared to mature human IL-15.

[0008] Although a reduction or change in glycosylation pattern can affect the activity of a protein in vitro and in vivo, and although glycosylation of a protein has generally been described to increase the half-life in vivo and prevent physical instability of the protein (Sola and Griebenow 2009), the inventors have surprisingly found that the combination of substitutions at G78 and N79 brings about some unexpected advantages. Specifically, the mutation of these two sites leads to reduced deamidation, reduced glycosylation and increased homogeneity of the IL-15 variant. At the same time, the substitution of these amino acid residues does not affect the potency of IL-15 activity, does not affect stability parameters other than deamidation, and even has an increased in vivo half-life in fusion proteins with interleukin-15 receptor alpha, compared to mature IL-15. The latter is particularly advantageous. Increasing the in vivo half-life of IL-15 or IL-15 / IL-15Rα superagonists is generally seen as beneficial since their half-life is very short, and researchers have adopted various principles to increase the in vivo half-life, such as forming a complex with soluble IL-15Rα (WO 2007 / 001677), coupling an IL-15 / IL-15Rα sushi conjugate to the Fc fragment (WO 2008 / 143794 A1), or PEGylating IL-15 (WO 2015 / 153753 A2). In addition, the inventors found that such combined substitutions lead to increased stability of IL-15 during the manufacturing process, while wild-type IL-15 is degraded under such conditions.

[0009] Thus, the present invention provides, inter alia, IL-15 variants and conjugates comprising such IL-15 variants, which can be used for the treatment of new oncology indications and patient populations.

[0010] Definitions, Abbreviations and Acronyms "Interleukin-15", "IL-15" or "IL15" refers to the human cytokine described by NCBI reference sequence NP_000576.1 or UniProt ID P40933 (SEQ ID NO: 1). Its precursor protein has 162 amino acids and has a long 48-aa peptide leader resulting in a 114-aa mature protein (SEQ ID NO: 2), however the mature refers to the IL-15 protein lacking the 48 amino acid signal peptide of SEQ ID NO: 1. The complete coding sequence of its mRNA is described in NCBI GenBank reference U14407.1.

[0011] "IL-15 variant" or "variant of IL-15" refers to a protein having a percentage of identity of at least 92%, preferably at least 96%, more preferably at least 98%, and most preferably at least 99% with the amino acid sequence of mature human IL-15 (114aa) (SEQ ID NO: 2). Preferably, the IL-15 variant has at least 10%, more preferably at least 25%, even more preferably at least 50%, and most preferably at least 80% of the activity of IL-15. More preferably, the IL-15 variant has at least 0.1%, preferably 1%, more preferably at least 10%, more preferably at least 25%, even more preferably at least 50%, and most preferably at least 80% of the activity of human IL-15. Interleukins are extremely potent molecules that act at very low concentrations, and even an activity as low as 0.1% of human IL-15 may still be sufficiently potent, especially if administered at a higher concentration or if an extended half-life compensates for the loss of activity.

[0012] The activity of IL-15 can be measured by induction of proliferation of kit225 cells as described by Hori et al. (1987). Preferably, methods such as colorimetry or fluorescence are used to measure proliferation activation by IL-2 or IL-15 stimulation, for example as described by Soman et al. (Soman et al., 2009) using CTLL-2 cells. As an alternative to cell lines such as kit225 cells, human peripheral blood mononuclear cells (PBMCs) or buffy coats can be used. A preferred bioassay for measuring the activity of IL-15 is the IL-2 / IL-15 bioassay kit (Promega catalog number CS2018B03 / B07 / B05) using STAT5-RE CTLL-2 cells.

[0013] IL-15 mutant proteins (muteins) can be produced by standard genetic engineering methods and are well known in the art, for example from WO 2005 / 085282, US 2006 / 0057680, WO 2008 / 143794, WO 2009 / 135031, WO 2014 / 207173, WO 2016 / 142314, WO 2016 / 060996, WO 2017 / 046200, WO 2018 / 071918, WO 2018 / 071919, US 2018 / 0118805. IL-15 variants may further be generated by chemical modifications known in the art, such as by PEGylation or other post-translational modifications (see WO 2017 / 112528 A2, WO 2009 / 135031 A1).

[0014] "IL-15Rα" refers to the human IL-15 receptor alpha or CD215, described by NCBI reference sequence AAI21142.1 or UniProt ID Q13261 (SEQ ID NO: 4). Its precursor protein has 267 amino acids, with a 30-aa peptide leader, resulting in a 231-aa mature protein. Its mRNA is described in NCBI GenBank reference number HQ401283.1. IL-15Rα sushi domain (or IL-15Rα sushi , SEQ ID NO:5) is the domain of IL-15Rα that is essential for binding to IL-15 (Wei et al., 2001). The sushi+ fragment (SEQ ID NO:6) contains the sushi domain and a portion of the hinge region defined as the 14 amino acids that follow the IL-15Rα sushi domain and are C-terminal to the sushi domain, i.e., the IL-15Rα hinge region starts with the first amino acid after the (C4) cysteine ​​residue and ends with the 14th amino acid (counting in the standard "N-terminal to C-terminal" orientation). The sushi+ fragment reconstitutes full binding activity to IL-15 (WO 2007 / 046006).

[0015] An "IL-15Rα derivative" refers to a polypeptide comprising an amino acid sequence having a percentage of identity of at least 92%, preferably at least 96%, more preferably at least 98%, even more preferably at least 99%, and most preferably 100% identical to the amino acid sequence of the sushi domain of human IL-15Rα (SEQ ID NO:5), preferably the amino acid sequence of the sushi+ domain of human IL-15Rα (SEQ ID NO:6). Preferably, an IL-15Rα derivative is an N-terminally and C-terminally truncated polypeptide, but lacking the signal peptide (amino acids 1-30 of SEQ ID NO:4) and the transmembrane domain and intracytoplasmic portion of IL-15Rα (amino acids 210-267 of SEQ ID NO:4). Thus, a preferred IL-15Rα derivative comprises at least the sushi domain (aa 33-93), but does not extend beyond the extracellular portion of mature IL-15Rα, which is amino acids 31-209 of SEQ ID NO:4. Particularly preferred IL-15Rα derivatives are the sushi domain of IL-15Rα (SEQ ID NO:5), the sushi+ domain of IL-15Rα (SEQ ID NO:6) and soluble forms of IL-15Rα (amino acid 31 to any of amino acids 172, 197, 198, 199, 200, 201, 202, 203, 204 or 205 of SEQ ID NO:4; see WO 2014 / 066527 (Giron-Michel et al., 2005)). Within the limits provided by this definition, IL-15Rα derivatives may include naturally occurring or introduced mutations. Natural variants and alternative sequences are described, for example, in UniProtKB entry Q13261 (https: / / www.uniprot.org / uniprot / Q13261). Furthermore, one skilled in the art can easily identify amino acids that are less conserved among mammalian IL-15Rα homologs or even primate IL-15Rα homologs in order to generate derivatives that are still functional. The respective sequences of mammalian IL-15Rα homologs are set out in WO 2007 / 046006, pages 18 and 19. Additionally or alternatively, one skilled in the art can easily make conservative amino acid substitutions.

[0016] Preferably, the IL-15Rα derivative has at least 10%, more preferably at least 25%, even more preferably at least 50%, and most preferably at least 80% of the binding activity of the human sushi domain for human IL-15, e.g., as measured in (Wei et al., 2001).

[0017] "IL-2Rγ" refers to the common cytokine receptor gamma or gamma shared by IL-4, IL-7, IL-9, IL-15, and IL-21. c Or CD132.

[0018] "RLI-15" or "RLI" refers to any IL-15 / IL-15Rα conjugate that is a receptor-linker-interleukin fusion protein of the human IL-15Rα sushi+ fragment and human IL-15. Suitable linkers are described in WO 2007 / 046006 and WO 2012 / 175222.

[0019] "RLI2" or "SO-C101" refers to a specific version of RLI-15, an IL-15 / IL-15Rα conjugate that is a receptor-linker-interleukin fusion protein of human IL-15Rα sushi+ fragment and human IL-15 using a linker having SEQ ID NO:7 (SEQ ID NO:8).

[0020] A conjugate as used herein relates to either a non-covalent or covalent complex of interleukin 15 (IL-15) or a derivative thereof with the sushi domain of interleukin 15 receptor alpha (IL-15Rα) or a derivative thereof. The non-covalent complex may be formed by co-expression of the two polypeptides or by separate expression, (partial) purification and subsequent combination of such polypeptides to form such a complex by affinity. Preferably, the conjugate is a fusion polypeptide or protein, in which at least two polypeptides are genetically fused and recombinantly expressed to result in a single polypeptide chain to form an intact complex.

[0021] According to the present invention, a fusion polypeptide or fusion protein comprises a conjugate having at least one fusion polypeptide non-covalently or preferably covalently linked to another polypeptide chain, for example an immunocytokine comprising an antibody (having two heavy chains and two light chains covalently linked via disulfide bonds) fused to IL-15 or a variant thereof, or an IL-15 / sushi domain fusion protein or an Fc domain of an antibody having a polypeptide chain comprising two CH2 / CH3 each fused to a sushi domain, each complexed with an IL-15 variant, or having a polypeptide comprising one CH2 / CH3 fused to a sushi domain and the other fused to IL-15.

[0022] An immunocytokine as used herein relates to a polypeptide comprising an antibody or a functional variant thereof genetically fused to a conjugate according to the invention.

[0023] "ALT-803" refers to Altor BioScience Corp.'s IL-15 / IL-15Rα conjugate, which contains two molecules of human IL-15 "supergonist" with an optimized amino acid substitution (N72D) and two molecules of the human IL-15α receptor "sushi" domain, which confer stability and potentiate the IL-15 receptor activity. N72D :IL-15Rα sushi -conjugates fused to a dimeric human IgG1 Fc, which extends the half-life of the Fc conjugate (see, e.g., US Patent Application Publication No. 2017 / 0088597).

[0024] "P-22339" refers to Hengrui Medicine's IL-15 / IL-15Rα conjugate, which is a fusion protein containing two fusions of IL-15 with the sushi domain of IL-15Rα via engineered disulfide bonds fused to the N-terminus of the Fc fragment.

[0025] "XmAb24306" refers to Xencor's IL-15 / IL-15Rα conjugate, which is a fusion protein in which the sushi domain of IL-15Rα and IL-15 are fused to the N-terminus of an Fc fragment.

[0026] "CUG105" refers to Cugene's IL-15 / IL-15Rα conjugate, which is a fusion protein in which the sushi domain of IL-15Rα and IL-15 are fused to the N-terminus of an Fc fragment.

[0027] "Heterodimeric IL-15:IL-Rα", "hetIL-15" or "NIZ985" refers to Novartis' IL-15 / IL-15Rα conjugate similar to IL-15, which circulates as a stable molecular conjugate with human IL-15 and soluble human IL-15Rα (sIL-15Rα), a recombinantly co-expressed non-covalent conjugate of the 170 amino acids of IL-15Rα without the signal peptide and the transmembrane and cytoplasmic domains (Thaysen-Andersen et al., 2016, see e.g., Table 1).

[0028] "IL-2 / IL-15Rβγ agonists" are those that act by inhibiting the T-cell proliferation and / or proliferation of the IL-2Rα and / or IL-15Rα receptors without binding to them. reg This refers to molecules or conjugates that primarily target the medium affinity IL-2 / IL-15Rβγ receptor, lacking stimulation of the IL-2 / IL-15Rβγ receptor. An example is IL-15 bound to at least the sushi domain of IL-15Rα, which has the advantage of not being dependent on transpresentation or cell-cell interactions, and of having a longer in vivo half-life due to the increased size of the molecule, which has been shown to be significantly more potent than native IL-15 in vitro and in vivo (Robinson and Schluns, 2017). This is in addition to IL-15 / IL-15Rα-based conjugates, which also target IL-2 / 15Rβ and γ receptors. c This can be achieved by mutated or chemically modified IL-2, which significantly reduces or timely delays binding to the IL-2α receptor without affecting binding to the receptor.

[0029] "NKTR-255" refers to a PEG-conjugated human IL-15-based IL-2 / IL-15Rβγ agonist (WO 2018 / 213341 A1) that retains binding affinity for IL-15Rα, exhibits reduced clearance, and provides a sustained pharmacodynamic response.

[0030] "THOR-924, -908, -918" refer to PEG-conjugated IL-15-based IL-2 / IL-15Rβγ agonists (WO 2019 / 165453 A1) that have unnatural amino acids used for site-specific PEGylation and reduced binding to IL-15Rα.

[0031] "AM0015" refers to a PEG-conjugated IL-15 mutein (WO 2017 / 112528).

[0032] "Percentage of identity" between two amino acid sequences, "percentage of identity" refers to the percentage of identical amino acids between the two sequences compared, obtained using the best alignment of those sequences, where this percentage is purely statistical, and the differences between these two sequences are spread randomly across the amino acid sequences. As used herein, "best alignment" or "optimal alignment" refers to the alignment with the highest determined percentage of identity (see below). Sequence comparison between two amino acid sequences is usually achieved by comparing these sequences that have been pre-aligned according to the best alignment. This comparison is performed on a comparison segment to identify and compare local regions of similarity. In addition to manual methods, the best sequence alignment for the comparison can be achieved by using the global homology algorithm developed by Smith and Waterman (1981), by using the local homology algorithm developed by Needleman and Wunsch (1970), by using the similarity method developed by Pearson and Lipman (1988), by using computer software that uses such algorithms (GAP, BESTFIT, BLAST P, BLAST N, FASTA, TFASTA in the Wisconsin Genetics software package, Genetics Computer Group, 575 Science Dr., Madison, WI, USA), by using the MUSCLE multiple alignment algorithm (Edgar 2004), or by using CLUSTAL (Goujon et al., 2010). To obtain the best local alignment, the BLAST software can be used, preferably with the BLOSUM62 matrix.The percentage of identity between two amino acid sequences is determined by comparing the two optimally aligned sequences, which may also include additions or deletions relative to the reference sequence in order to obtain optimal alignment between the two sequences. The percentage of identity is calculated by determining the number of identical positions between the two sequences, dividing this number by the total number of positions compared, and multiplying the result by 100 to obtain the percentage of identity between the two sequences.

[0033] A "conservative amino acid substitution" refers to an amino acid substitution in which an aliphatic amino acid (i.e., glycine, alanine, valine, leucine, isoleucine) is replaced by another aliphatic amino acid, a hydroxyl or sulfur / selenium containing amino acid (i.e., serine, cysteine, selenocysteine, threonine, methionine) is replaced by another hydroxyl or sulfur / selenium containing amino acid, an aromatic amino acid (i.e., phenylalanine, tyrosine, tryptophan) is replaced by another aromatic amino acid, a basic amino acid (i.e., histidine, lysine, arginine) is replaced by another basic amino acid, or an acidic amino acid or its amide (aspartic acid, glutamic acid, asparagine, glutamine) is replaced by another acidic amino acid or its amide.

[0034] "Antibodies", also known as immunoglobulins (Ig), are large Y-shaped proteins that, in humans and most mammals, consist of two heavy chains (HC) and two light chains (LC) linked by disulfide bonds. The light chains contain one variable domain, V L and one constant domain C L The heavy chain consists of one variable domain, V H and three constant domains C H 1. C H 2. C H Structurally, antibodies each contain one V L , V H , C L , and C HTwo antigen-binding fragments (Fab) containing one domain, as well as two C domains of two heavy chains H 2 and C H It is also divided into an Fc fragment or domain comprising 3.

[0035] "Antibody variants" or "antibody functional variants" as used herein refer to antibodies with modifications, for example, to modulate the effector functions of the antibody, to modulate the antibody stability and in vivo half-life, and / or to induce heterodimerization of the antibody Fc domain. Such variants may be achieved by mutation and / or post-translational modifications. Antibody variants also include antibody heavy chains with shortening (truncating) of the N-terminal lysine on one or preferably both heavy chains. Other included variations are N- or C-terminal tags of the heavy and / or light chains for chemical or enzymatic coupling to other moieties, such as dyes, radionuclides, toxins or other binding moieties. Furthermore, antibody variants may include chemical modifications, modifications of their glycosylation, or substitutions with artificial amino acids for chemical linking to other moieties. Antibody variants as used herein also refer to immunoglobulin gamma (IgG)-based bispecific antibodies that potentially recognize two or more different epitopes. Various formats of bispecific antibodies are known in the art and are reviewed, for example, by Godar et al. (2018) and Spiess et al. (2015). The bispecific formats of the present invention include an Fc domain. For the immunocytokines of the present invention, the two RLI conjugates may be fused to either the C-terminus of both light chains or the C-terminus of both heavy chains, if not otherwise linked to a moiety, or one RLI conjugate may be fused to the C-terminus of one heavy chain for a heterodimeric bispecific format, or to the heavy chain or one light chain of a heterodimeric bispecific format with different light chains. An antibody functional variant can bind to the same epitope or target as the corresponding unmodified antibody.

[0036] "In vivo half-life", T 1 / 2or terminal half-life refers to the elimination half-life or terminal half-life, i.e., after administration, the in vivo half-life is the time required for the plasma / blood concentration to decrease by 50% after reaching pseudo-equilibrium of distribution (Toutain and Bousquet-Melou, 2004). Measurement of the drug in blood / plasma, here the polypeptide IL-2 / IL-15βγ agonist, is typically performed by a polypeptide-specific ELISA.

[0037] "Immune checkpoint inhibitors" or "checkpoint inhibitors" for short refer to a class of drugs that block certain proteins (immune checkpoint proteins) made by some types of immune system cells, such as T cells, and some cancer cells. These proteins are important in maintaining peripheral tolerance and preventing excessive immune responses. In malignant diseases, these proteins can be used by tumor cells to prevent T cells from killing cancer cells. When these proteins are blocked by checkpoint inhibitors, the "brakes" on the immune system are released and T cells can kill cancer cells again. Thus, checkpoint inhibitors are antagonists of immune inhibitory checkpoint molecules or antagonists of agonist ligands of inhibitory checkpoint molecules. Examples of checkpoint proteins found on T cells or cancer cells include PD-1 / PD-L1 and CTLA-4 / B7-1 / B7-2 (see National Cancer Institute of the National Institutes of Health definition, https: / / www.cancer.gov / publications / dictionaries / cancer-terms / def / immune-checkpoint-inhibitor), as reviewed, for example, by Darvin et al. (2018). Examples of checkpoint inhibitors are anti-PD-L1 antibodies, anti-PD-1 antibodies, anti-CTLA-4 antibodies, but also antibodies against LAG-3 or TIM-3, or blockers of BTLA currently being tested in the clinic (De Sousa Linhares et al., 2018). Further promising checkpoint inhibitors are anti-TIGIT antibodies (Solomon and Garrido-Laguna, 2018).

[0038] "Anti-PD-L1 antibody" refers to an antibody or antibody fragment thereof that binds to PD-L1. Examples are avelumab, atezolizumab, durvalumab, KN035, MGD013 (bispecific for PD-1 and LAG-3).

[0039] "Anti-PD-1 antibody" refers to an antibody or antibody fragment thereof that binds to PD-1. Examples are pembrolizumab, nivolumab, cemiplimab (REGN2810), BMS-936558, SHR1210, IBI308, PDR001, BGB-A317, BCD-100, and JS001.

[0040] "Anti-PD-L2 antibody" refers to an antibody or antibody fragment thereof that binds to anti-PD-L2. An example is sHIgM12.

[0041] "Anti-CTLA4 antibody" refers to an antibody or antibody fragment thereof that binds to CTLA-4. Examples are ipilimumab and tremelimumab (ticilimumab).

[0042] "Anti-LAG-3" antibody refers to an antibody or antibody fragment thereof that binds to LAG-3. Examples of anti-LAG-3 antibodies are leratolimab (BMS986016), Sym022, REGN3767, TSR-033, GSK2831781, MGD013 (bispecific for PD-1 and LAG-3), and LAG525 (IMP701).

[0043] "Anti-TIM-3 antibody" refers to an antibody or antibody fragment thereof that binds to TIM-3. Examples are TSR-022 and Sym023.

[0044] "Anti-TIGIT antibody" refers to an antibody or antibody fragment thereof that binds to TIGIT. Examples are tiragolumab (MTIG7192A, RG6058) and etigilimab (WO 2018 / 102536).

[0045] "Therapeutic antibody" or "tumor-targeting antibody" refers to an antibody or antibody fragment thereof that has a direct therapeutic effect on tumor cells through the binding of the antibody to a target expressed on the surface of the tumor cells to be treated. Such therapeutic activity may result from receptor binding leading to alteration of signal transduction in the cells, direct cell death induction, antibody-dependent cellular cytotoxicity (ADCC), complement-dependent cytotoxicity (CDC) or other antibody-mediated killing of tumor cells.

[0046] "Anti-CD38 antibody" refers to an antibody or antibody fragment thereof that binds to CD38, also known as cyclic ADP-ribose hydrolase. Examples of anti-CD38 antibodies are daratumumab, isatuximab (SAR650984), MOR-202 (MOR03087), TAK-573 or TAK-079 (Abramson, 2018) or GEN1029 (HexaBody®-DR5 / DR5).

[0047] When described as "administered in combination," this typically does not mean that two agents are co-formulated and administered simultaneously, but rather that one agent has a label that specifies its use in combination with the other agent. So, for example, an IL-2 / IL-15Rβγ agonist is for use in the treatment or management of cancer, which use includes administering the IL-2 / IL-15Rβγ agonist and an additional therapeutic agent simultaneously, separately, or sequentially, or vice versa. However, nothing in this application should preclude the two combined agents from being provided as a package or kit, or even from being co-formulated and administered together when the dosing schedules match. So, "administered in combination" includes (i) the agents are administered together, such as by intra-articular infusion, intra-articular injection, etc.; (ii) the agents are administered separately but in parallel according to the given method of administration of each agent; and (iii) the agents are administered separately and sequentially. Concurrent administration in this context preferably means that both treatments are started together, for example, the first dose of each drug in the treatment regimen is administered on the same day.Considering the potential different treatment schedules, it is clear that administration does not always have to occur on the same day in the course of the following days / weeks / months.Generally, concurrent administration aims to have both drugs present in the body at the same time at the beginning of each treatment cycle. Sequential administration in this context preferably means that both treatments are initiated sequentially, e.g., the first administration of the first drug occurs at least one day, preferably several days or a week, before the first administration of the second drug to allow for the body's pharmacodynamic response to the first drug before the second drug becomes active. Thereafter, the treatment schedules may overlap or be intermittent with each other, or may follow directly after each other.

[0048] The term "resistant to checkpoint inhibitor therapy" refers to a patient who never exhibits a therapeutic response when receiving a checkpoint inhibitor.

[0049] The term "refractory to checkpoint inhibitor therapy" refers to patients who initially demonstrated a therapeutic response to checkpoint inhibitor therapy, but that therapeutic response was not sustained over time.

[0050] The term "about" when used in conjunction with a value means ±10% of that value, preferably ±5% and especially ±1% of that value.

[0051] Where the term "comprising" is used in the present description and claims, it does not exclude other elements. For the purposes of the present invention, the term "consisting of" is considered to be a preferred embodiment of the term "comprising of". In the following, when a group is defined as comprising at least a certain number of embodiments, this is to be understood as also disclosing a group that preferably consists only of these embodiments.

[0052] Where an indefinite or definite article such as "a", "an" or "the" is used when referring to a singular noun, this includes a plural of that noun, unless something else is specifically stated.

[0053] Thus, the terms "at least one," "at least one," as in "at least one chemotherapeutic agent," may be meant to mean one or more chemotherapeutic agents. The term "a combination thereof," in the same context, refers to a combination including more than one chemotherapeutic agent.

[0054] Technical terms are used with their common meaning. Where a specific meaning is conveyed to a particular term, the definition of the term is given below in the context in which the term is used.

[0055] "qxw", derived from the Latin quaque / each, every, means every x weeks, for example q2w means every two weeks. "sc" or "SC" stands for subcutaneous. "iv" or "IV" stands for intravenous. "ip" or "IP" stands for intraperitoneal. C max represents the maximum concentration. AUC stands for area under the curve.

[0056] [Table 1(1)] [Table 1(2)] [Table 1(3)] [Table 1(4)] [Table 1(5)] [Table 1(6)] [Table 1(7)] [Brief description of the drawings]

[0057] [Figure 1A] LMW SDS-PAGE and Western blot (anti-RLI-15) analysis of RLI2 (RLI2 wt), RLI2 with a G78A substitution (RLI2 A) and RLI2 with a G78A / N79Q substitution (RLI2 AQ) under non-reducing conditions. For Coomassie staining, 0.5 or 2 μg of protein were used (lanes 2, 4, 6, 8, 10 and 12) and for Western blotting, 25 ng of protein was used (lanes 3, 7, 11). [Figure 1B]Capillary electrophoresis, denaturation and analysis of RLI2 (RLI2 wt), RLI2 with G78A substitution (RLI2 A) and RLI2 with G78A / N79Q substitutions (RLI2 AQ) under reducing (R) and non-reducing (NR) conditions. Dashed box 1 represents the band at glycosylation site number 2 (major), box 2 represents the band at glycosylation site number 1 (minor), and dashed box 3 represents the de novo glycosylation site in RLI2A. Unnamed lanes are markers for 16, 21, 30, 48 and 68 kDa. [Diagram 2] Analysis of three deglycosylated RLI variants expressed in CHO cells by SDS-PAGE (7.5–18%) stained with Coomassie blue (left pane), silver nitrate (middle pane) and detected by anti-IL15 Western blot (right pane): lane 1: molecular weight marker; lane 2: RLI2N176Q, lane 3: RLI2N168S / N176Q / N209S, lane 4: RLI1N168S / N176Q / N209S. [Diagram 3] Potency of RLI2 and RLI2AQ from supernatants determined by activation of 32Db or kit225 cells: (A) 32Db cells, 21 hours (h), (B) kit225 cells, 4 hours. [Figure 4] Relative potency of purified RLI2 or RLI2 from supernatant compared to RLI2AQ from supernatant as determined by activation of kit225 cells. [Figure 5A] Comparison of hyperglycosylated and hypoglycosylated RLI2. CPI HIC elution profiles depending on the concentration of buffer B measured at 280 nm. The left box shows pooled fractions 2B1 1-3 for hyperglycosylated RLI2 ("RLI-15-HG") and the right box shows pooled fractions 4B1 1-3 for hypoglycosylated RLI2 ("RLI-15-LG"). [Figure 5B] Comparison of hyperglycosylated and hypoglycosylated RLI2. SDS PAGE of fractions 2B1 1-3 of RLI-15-HG, RLI2 reference standard and molecular weight ladder at the given kDa. [Figure 5C]Comparison of hyperglycosylated and hypoglycosylated RLI2. SDS PAGE of fractions 4B1 1-3 of RLI-15-LG, RLI2 reference standard and molecular weight ladder at the given kDa. [Figure 6] In vitro mixed lymphocyte reaction (hPBMC donor): Relative IFNγ production is shown for PEM (pembrolizumab) and RLI-15 (RLI2) compared to the immunocytokine PEM LY-RLI NA x1 (IL-15 N65A mutant also carrying the AQ mutation). [Figure 7] In vivo hPD1 single KI HuGEMM mice implanted with HuCell MC38-hPD-L1 tumor cell line were used as animal tumor models. Tumor volumes are shown for control (triangles), pembrolizumab administered at 5 mg / kg on days 0 (D0), 3 (D3), 6 (D6) and 9 (D9) (grey circles), and PEM-RLI NA x1 administered at 20 mg / kg on D0 (black circles). [Figure 8] Comparison of ADCC activity of immunocytokines based on hCl1a antibodies with unmodified effector function to immunocytokines and antibodies hCl1a and zolbetuximab with reduced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). [Figure 9A] Comparison of ADCC activity of immunocytokines based on hCl1a antibody with unmodified effector function to immunocytokines and antibodies hCl1a and zolbetuximab with enhanced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). (A) DLE mutations. [Figure 9B]Comparison of ADCC activity of immunocytokines based on hCl1a antibodies with unmodified effector function to immunocytokines and antibodies hCl1a and zolbetuximab with enhanced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). (B) DE mutations. [Figure 9C] Comparison of ADCC activity of immunocytokines based on hCl1a antibodies with unmodified effector function to immunocytokines and antibodies hCl1a and zolbetuximab with enhanced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). (C) AAA mutation. [Figure 9D] Comparison of ADCC activity of immunocytokines based on hCl1a antibodies with unmodified effector functions to immunocytokines and antibodies hCl1a and zolbetuximab with enhanced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). (D) TL mutations. [Figure 9E] Comparison of ADCC activity of immunocytokines based on hCl1a antibodies with unmodified effector function to immunocytokines and antibodies hCl1a and zolbetuximab with enhanced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). (E) IE mutation. [Figure 9F]Comparison of ADCC activity of immunocytokines based on hCl1a antibody with unmodified effector function to immunocytokines and antibodies hCl1a and zolbetuximab with enhanced ADCC activity. ADCC target cells were A549 cells overexpressing CLDN18.2 (A549-CLDN18.2) or PA-TU-8988S cells endogenously expressing CLDN18.2 (PATU). (F) Afucosylated immunocytokines. [Figure 10A] The % of PD-1 / PD-L1 blockade is shown as a function of increasing pM concentrations of Keytruda and SOT201. [Figure 10B] Percentage of Ki67+ NK cells and CD8+ T cells determined by flow cytometry after 7 days of in vitro stimulation of human PBMCs from healthy donors with increasing amounts of SOT201 or SOT201 wt carrying an IL-15 moiety without reducing binding to IL-2 / IL-15Rβγ. [Figure 10C] Cell proliferation (Ki67+) of CD8+ T cells or NK cells detected by flow cytometry in the spleens of healthy C57BL / 6 mice (n=2 / group) 5 days after IV injection of equimolar amounts of compound to 5 mg / kg of the murine surrogate molecule mSOT201 (anti-mouse PD-1 antibody RMP1-14 fused RLI-15AQA) compared to anti-mouse PD-1 antibody alone or anti-human PD1 mouse IgG1-RLI-15AQA (hPD1-mSOT201) as single activity controls. [Figure 11A] Tumor volume in mm3 over a 17-day time course in C57BL / 6 mice bearing syngeneic MC38 tumor cells treated IV on day 1 (randomization day with tumor volume between 80 and 100 mm3) with a single injection of control (NaCl), mSOT201, mSOT201 (5 mg / kg) and an equimolar amount of hPD1-mSOT201 or mPD1 (n=10 mice / group). [Figure 11B] Corresponding percentage of MC38 tumor-bearing mice surviving up to 100 days after treatment. [Figure 12A(1)]Relative expression levels of indicated adaptive and innate immune cell and cancer-associated fibroblast (CAF)-related gene sets across mSOT201-treated tumor samples (N=3) and control samples (n=4) of MC38 tumor-bearing mice as determined by metagene on RNA-seq data. Box plots: minimum, median, maximum. [Figure 12A(2)] Continued from Figure 12A(1). [Figure 12A(3)] Continuation of Figures 12A(1) and 12A(2). [Figure 12B] Cell proliferation determined by % Ki67+ cells by flow cytometry of the indicated cells in the spleen or lymph nodes of MC38 tumor-bearing mice 7 days after mSOT201 (5 mg / kg) IV treatment of established tumors (80–100 mm3) (n=2). [Figure 13A] Tumor volume in mm3 over a 21-day time course in C57BL / 6 mice bearing MC38 tumors treated IV with a single injection of control (NaCl), mSOT201, mPD1-IL-2βγ agonist that abolishes CD25 binding (IL-2v fused to anti-mouse PD-1 antibody RMP1-14), or the combination of RLI-15AQA and anti-mouse PD-1 antibody mPD1 (RMP1-14). [Figure 13B] Cell proliferation as determined by % Ki67+ cells of CD8+ T cells and NK cells detected by flow cytometry on days 5 and 8 after IV administration in healthy C57 / BL6 mice. [Figure 13C] %Ki67+ cells of CD8+ T cells in spleen or lymph nodes of MC38 tumor-bearing C57BL / 6 mice treated IV with mSOT201, mPD1-IL-2v or a combination of RLI-15AQA and mPD-1 on day 7. Randomization day 1, tumor volume 100 mm3 (n=10 / group). [Figure 14A]Fold change in % and absolute cell counts of Ki67+ NK and CD8+ T cells in blood of cynomolgus monkeys following a single IV dose of 0.6 mg / kg SOT201 on day 1, as determined by flow cytometry and hematology on the days indicated. Each graph curve represents one animal. [Figure 14B] Percent Ki67+ of NK cells and CD8+ T cells in blood of cynomolgus monkeys following administration of 0.3 mg / kg IV SOT201 on days 1 and 21 (indicated by arrows) as determined by flow cytometry on the days indicated. Each graph curve represents one animal. [Figure 15A] Proliferation of NK cells and CD8+ T cells after treatment with murine SOT201 surrogates in vivo. Proliferation of CD8+ T cells and NK cells in the spleen of healthy C57BL / 6 mice on days 5 and 8 after treatment with hPD1-mSOT201, mPD-1, mSOT201, mSOT201 wt, and mPD1-IL2v. Expression of Ki67 in CD8+ T cells and NK cells was detected by flow cytometry. The molecules were administered iv on day 1 at 5.37 mg / kg hPD1-mSOT201, 4.51 mg / kg mPD-1, equimolar to 5 mg / kg mSOT201, and 0.26 mg / kg mPD1-IL2v, equimolar to 0.25 mg / kg mSOT201 wt. Flow cytometry analysis was performed on days 5 and 8. Data represent the mean±SEM of two individuals per group per day. [Figure 15B]Proliferation of NK cells and CD8+ T cells after treatment with murine SOT201 surrogates in vivo. Proliferation of CD8+ T cells and NK cells in the spleen of healthy C57BL / 6 mice on days 5 and 8 after treatment with hPD1-mSOT201, mPD-1, mSOT201, mSOT201 wt, and mPD1-IL2v. Expression of Ki67 in CD8+ T cells and NK cells was detected by flow cytometry. The molecules were administered iv on day 1 at 10.74 mg / kg hPD1-mSOT201, 9.02 mg / kg mPD-1, equimolar to 10 mg / kg mSOT201, and 0.1 mg / kg mPD1-IL2v, equimolar to 0.1 mg / kg mSOT201 wt. Flow cytometry analysis was performed on days 5 and 8. Data represent the mean±SEM of two individuals per group per day. [Figure 16]Mouse SOT201 surrogates in PD-1-sensitive and PD-1-resistant tumor models in vivo. (A) Anti-PD-1-sensitive tumor models. MC38 / C57BL / 6 mouse model: single iv administration of 4.51 mg / kg mPD-1 (selected as equimolar to mSOT201, suboptimal dose compared to literature), 5 mg / kg mSOT201 or 5.37 mg / kg hPD1-mSOT201 (equimolar to mSOT201) on day 0; DO = randomization day when tumor volume was approximately 80-100 mm3, 10 mice / group. CT26 / BALB / c mouse model: 9.02 mg / kg mPD-1 (effective dose compared to literature), 10 mg / kg mSOT201, 10.74 mg / kg hPD1-mSOT201 (equimolar amount to mSOT201) ip administered four times on days 0, 3, 6 and 9; DO = randomization day when tumor volume was approximately 100 mm3, 10 mice / group. (B) Anti-PD-1 resistant tumor model. CT26 STK11 ko mouse model: 4 doses of ip administration on days 0, 3, 6 and 9 with 9.02 mg / kg mPD-1 (effective dose compared to literature), 10 mg / kg mSOT201, 10.74 mg / kg hPD1-mSOT201 (equimolar amount to mSOT201); DO = randomization day when tumor volume is approximately 100 mm3, 10 mice / group. B16F10 / C57BL / 6 mouse model: 4 doses of ip administration on days 0, 3, 6 and 9 with 9.02 mg / kg mPD-1 (effective dose compared to literature), 10 mg / kg mSOT201, 10.74 mg / kg hPD1-mSOT201 (equimolar amount to mSOT201); DO = randomization day when tumor volume is approximately 100 mm3, 10 mice / group. Cut-off date for all mice present in the control group, CR = complete response. [Figure 17]Comparison of mSOT201 vs. RLI-15AQA mutein + anti-PD-1 in vivo. MC38 / C57BL / 6 mouse model with the following groups: G1 sham control G4: Single sc dose of 0.64 mg / kg RLI-15AQA on day 0 + single ip dose of 4.51 mg / kg mPD-1 on day 0. G2 single dose of mSOT201 5 mg / kg iv on day 0 G3 single dose of mSOT201 2 mg / kg iv on day 0 G6 single dose of mPD1 alone 4.51 mg / kg ip on day 0 (selected as equimolar to mSOT201, suboptimal dose compared to literature), G11 single dose of hPD1-mSOT201 5 mg / kg iv on day 0 + single dose of mPD-1 4.36 mg / kg ip on day 0, day 0 = randomization day with tumor volume ≈80-100 mm3; cutoff day for all mice present in control group with 10 mice / group, CR = complete response. [Figure 18] MC38 / C57BL / 6 mouse model - D0 = randomization day, approximately 80-100 mm3, 10 mice / group. CR = complete response G1 Sham control G2 A single dose of mSOT201 5 mg / kg iv on day 0 G3 A single dose of mSOT201 2 mg / kg iv on day 0 G7 4 doses of RLI2AQ 1 mg / kg sc on days 0, 1, 2 and 3 G5 A single dose of RLI2AQ 1 mg / kg sc on day 0 + a single dose of mPD1 5 mg / kg ip on day 0 G8 4 doses of RLI2AQ 1 mg / kg sc on days 0, 1, 2 and 3 + a single dose of mPD1 5 mg / kg ip on day 0 G9 4 doses of RLI2AQ 1 mg / kg sc on days 0, 1, 2 and 3 + 4 doses of mPD1 5 mg / kg ip on days 0, 3, 6 and 9 G6 A single dose of mPD1 5 mg / kg ip on day 0 G10 Four ip doses of 5 mg / kg mPD1 on days 0, 3, 6, and 9. Cut-off date for all mice in the control group. [Figure 19A]Comparison of mSOT201 vs. RLI2AQ + anti-PD-1 tumor growth in vivo. MC38 / C57BL / 6 mouse model. Mean tumor volumes in mm3 shown for individual animals depending on time and on day 16. Horizontal lines indicate mean tumor volumes. G1 mock control G2 single dose iv of mSOT201 2 mg / kg on day 0, G3 two sc doses of RLI2AQ 2 mg / kg on days 0 and 1 + four ip doses of mPD1 2 mg / kg on days 0, 3, 6 and 9. One experiment only. D0 = day of randomization with tumor volume approx. 80-100 mm3; 10 mice / group. CR = complete response. Relative proliferation of NK cells, CD8+ T cells and cells expressing αβTCR and γδTCR (T cells) was examined in spleen, lymph nodes and tumors on day 7 after SOT201 (G2 from above) and RLI2AQ + anti-PD-1 (G3 from above) treatment using flow cytometry. Three tumor samples were pooled and three spleen and lymph node samples were analyzed separately. [Figure 19B] Comparison of mSOT201 vs. RLI2AQ + anti-PD-1 tumor growth in vivo in MC38 / C57BL / 6 mouse model. Parental frequencies in % (relative percentage compared to the parental population) are shown for CD8+ T cells (top) and NK cells (bottom) from lymph nodes, spleen and tumors. [Figure 19C] Comparison of mSOT201 vs. RLI2AQ + anti-PD-1 tumor growth in vivo. MC38 / C57BL / 6 mouse model. Parental frequencies in % for αβTCR+ CD3+ T cells (top) and βγTCR+ CD3+ T cells (bottom) from lymph nodes, spleen and tumor are shown. [Figure 20](A) Immunogenicity in DC-T cell based assay. T cell response to PEM-RLI-15 candidate molecules shown as %CFSElow stained CD4+ T cells after loading the candidate molecules into iDCs, incubating with autologous CD4+ T cells pre-stained with CFSE, and detecting CFSE staining. CFSElow is used as a surrogate for cycling cells. Mean ± SEM of 11 donors is shown. Significant differences compared to control DCs incubated without protein, thus inducing non-specific T cell proliferation, are shown. *p≦0.05, ***p≦0.001. (B) FluoroSpot assay for IFN-γ and TNF-α of RLI-15 peptides across the introduced substitutions N65A and G175A / N176Q. Estimation with 95% confidence interval (CI) of the effect of Mut2 or Mut3 peptides versus the respective wild type peptide on the mean dSFU in a study population of 40 donors. SFU = spot forming units, dSFU = SFU of restimulated wells - SFU of non-restimulated wells. [Figure 21] Comparison of the ability of SOT202 molecules with altered effector functions to induce proliferation of hPBMCs. Proliferation of isolated hPBMCs was assessed for SOT202-DANA, SOT202-afuc-DANA, SOT202-DLE-DANA, SOT202-DE-DANA and SOT202-LALAPG-DANA. Cells were stimulated in vitro for 7 days. Means ± SEM of 6 donors are shown. Proliferation of NK cells (top) and CD8+ T cells (bottom) was measured by counting Ki67+ cells by flow cytometry. [Figure 22] Comparison of the ability of SOT202 molecules and SOT201 to induce proliferation of hPBMC. Proliferation of isolated hPBMC was assessed for SOT202, SOT202-afuc, SOT201-DANA, SOT202-DANA and SOT202-afuc-DANA. Proliferation of NK cells (top) and CD8+ T cells (bottom) was measured by counting Ki67+ cells by flow cytometry. [Figure 23]Comparison of the ability of SOT202-DANA molecules with altered effector functions and SOT201-DANA to induce proliferation of hPBMCs. Proliferation of isolated hPBMCs was assessed for SOT201-DANA, SOT202-DANA, SOT202-afuc-DANA, SOT202-LALAPG-DANA and hCl1a (also referred to as SOT202-mab). Proliferation of NK cells (top) and CD8+ T cells (bottom) was measured by counting Ki67+ cells by flow cytometry. [Figure 24A] Cell proliferation (Ki67+) of CD8+ T cells or NK cells detected in the spleen of healthy C57BL / 6 mice after stimulation with mSOT202. Cell proliferation was detected by Ki67 staining and measured by flow cytometry 5 days after IV injection of 5, 10 or 20 mg / kg of mSOT202 (hCl1a-mIgG2a-NA 1x) or hCl1a-mIgG2a compound. [Figure 24B] Percentages of NK cells and CD8+ T cells under the same experimental conditions as in (A). [Figure 25A] Cell proliferation of NK cells detected in the spleen of healthy C57BL / 6 mice after stimulation with mSOT202, mSOT202-LALAPG and hCl1a-mIgG2a. Top: Cell proliferation was detected by Ki67 staining and measured by flow cytometry 5 and 10 days after IV injection of 5 mg / kg of the compounds. Bottom: Percentage of NK cells. [Figure 25B] Cell proliferation of CD8+ T cells detected in the spleen of healthy C57BL / 6 mice after stimulation with mSOT202, mSOT202-LALAPG and hCl1a-mIgG2a. Top: Cell proliferation was detected by Ki67 staining and measured by flow cytometry 5 and 10 days after IV injection of the compounds at 5 mg / kg. Bottom: Percentage of CD8+ T cells.

[0058] array [ka] [ka] [ka] [ka] [ka] [ka] [ka] [ka] [ka] [ka] [ka] [ka] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0059] In a first aspect, the present invention relates to an Interleukin-15 (IL-15) variant comprising amino acid substitutions at positions G78 and N79 of mature human IL-15 (SEQ ID NO: 2). Preferably, the substituted amino acids are naturally occurring amino acids.

[0060] The inventors have successfully generated IL-15 variants with particularly high homogeneity and reduced glycosylation by substituting sites G87 and N79, while the potency and stability of the IL-15 variants were not affected. This was surprising, since glycosylation is the main source of microheterogeneity in proteins (glycoforms), which reflect the complexity at both molecular and cellular levels. There are many potential functions of glycosylation, such as protein folding, trafficking, packing, stabilization, protease protection, organization of quaternary structure or water structure. For example, changes in sugar motifs may reflect and result in physiological changes, for example, in cancer and rheumatoid arthritis. Therefore, those skilled in the art are hesitant to modify the glycosylation of therapeutic proteins, especially for pharmaceutical applications.

[0061] In one embodiment, the IL-15 variant comprises the amino acid substitutions G78A, G78V, G78L or G78I, and N79Q, N79H or N79M, preferably G78A and N79Q. The G78A / N79Q double substitution resulted in a superior IL-15 variant when tested in the context of the RLI2 fusion protein (where the respective numbering would be G175A / N176Q) with respect to homogeneity, stability and in vivo half-life.

[0062] Preferably, the IL-15 variant is expressed in a mammalian cell line, preferably the mammalian cell line is selected from CHO cells, HEK293 cells, COS cells, PER.C6 cells, SP20 cells, NSO cells or any cells derived therefrom, more preferably CHO cells. Although various eukaryotic or preferably mammalian expression systems can be used, expression in CHO cells is the most established expression system and gives good yields.

[0063] The amino acid substitutions in the IL-15 variants preferably reduce deamidation at N77 and glycosylation at N79 of the IL-15 variants compared to mature human IL-15 without such substitutions. More preferably, there is less than 30% glycosylated IL-15 variants, in particular less than 25% glycosylated IL-15 variants, as measured in RLI2 fusions. For comparison, RLI2 (without AQ substitutions) has a maximum of 40% glycosylation. In one embodiment, less than 30% of the IL-15 variants are glycosylated. In a further embodiment, less than 25% of the IL-15 variants are glycosylated. Preferably, N71 is more glycosylated compared to IL-15 without such substitutions (human mature IL-15). Thus, while the overall glycosylation of RLI2 AQ is reduced compared to RLI2, glycosylation on the secondary glycosylation site N71 (IL-15 numbering) / N168 (RLI numbering) appears to be increased to 20%, likely due to the proximity of the two glycosylation sites N168 and N176, resulting in interference with the predominant / preferential glycosylation of N176. This interference is lifted by the N176Q substitution, resulting in increased glycosylation at N168.

[0064] In a preferred embodiment, the amino acid substitutions of the IL-15 variants do not substantially reduce the IL-15 activity of the IL-15 variants on proliferation induction of kit225 cells, 32Db cells, human PBMCs or in the Promega IL-15-bioassay. Substantially in this context means that the activity is not reduced by more than 20%, preferably by more than 10%, compared to IL-15 without such substitutions. kit225 cells (Hori et al., 1987) are commonly used to measure the induction of proliferation by IL-15 and IL-15 superagonists. Preferably, methods such as colorimetry or fluorescence are used to measure proliferation activation by IL-2 or IL-15 stimulation, as described, for example, by Soman et al. (Soman et al., 2009) using CTLL-2 cells. As an alternative to cell lines such as kit225 cells, 32Db cells (ThermoFisher), human peripheral blood mononuclear cells (PBMCs) or buffy coats can be used. A preferred bioassay for determining IL-15 activity is the IL-2 / IL-15 bioassay kit (Promega catalog number CS2018B03 / B07 / B05) using STAT5-RE CTLL-2 cells.

[0065] In another embodiment, the IL-15 variant has no substitutions at position N71 and / or position N77. The inventors have found that substituting the next glycosylation site results in low expression and glycosylation at other sites. In addition, each additional mutation / substitution introduced increases the risk of immunogenicity, which should be avoided.

[0066] In a preferred embodiment, the IL-15 variant is IL-2 / IL-15Rβ and / or γ c Contains at least one additional substitution that reduces binding to the IL-2 / IL-15Rβ and / or IL-15Rγ receptors. cBased on the extremely high affinity of IL-15 for its receptor in terms of binding to the receptor, administered IL-15, and likewise IL-15 / IL-15Rα conjugates, exhibit a very short half-life, mainly due to target-mediated drug deposition (TMDD) (Hangasky et al., 2020), where the drug is bound by its target immune cells and thus consumed and cleared. Thus, a single iv injection would result in a high C max and immediate, very short half-life leading to rapid decline, resulting in a rather small AUC and therefore a suboptimal pharmacokinetic (PK) profile. However, potent immune cell proliferation requires repeated and / or longer IL-15 exposure above a certain threshold, i.e. a higher AUC. There are several methods used to achieve a more favorable PK profile, including (i) continuous iv infusion, although this is inconvenient; (ii) increasing the size of the molecule, for example by PEGylation (e.g. NKTR-255, THOR-924, AM0015), complexing the molecule to a portion of the IL-15Rα (RLI-15, hetIL-15, ALT-803, P-22339, XmAb24306 or CUG105), or complexing / fusing the molecule to the Fc portion of an antibody (ALT-803, P-22339, XmAb24306 or CUG105); (iii) sc administration, which results in some delayed absorption from the subcutaneous depot; and / or (iv) by decreasing the binding affinity of IL-15 to its receptor, thereby reducing TMDD. Such reduced binding of IL-15 to its receptor is accompanied by a reduced potency in activating its target immune cells in vitro (wherein TMDD does not play a major role, e.g., as measured in kit225 cells), but is compensated in vivo by its better PK profile due to its extended in vivo half-life (US Patent Application Publication No. 2018 / 0118805 A1) (Bernett et al., 2018).

[0067] IL-2 / IL-15Rβ or γ cSuitable amino acid substitutions that reduce binding to the receptor are preferably IL-2Rβ or γ c Located at the interface. IL-2 / IL-15Rβ and / or γ cSeveral sites for further substitutions that reduce binding to receptors have been described in the prior art.The amino acid substitutions may be one or more sites selected from the list consisting of N1, N4, S7, D8, K10, K11, D30, D61, E64, N65, L69, N72, E92, Q101, Q108, I111, preferably one or more sites selected from positions D61, N65 and Q101 (see WO 2005 / 085282, WO 2006 / 020849A2, WO 2008 / 143794A1, WO 2014 / 207173A1, US 2018 / 0118805A1) (Ring et al., 2012), especially N65. Specifically, the one or more substitutions are N1D, N1A, N1G, N4D, S7Y, S7A, D8A, D8N, K10A, K11A, D30N, D61A, D61N, E64Q, N65D, N65A, N65E, N65R, N65K, L69R, N72R, Q101D, Q101E, Q108D, Q108A, Q108E, Q108R and preferably selected from the list consisting of D8A, D8N, D61A, D61N, N65A, N65D, N72R, Q101D, Q101E and Q108A, more preferably selected from the substitutions D61A ("DA" mutation), N65A ("NA" mutation), Q101D ("QD" mutation), especially N65A. N65K and L69R have been reported to abrogate IL-2 / IL-15Rβ binding (WO 2014 / 207173 A1), whereas Q101D and Q108D inhibit the function of IL-15 (WO 2006 / 020849 A2) and are preferred substitutions. Q108D has been specifically described to increase affinity for CD122 and impair recruitment of CD132 to inhibit IL-2 and IL-15 effector function, whereas N65K has been described to abrogate CD122 affinity (WO 2017 / 046200A1).N1D, N4D, D8N, D30N, D61N, E64Q, N65D and Q108E were described to gradually reduce the activity of the respective IL-15 / IL-15Rα conjugates on NK cell and CD8 T cell activation (see Figure 51, WO 2018 / 071918A1, WO 2018 / 071919A1). S7Y, S7A, K10A, K11A have been identified to reduce IL-2 / IL-15Rβ binding (Ring et al., 2012). Preferred combinations are D8N / N65A, D61A / N65A ("DANA" mutations), N1D / D61N, N1D / E64Q, N4D / D61N, N4D / E64Q, D8N / D61N, D8N / E64Q, D61N / E64Q, E64Q / Q108E, D61A / N65A / Q101D ("DANAQD" mutations), N1D / N4D / D8N, D61N / E64Q / N6SD (the "NQD" mutations), N1D / D61N / E64Q, N1D / D61N / E64Q / Q108E, or N4D / D61N / E64Q / Q108E, more preferably D8N / N65A, D61A / N65A or D61A / N65A / Q101D, especially D61A / N65A.

[0068] A number of substitutions that reduce binding to IL-2 / IL-15Rβγ have been described in the prior art. However, adequate data on their effects on pharmacokinetics in mammals are lacking and are largely unpredictable. The present inventors have identified a suitable range of IL-15 variants with AQ mutations, which have additional single substitutions that significantly reduce potency when tested in fusion protein with sushi+ fragment of IL-15Rα (RLI2). As shown in Table 11, D61A substitution results in about 8-fold reduction, N65D substitution results in about 20-fold reduction, and N65A substitution results in 48-fold reduction.

[0069] Similarly, RLI2 fused to the C-terminus of one or both heavy chains or both light chains of an antibody AQAn immunocytokine based on the anti-PD-1 antibody pembrolizumab with the EC50 was generated (see Example 11). The single substitutions again covered a range of reduced potency compared to wt RLI2 (set at 100%) as EC50 on kit225 cells. Fusion to the antibody already reduced potency to about 50% (two RLI2 molecules fused (x2)) or about 15% (one RLI2 (x1) molecule fused by KIH technology), while a range of about 40% to about 0.4% was observed for the N65A substitution. The NQD mutation had the lowest potency in this assay, below the detection limit for 1x molecules and about 0.04% for x2 molecules. Furthermore, RLI2 with a mutation reducing binding to IL-2Rβγ fused to the light chain of the antibody was also shown to be effective in reducing the binding to IL-2Rβγ. AQ The immunocytokine based on the anti-PD-1 antibody pembrolizumab with RLI2 AQ The homodimeric light chain fusions were compared with the respective immunocytokines in which RLI2 was fused to the C-terminus of one heavy chain of the antibody (see Example 12). AQ The variants showed similar or slightly improved EC50 values ​​compared to the heterodimeric heavy chain fusions.

[0070] RLI2 AQ NA(RLI-15 AQA Compared to the mutant protein (also called β-lactamase), the QDQA (Q101D / Q108A) double substitution reduced potency in kit225 cells to approximately 50%, the NQD (D30N / E64Q / N65D) triple mutation reduced it to approximately 7%, and the DANA (D61A / N65A) double substitution reduced it to approximately 1%.

[0071] One RLI2 fused to a pembrolizumab derivative AQ PEM-RLI NA x1 constructs carrying NA (see SEQ ID NO:22, SEQ ID NO:23 and SEQ ID NO:24, but without the L235E substitution in the heavy chain) were shown to strongly reduce tumor volume in mouse tumor models compared to control untreated groups (p-value was <0.05) and similar to pembrolizumab treated groups (see Example 14).

[0072] In a further embodiment, the IL-15 variant comprises at least one further substitution that activates IL-15. Preferably, the activating mutation is at position N72, in particular N72D. AQ substitutions may also be used to reduce heterogeneity in conjugates comprising IL-15 variants with activating mutations at position N72, such as N72D as used in the clinical candidate IL-2 / IL-15Rβγ agonist ALT-803.

[0073] In a further embodiment, the IL-15 variant comprises at least one further substitution that reduces binding to IL-15Rα, preferably the site of the amino acid substitution that reduces binding to IL-15Rα may be one or more sites selected from the list consisting of L44, L45, E46, L47, V49, I50, S51, L52, E64, L66, I67, I68 or L69. L44, E46, L47, V49, I50, S51, L66 and I67 are preferred. The one or more substitutions are preferably selected from the list consisting of L44D, E46K, E46G, L47D, V49D, V49R, I50D, L66D, L66E, I67D and I67E. L44D, E46K, L47D, V49D, I50D, L66D, L66E, I67D, and I67E were specifically described for reducing binding to IL-15Rα (WO 2016 / 142314 A1), whereas replacement of L45, S51 and / or L52 by D, E, K or R, and replacement of E64, I68 and L69 by D, E, R or K increased binding to IL-15Rα (WO 2005 / 085282 A1). Similarly, IL-15 variants containing amino acid substitutions at positions V49 and I51 or V49, I50 and S51 and further containing one or more amino acid substitutions at positions N1, N4, S7, K10, K11, Y26, S29, D30, V31, H32, E53, G55, E64, I68, L69, E89, L91, M109, and / or I111 have been described that have reduced or no binding to the IL-15Rα and IL-2 / IL-15βγ receptors.

[0074] Preferred substitution combinations that reduce binding to IL-15Rα are E46G / V49R, N1A / D30N / E46G / V49R, N1G / D30N / E46G / V49R / E64Q, V49R / E46G / N1A / D30N and V49R / E46G / N1G / E64Q / D30N (WO 2019 / 166946A1). Similarly, amino acid sites L45, S51, L52, E64, I68, L69 have been described to reduce binding to IL-15Rα. Preferably, L45, S51 and / or L52 are substituted with D, E, K or R, and E64, I68, L69 are substituted with D, E, R or K (WO 2005 / 085282 A1).

[0075] In another embodiment, furthermore N71 is replaced by S, A or N, N72 is replaced by S, A or N, and N79 is replaced by S, A or G to reduce deamidation (WO 2009 / 135031 A1).

[0076] WO 2016 / 060996 A2 defines certain regions of IL-15 as suitable for substitution (see paragraphs 0020, 0035, 00120 and 00130) and specifically provides guidance on how to identify potential substitutions to provide anchors for PEG or other modifications (see paragraph 0021).

[0077] Additionally or alternatively, one of skill in the art can readily make conservative amino acid substitutions.

[0078] In another aspect, the present invention relates to a conjugate comprising an IL-15 variant of the present invention. IL-15 or IL-15 variants are used in various non-covalent or covalent conjugates in clinical or pre-clinical stages. RLI2 / SO-C101 / SOT101 (Cytune Pharma) is a covalent fusion protein of sushi+ fragment of IL-15Rα, a linker and IL-15. NIZ985 (Novartis) is a non-covalent conjugate of the heterodimer of IL-15 and soluble IL-15Rα. ALT-803 (Immunity-Bio / formerly Altor) is a homodimeric non-covalent conjugate of two IL-15 N72D variants non-covalently bound to an IL-15Rα sushi domain, each of which is fused N-terminally to an IgG1-Fc chain. P-22339 (Hengrui Medicine) is a homodimeric covalent conjugate of two IL-15 variants that have cysteine ​​substitutions to form an artificial disulfide bridge linking the IL-15 variant to two IL-15Rα sushi domains that are both fused N-terminally to an IgG1-Fc chain and also have cysteine ​​substitutions. XmAb24306 (Xencor, Genentech) is a heterodimeric covalent conjugate of an IL-2 / IL-15Rβγ binding reduced IL-15 variant fused N-terminally to one Fc chain and an IL-15Rα sushi domain fused N-terminally to the other Fc chain. CUG105 (Cugene) is a heterodimeric covalent conjugate of IL-15 fused N-terminally to one Fc chain and an IL-15Rα sushi domain fused N-terminally to the other Fc chain.Additionally, IL-15 or IL-15 variants have been used as conjugates with PEG, for example AM0015 (Armo Bio, Eli Lilly), THOR-924, 908, 918 (Synthorx, Sanofi) or NKTR-255 (Nektar Therapeutics). AQ mutations include RLI2. AQ and RLI2 AQ It is expected that similar improvements in the heterogeneity of such conjugates will be achieved, as shown by the inventors for the base immunocytokine.

[0079] In one embodiment, the conjugate further comprises a sushi domain of IL-15Rα or a derivative thereof. Complex formation of a polypeptide comprising a sushi domain with IL-15 occupies the IL-15Rα binding site of IL-15, thus on the one hand abolishing binding to IL-2 / IL-15Rαβγ and increasing the binding affinity to IL-2 / IL-15Rβγ (compared to IL-15 alone), and circumventing the need for transpresentation to IL-2 / IL-15Rβγ expressing cells, thereby making such a conjugate an IL-2 / IL-15Rβγ superagonist. As mentioned above, this concept has been adopted by several different approaches, including RLI2 / SO-C101 / SOT101, NIZ985, ALT-803, P-22339, XmAb24306 and CUG105. Some have used only the sushi domain, the minimal binding domain of IL-15Rα, to bind IL-15 (e.g., ALT-803), some have used the sushi+ fragment, an extended sushi domain with full binding activity to IL-15 (RLI2 / SO-C101 / SOT101), and others have used soluble IL-15Rα, a much larger polypeptide without its transmembrane domain (NIZ985). Derivatives of the sushi domain should either retain binding to IL-15 (retain at least 25%, preferably at least 50% of the binding of the respective sushi domain) or block binding to IL-2 / IL15Rαβγ in the conjugate (i.e. reduce the binding affinity to IL15Rαβγ by at least 1 log, preferably at least 2 logs). For example, WO 2016 / 095642 discloses sushi derivatives with cysteine ​​substitutions at positions K34, L42, A37, G38 or S40, preferably sushi S40C variant pairing with IL-15 variants with L52C substitution, to introduce artificial disulfide bonds with IL-15 variants with cysteine ​​substitutions at L45, Q48, V49, L52, E53, C88 or E89.

[0080] In another aspect of the invention, the invention relates to a fusion protein comprising an IL-15 variant of the invention. Fusion proteins are preferred conjugates according to the invention, which, compared to non-covalent conjugates, do not have the risk of dissociation of the conjugate after strong dilution upon administration to a patient. Also, expression of fusion proteins is typically more efficient and leads to a more homogenous product than the simultaneous expression of multiple polypeptide chains, or even the in vitro assembly of the polypeptides after individual purification. Fusion proteins comprising an IL-15 variant fused to the C-terminus of an antibody heavy chain are disclosed, for example, in WO 2019 / 166946 A1 (Pfizer) or WO 2018 / 184964 A1 (Roche), or fusion proteins comprising an IL-15 variant fused to each C-terminus of an antibody heavy chain are disclosed, for example, in WO 2016 / 142314 A1 (DKFZ, Univ. Tuebingen).

[0081] In one embodiment, the fusion protein of the invention further comprises a sushi domain of IL-15Rα or a derivative thereof, a targeting moiety, and / or a half-life extending moiety, and optionally one or more linkers. As mentioned above, fusion with a sushi domain of IL-15Rα or a derivative thereof is preferred because the resulting fusion protein does not bind to IL-15Rαβγ, does not require transpresentation of IL-15Rα, and has optimized targeting to IL-2 / IL-15Rβγ. Furthermore, IL-15 variants may be fused to a targeting moiety. The targeting moiety is primarily an antibody or functional fragment that binds to the same target, and the IL-15 or IL-15 / IL-15Rα fusion protein may be fused preferably to the C-terminus of one or both heavy chains (one heavy chain requires heterodimerization mutations in the Fc domain, such as KiH technology) or to both light chains. Other targeting moieties may be short binding tags, such as RGD motifs (see, e.g., WO 2017 / 000913), albumin binding domains (ABD) (see, e.g., WO 2018 / 151868A2), TCRs (see, e.g., WO 2008 / 143794), or antibody mimetics, such as anticalins, affibodies, adectins, aptamers, affimers, affitins, avimers, fynomers, armadillo repeat proteins, and knottins (Yu et al., 2017). The IL-15 variants may also be fused to half-life extending moieties, such as Fc domains or human serum albumin. By increasing the size of the protein, thereby slowing its clearance from the bloodstream, it increases its in vivo half-life and targets reactive immune cells, primarily NK cells and CD8 + Prolonging the stimulation of T cells is a common strategy in IL-15 development: fusion to the Fc domain has been employed, for example, in development candidates P-22339, XmAb24306 and CUG105.

[0082] In a preferred embodiment, the fusion protein of the invention comprises, preferably in N- to C-terminal order, a human IL-15Rα sushi domain, a linker, and an IL-15 variant of the invention. The order of receptor-linker-interleukin ("RLI") has been shown to be beneficial compared to the reversed order of ILR. Preferably, the human IL-15Rα sushi domain comprises the sequence of SEQ ID NO:5, the linker has a length of 18-22 amino acids and is composed of glycine or serine and glycine, and the IL-15 variant of the invention. The human sequence is preferred for human patients. A linker of 18-22 amino acids in length has been shown to be beneficial, with glycine or serine and glycine being preferred amino acids for the linker sequence to make the linker flexible and non-immunogenic. RLI2 / SO-C101 / SOT101 is a clinical stage fusion protein with the sushi+ fragment of IL-15Rα, which has been improved to have excellent homogeneity by introducing AQ substitutions. Thus, RLI2 AQ (SEQ ID NO: 9) is a preferred embodiment. Another preferred RLI molecule with a less potent IL-15 variant is RLIAQ N65A / RLI-15 AQA (SEQ ID NO: 10). Typically, the linker used is composed of glycine or serine and glycine, and has a length of 10 to 40 amino acids.

[0083] In another embodiment, the targeting moiety is preferably an antibody or functional variant thereof that binds to a tumor antigen, a tumor extracellular matrix antigen, or a tumor angiogenesis antigen, or is an immunomodulatory antibody.

[0084] The tumor antigen is preferably selected from EGFR, HER2, FGFR2, FOLR1, CLDN18.2, CEA, GD2, O-acetyl-GD-2, GM1, CAIX, EPCAM, MUC1, PSMA, c-Met, CD19, CD20, CD38. The tumor extracellular matrix antigen is preferably selected from FAP, the EDA domain of fibronectin, the EDB domain of fibronectin and LRRC15, preferably FAP and the EDB domain of fibronectin.

[0085] Preferably, the angiogenic antigen is selected from VEGF or endoglin (CD105).

[0086] The immunomodulatory antibody or functional variant thereof may be an immunomodulatory antibody that stimulates a costimulatory receptor, preferably selected from a CD40 agonist, a CD137 / 4-1 BB agonist, a CD134 / OX40 agonist and a TNFRSF18 / GITR agonist, or the immunomodulatory antibody may inhibit an immunoinhibitory receptor, preferably selected from a PD-1 antagonist, a CTLA-4 antagonist, a LAG3 antagonist, a TIGIT antagonist, an inhibitory KIR antagonist, a BTLA / CD272 antagonist, a HAVCR2 / TIM-3 / CD366 antagonist and an ADORA2A antagonist, more preferably a PD-1 antagonist.

[0087] Antibodies against the targets listed above are well known in the art or can be generated by standard immunization or phage display protocols. Non-human antibodies can be humanized. Examples of anti-EGFR antibodies are cetuximab, panitumumab, zalutumumab, nimotuzumab, and matuzumab. Examples of anti-HER2 antibodies are trastuzumab, pertuzumab, or margetuximab. Examples of anti-CLDN18.2 antibodies are zolbetuximab and the antibodies of the present invention below. An example of an anti-CEA antibody is arcitumomab. An example of an anti-GD2 is hu14.18K322A. An example of an anti-O-acetyl-GD-2 is c.8B6. Examples of anti-CD20 antibodies are rituximab, ocrelizumab, obinutuzumab, ofatumumab, ibritumomab, tositumomab and ublituximab. Examples of anti-CD38 antibodies are daratumumab, MOR202 and isatuximab.

[0088] Examples of anti-FAP antibodies are sibrotuzumab and B12 (US Patent Publication No. 2020-0246383A1). An example of an anti-EDA domain antibody of fibronectin is the F8 antibody (Villa et al., 2008; WO 2010 / 078945; WO 2014 / 174105), an example of an anti-EDB domain of fibronectin is the L19 antibody (Pini et al., 1998; WO 1999 / 058570), and an example of an anti-LRRC15 antibody is Samrotamab / huM25 (WO 2017 / 095805).

[0089] Examples of anti-VEGF antibodies are bevacizumab and ranibizumab. An example of an anti-endoglin antibody is TRC105 (WO2010039873A2).

[0090] Examples of anti-CD40 agonist antibodies are selicrelumab, APX005M, ChiLob7 / 4, ADC-1013, SEA-CD40 and CDX-1140 (Vonderheide 2020). Examples of anti-CD137 / 4-1 BB agonist antibodies are urelumab and utomilumab (Chester et al., 2018). Examples of anti-CD134 / OX40 agonist antibodies are PF-04518600, MEDI6469, MOXR0916, MEDI0562, INCAGN01949 (Fu et al., 2020). An example of an anti-TNFRSF18 / GITR agonist antibody is DTA-1.

[0091] Examples of PD-1 antagonists are anti-PD-1 antibodies, anti-PD-L1 antibodies, or anti-PD-L2 antibodies. Examples of anti-PD-1 antagonist antibodies (antagonistic antibodies) are pembrolizumab, nivolumab, pidilizumab, toripalimab, and tislelizumab (Dolgin, 2020). Examples of anti-PD-L1 antagonist antibodies are atezolizumab and avelumab. An example of an anti-CTLA-4 antagonist antibody is ipilimumab. An example of an anti-LAG3 antagonist antibody is leratolimab. Examples of anti-TIGIT antagonist antibodies are tiragolumab, vibostolimab, domvanalimab, etigilimab, BMS-986207, EOS-448, COM902, ASP8374, SEA-TGT, BGB-A1217, IBI-939 and M6223 (Dolgin, 2020).

[0092] An example of an anti-BTLA antagonist antibody is TAB004. Examples of anti-HAVCR2 / TIM-3 antagonist antibodies are LY3321367, MBG453 and TSR-022.

[0093] In a preferred embodiment, the fusion protein is fused to the C-terminus of at least one heavy chain of the antibody or to the C-terminus of both light chains of the antibody. Various immunocytokines, i.e., antibodies fused to cytokines, are shown in Examples 7-14 as RLI2 without a linker. AQ RLI2 has been produced and tested by fusing it to the C-terminus of one heavy chain (e.g., SEQ ID NO: 22) or both heavy chains (e.g., SEQ ID NO: 25), or to both light chains of a pembrolizumab-derived antibody (e.g., SEQ ID NO: 30). AQA linker may be used to fuse RLI2AQ to the C-terminus of one or both heavy chains. Such linkers are preferably composed of glycine or glycine and serine, more preferably composed of 30-50 amino acids in length from GGGGS units, in particular the L40 linker of SEQ ID NO: 31. Exemplary immunocytokines based on anti-CD20 antibodies with RLI2AQ fused to both heavy chains with L40 linkers were generated (SEQ ID NO: 32, SEQ ID NO: 34). To generate heterodimeric immunocytokines with one RLI molecule fused to one heavy chain, the KiH technology was applied, using a T366W mutation (knob) in one chain and T366S / L368A / Y407V (hole) in the other chain (Elliott et al., 2014). Other heterodimerization techniques, such as KiH S-S(T366W / S354C - T366S / L368A / Y407V / Y349C, (Merchant et al. 1998, Leaver-Fay et al. 2016)), HA-TF(S364H / F405A - Y349T / T394F, (Moore et al., 2011)), ZW1(T350V / L351Y / F405A / Y407V - T350V / T366L / K392L / T394W, (Von Kreudenstein et al., 2013)), 7.8.60(K360D / D399M / Y407A - E345R / Q347R / T366V / K409V, (Leaver-Fay et al., 2016)), DD-KK (K409D / K392D - D399K / E356K, (Gunasekaran et al., 2010)), EW-RVT (K360E / K409W - Q347R / D399V / F405T, (Choi et al. 2013, Choi et al. 2015)), EW-RVTS-S (K360E / K409W / Y349C - Q347R / D399V / F405T / S354C, (Choi et al., 2015)) are known in the art. SEED f (IgA-derived 45 residues on IgG1 CH3 - IgG1-derived 57 residues on IgA CH3 (Davis et al., 2010)), A107 (K370E / K409W - E357N / D399V / F405T (Choi et al., 2015)). The IgG4-based Fc domain of the immunocytokine was modified by L235E mutation to further reduce ADCC activity (Alegre et al., 1992) and / or by M252Y / S254T / T256E mutation to increase FcRn binding to extend in vivo half-life (Dall'Acqua et al., 2002). In another embodiment, antibodies targeting checkpoint inhibitors such as PD-1 or CTLA-4 may be in an IgG1 format engineered to strongly reduce or silence ADCC and / or CDC activity, e.g., to reduce FcγR and C1q binding. Suitable Fc modifications for immunocytokines are listed in Table 2.

[0094] [Table 2(1)] [Table 2(2)]

[0095] Different IL-15 variants (all with the AQ mutation) with additional mutations that reduce IL-2Rβγ binding were used in the RLI conjugate.

[0096] The N65A substitution in IL-15 was identified as the single mutation that modulates RLI-15 activity to a level suitable for many antibodies. AQA A fusion protein comprising is a preferred embodiment of the present invention.

[0097] One preferred embodiment is a PD-1 targeting fusion protein comprising the sequence and an antibody comprising the pembrolizumab-derived heavy chain knob sequence of SEQ ID NO:22 (fused to SEQ ID NO:10), the pembrolizumab-derived heavy chain hole sequence of SEQ ID NO:23, and the light chain sequence of SEQ ID NO:24, which is fused to the C-terminal heavy chain knob sequence without a linker. In a more preferred embodiment, the PD-1 targeting fusion protein comprises an antibody (SOT201) comprising SEQ ID NO:22, SEQ ID NO:38, and SEQ ID NO:24.

[0098] One preferred embodiment is a conjugate of the sequence of SEQ ID NO: 10 with an anti-CLDN18.2 heterodimeric IgG1 antibody variant having the VH and VL domain sequences of SEQ ID NO: 46 and SEQ ID NO: 47, respectively, which is heterodimeric via KiH mutations (T366W mutations (knob) in one chain and T366S / L368A / Y407V (hole) in the other chain). In a preferred embodiment, the conjugate comprises SEQ ID NO: 66, SEQ ID NO: 67 and SEQ ID NO: 68 (SOT202).

[0099] A further embodiment is any polypeptide that comprises an IL-15 variant listed in Table 1.

[0100] In another aspect of the invention, the invention relates to a nucleic acid encoding an IL-15 variant of the invention, a conjugate of the invention, or a fusion protein of the invention.

[0101] Furthermore, one aspect of the present invention relates to a vector comprising the nucleic acid of the present invention.

[0102] Furthermore, one aspect of the present invention relates to a host cell comprising the nucleic acid of the invention or the vector of the invention.

[0103] Another aspect of the present invention relates to an IL-15 variant of the present invention, a conjugate of the present invention, or any fusion protein of the present invention, a nucleic acid of the present invention, or a vector of the present invention for use in therapy. The IL-15 of the present invention, and thus the IL-15 variants, are potent cytokines that are used and / or tested clinically or preclinically as medicines for the treatment of neoplastic diseases (Robinson and Schluns, 2017) and infectious diseases.

[0104] Another aspect of the invention relates to a pharmaceutical composition comprising an IL-15 variant of the invention, a conjugate of the invention or a fusion protein of the invention, a nucleic acid of the invention or a vector of the invention and a pharma- ceutical acceptable carrier. In addition, the pharmaceutical composition may contain pharma- ceutical acceptable excipients such as surfactants, salts and / or cryoprotectants.

[0105] Yet another aspect of the present invention relates to an IL-15 variant of the present invention, a conjugate of the present invention, or a fusion protein of the present invention, a nucleic acid of the present invention or a vector of the present invention for use in the treatment of a subject suffering from, at risk of developing, and / or diagnosed with a neoplastic or infectious disease.

[0106] In one embodiment, the neoplastic disease is selected from solid tumors or hematological diseases. Examples of solid tumors are colorectal cancer, gastric cancer, melanoma, ocular melanoma, Merkel cell carcinoma, cutaneous squamous cell carcinoma, anal cancer, renal cell carcinoma, bladder cancer, adenocarcinoma, carcinoid tumor, leiomyosarcoma, breast cancer, triple-negative breast cancer, osteosarcoma, thyroid cancer, thymic cancer, bile duct cancer (cholangiocarcinoma), salivary gland cancer, adenoid cystic carcinoma, gastric cancer, head and neck squamous cell carcinoma, non-small cell lung cancer, small cell lung cancer, hepatocellular carcinoma, ovarian cancer, cervical cancer, biliary tract cancer, urothelial carcinoma and mesothelioma. In one embodiment, high-frequency microsatellite instability solid tumors are preferred. Examples of hematological cancers are leukemias such as acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL), chronic myelogenous leukemia (CML) and acute monocytic leukemia (AMoL), lymphomas such as Hodgkin's lymphoma, non-Hodgkin's lymphoma, and myeloma. In one embodiment, the infectious disease is selected from HIV, hepatitis A, B, or C, and herpes virus infection.

[0107] In one aspect, the present invention relates to a method of treating a subject, comprising administering a therapeutically effective amount of an IL-15 variant of the invention, a conjugate of the invention, or a fusion protein of the invention, a nucleic acid of the invention, or a vector of the invention to a subject in need thereof.

[0108] In one embodiment, the present invention relates to a polypeptide comprising the amino acid sequence set forth in SEQ ID NO:9.

[0109] In another embodiment, the present invention relates to a polypeptide comprising the amino acid sequence set forth in SEQ ID NO:10.

[0110] The present invention is further illustrated by the following embodiments.

[0111] 1. An IL-15 variant that contains an amino acid substitution at position G78 and position N79 of mature human IL-15.

[0112] 2. An IL-15 variant comprising SEQ ID NO:3.

[0113] 3. The IL-15 variant of embodiment 1 or embodiment 2, wherein said IL-15 variant is glycosylated.

[0114] 4. An IL-15 variant according to any one of embodiments 1 to 3, wherein said IL-15 variant has reduced glycosylation compared to glycosylated mature human IL-15.

[0115] 5. The IL-15 variant of any one of embodiments 1 to 4, wherein glycosylation of said IL-15 variant is increased at N71 of said IL-15 variant compared to glycosylated mature human IL-15.

[0116] 6. An IL-15 variant according to any one of embodiments 1 to 5, wherein said IL-15 variant is obtained by expression of a nucleic acid encoding said IL-15 variant in a mammalian cell.

[0117] 7. The IL-15 variant of embodiment 6, wherein the mammalian cell is a CHO cell.

[0118] 8. The IL-15 variant of any one of embodiments 1 to 7, wherein said IL-15 variant exhibits increased homogeneity compared to mature human IL-15.

[0119] 9. The IL-15 variants are those described herein that are IL-2 / IL-15Rβ and / or γ c The IL-15 variant of any one of embodiments 1 to 8, further comprising an amino acid substitution that reduces binding to the receptor and / or IL-15Rα.

[0120] 10. The IL-15 variant of any one of embodiments 1 to 9, wherein said IL15 variant comprises G78A and N79Q.

[0121] 11. A composition comprising an IL-15 variant according to any one of embodiments 1 to 10, wherein less than 30%, preferably less than 25%, of the IL-15 variants in the composition are glycosylated.

[0122] 12. A composition comprising an IL-15 variant described in any one of embodiments 1 to 11, wherein more than 15% and less than 25% of the IL-15 variants in the composition are glycosylated at N71.

[0123] 13. The composition of embodiment 11 or embodiment 12, wherein the composition exhibits increased homogeneity compared to a composition comprising mature human IL-15.

[0124] 14. The composition of embodiment 11 or embodiment 12, wherein the composition exhibits a more homogenous glycosylation pattern compared to a composition comprising mature human IL-15.

[0125] 15. A conjugate comprising an IL-15 variant according to any one of embodiments 1 to 10 and the sushi domain of IL-15Rα or a derivative thereof.

[0126] 16. A fusion protein comprising an IL-15 variant according to any one of embodiments 1 to 10 and the sushi domain of IL-15Rα or a derivative thereof.

[0127] 17. An immunocytokine comprising an IL-15 variant according to any one of embodiments 1 to 10, a conjugate according to embodiment 15 or a fusion protein according to embodiment 16, and an antibody or a functional variant thereof.

[0128] 18. The immunocytokine described in embodiment 17, wherein the antibody is an antibody described herein or a functional variant thereof.

[0129] 19. The immunocytokine of embodiment 17, wherein the antibody is an immunomodulatory antibody or a functional variant thereof, preferably an antibody against PD-1, PD-L1 or PD-L2 or a functional variant thereof.

[0130] 20. A nucleic acid encoding an IL-15 variant according to any one of embodiments 1 to 10, a conjugate according to embodiment 15, a fusion protein according to embodiment 16 or an immunocytokine according to any one of embodiments 17 to 19.

[0131] 21. A vector comprising the nucleic acid described in embodiment 20.

[0132] 22. A host cell comprising the nucleic acid of embodiment 20 or the vector of embodiment 21.

[0133] 23. A method for preparing an IL-15 variant according to any one of embodiments 1 to 10, a conjugate according to embodiment 15, a fusion protein according to embodiment 16 or an immunocytokine according to any one of embodiments 17 to 19.

[0134] 24. An IL-15 variant according to any one of embodiments 1 to 10, a conjugate according to embodiment 15, a fusion protein according to embodiment 16 or an immunocytokine according to any one of embodiments 17 to 19 for use in therapy.

[0135] 25. An IL-15 variant according to any one of embodiments 1 to 10, a conjugate according to embodiment 15, a fusion protein according to embodiment 16 or an immunocytokine according to any one of embodiments 17 to 19 for use in the treatment of a tumor or infectious disease.

[0136] 26. A polypeptide comprising SEQ ID NO:9 or SEQ ID NO:10.

[0137] The present invention is further illustrated by the following embodiments.

[0138] 1. An interleukin-15 (IL-15) variant containing amino acid substitutions at positions G78 and N79 of mature human IL-15.

[0139] 2. An IL-15 variant according to embodiment 1, wherein said IL-15 variant comprises the amino acid substitutions G78A, G78V, G78L or G78I, and N79Q, N79H or N79M, preferably G78A and N79Q.

[0140] 3. The IL-15 variant according to embodiment 1 or embodiment 2, wherein said IL-15 variant is expressed in a mammalian cell line, preferably said mammalian cell line is selected from CHO cells, HEK293 cells, COS cells, PER.C6 cells, SP20 cells, NSO cells or any cells derived therefrom, more preferably CHO cells.

[0141] 4. The above amino acid substitutions are (a) reducing deamidation at N77 and glycosylation at N79 of said IL-15 variants compared to mature human IL-15; (b) resulting in less than 30% glycosylated IL-15 variants, preferably less than 25% glycosylated IL-15 variants; and / or (c) increasing glycosylation at N71 of the IL-15 variants compared to mature human IL-15; An IL-15 variant according to any one of embodiments 1 to 3.

[0142] 5. An IL-15 variant according to any one of embodiments 1 to 4, wherein said amino acid substitutions do not substantially reduce the IL-15 activity of said IL-15 variant on inducing proliferation of kit225 cells, 32Db cells, human PBMCs or in the Promega IL-15-bioassay.

[0143] 6. An IL-15 variant according to any one of embodiments 1 to 5, wherein said IL-15 variant does not have a substitution at position N71 and / or N77.

[0144] 7. The IL-15 variant is IL-2 / IL-15Rβ and / or γ c 7. An IL-15 variant according to any one of embodiments 1 to 6, comprising at least one further substitution which reduces binding to the receptor and / or IL-15Rα.

[0145] 8. (a) the IL-2 / IL-15Rβ and / or the gamma c The site of said further substitution that reduces binding to the receptor is selected from the list consisting of N1, N4, S7, D8, K10, K11, D30, D61, E64, N65, L69, N72, E92, Q101, Q108 and I111, preferably from the list consisting of D61, N65 and Q101, most preferably N65; or (b) the IL-2 / IL-15Rβ and / or the gamma c said further substitutions decreasing the binding to the receptor are selected from the list consisting of N1D, N1A, N1G, N4D, S7Y, S7A, D8A, D8N, K10A, K11A, D30N, D61A, D61N, E64Q, N65D, N65A, N65E, N65R, N65K, L69R, N72R, Q101D, Q101E, Q108D, Q108A, Q108E and Q108R, preferably from the list consisting of D8A, D8N, D61A, D61N, N65A, N65D, N72R, Q101D, Q101E and Q108A, more preferably from the list consisting of D61A, N65A and Q101, most preferably N65A; or (c) the IL-2 / IL-15Rβ and / or the gamma c The further substitutions that reduce receptor binding are combination substitutions and are selected from the list consisting of D8N / N65A, D61A / N65A and D61A / N65A / Q101D. An IL-15 variant as described in embodiment 7.

[0146] 9. (a) the site of the additional substitution that reduces binding to IL-15Rα is selected from the list consisting of L44, L45, E46, L47, V49, I50, S51, E64, L66, I67, I68, and L69; or (b) the additional substitution that reduces binding to IL-15Rα is selected from the list consisting of L44D, E46K, E46G, L47D, V49D, V49R, I50D, L66D, L66E, I67D, and I67E; or (c) the further substitutions that reduce binding to IL-15Rα are a combination of substitutions selected from the list consisting of E46G / V49R, N1A / D30N / E46G / V49R, N1G / D30N / E46G / V49R / E64Q, V49R / E46G / N1A / D30N and V49R / E46G / N1G / E64Q / D30N. An IL-15 variant as described in embodiment 7.

[0147] 10. A conjugate comprising an IL-15 variant according to any one of embodiments 1 to 9.

[0148] 11. The conjugate of embodiment 10, wherein said conjugate further comprises a sushi domain of IL-15Rα or a derivative thereof.

[0149] 12. A fusion protein comprising an IL-15 variant described in any one of embodiments 1 to 9.

[0150] 13. The fusion protein of embodiment 12, wherein the fusion protein further comprises the sushi domain of IL-15Rα or a derivative thereof, a targeting moiety, and / or a half-life extending moiety, and optionally one or more linkers.

[0151] 14. The fusion protein comprises, preferably in order from N-terminus to C-terminus, a human IL-15Rα sushi domain, a linker and an IL-15 variant according to any one of embodiments 1 to 9, preferably wherein the human IL-15Rα sushi domain comprises the sequence of SEQ ID NO: 5; The linker has a length of 18 to 22 amino acids and is composed of serine and glycine, More preferably, the fusion protein is SEQ ID NO: 9 or SEQ ID NO: 10. 14. The fusion protein of embodiment 13.

[0152] 15. A fusion protein described in any one of embodiments 12 to 14, wherein the targeting moiety is preferably an antibody or a functional variant thereof that binds to a tumor antigen, a tumor extracellular matrix antigen, or a tumor angiogenesis antigen, or is an immunomodulatory antibody.

[0153] 16. The fusion protein of embodiment 15, wherein the fusion protein is fused to the C-terminus of at least one heavy chain of the antibody or to the C-terminus of both light chains of the antibody.

[0154] 17. A nucleic acid encoding an IL-15 variant according to any one of embodiments 1 to 9, a conjugate according to embodiment 10 or embodiment 11, or a fusion protein according to any one of embodiments 12 to 16.

[0155] 18. A vector comprising the nucleic acid described in embodiment 17.

[0156] 19. A host cell comprising the nucleic acid of embodiment 17 or the vector of embodiment 18.

[0157] 20. An IL-15 variant according to any one of embodiments 1 to 9, a conjugate according to embodiment 10 or embodiment 11, or a fusion protein according to any one of embodiments 12 to 15, a nucleic acid according to embodiment 17, or a vector according to embodiment 18, for use in therapy.

[0158] 21. A pharmaceutical composition comprising an IL-15 variant according to any one of embodiments 1 to 9, a conjugate according to embodiment 10 or embodiment 11, or a fusion protein according to any one of embodiments 12 to 15, a nucleic acid according to embodiment 17, or a vector according to embodiment 18, and a pharma- ceutical acceptable carrier.

[0159] 22. An IL-15 variant according to any one of embodiments 1 to 9, a conjugate according to embodiment 10 or embodiment 11, or a fusion protein according to any one of embodiments 12 to 15, a nucleic acid according to embodiment 17 or a vector according to embodiment 18, for use in the treatment of a subject suffering from, at risk of developing and / or diagnosed with a neoplastic or infectious disease. EXAMPLES

[0160] 1. Expression and purification, general materials and methods RLI2 (RLI2 wt), RLI2 with a G78A substitution (RLI2 A) and RLI2 with a G78A / N79Q substitution (RLI2 AQ) were transiently expressed in CHO cells and purified from the supernatant by supernatant thawing, concentration and diafiltration, optional clarification, a Q-sepharose chromatography step, a phenyl-sepharose chromatography step, buffer exchange (dialysis) and concentration as detailed below.

[0161] Concentration and diafiltration with TFF1 After thawing, the sterile filtered CHO supernatant (875 mL for RLI2 wt or about 2800 mL for mutants) was concentrated and diafiltered for buffer exchange. The CHO supernatant was concentrated from a factor of 2.5 (for RLI wt) or about 5.5 (for RLI mutants) and diafiltered for buffer exchange (with buffer 25 mM Tris-HCl pH 7.5) with about 7 volumes of diafiltration buffer. If necessary, the material was then clarified by centrifugation at 15000g for 30 minutes at 20°C, then filtered through 0.45 μm and 0.22 μm PES membrane filters and immediately poured onto Q-Sepharose resin.

[0162] Capture by anion exchange chromatography (AEX) on Q-Sepharose resin Each diafiltered CHO supernatant was loaded at 200 cm / h (50.7 mL / min; residence time 3 min) onto a 150 mL column of Q-Sepharose (diameter 44 mm, bed height 10 cm) after pre-equilibration in buffer B (25 mM Tris HCl pH 7.5, 1 M NaCl) and then in buffer A (25 mM Tris HCl pH 7.5). After loading, the column was washed with 10 CV of buffer A at the same flow rate. Proteins were eluted from the column by increasing salt concentration. A first 15 CV linear gradient was applied from 0% to 25% buffer B (25 mM Tris HCl pH 7.5, 1 M NaCl), followed by a 5 CV step at 25% buffer B (step 1) and a 10 CV step at 100% buffer B (step 2). Finally, a 10 CV re-equilibration step was applied with buffer A. The purification was followed using the UV signal at 280 nm.

[0163] The linear gradient elution was fractionated and the first 10 CV were collected in 40 mL fractions, then 5 CV in fraction F5. The 250 mM NaCl step, 1 M NaCl step and re-equilibration were collected in fractions F6, F7 and F8, respectively. Purified fractions were analyzed by SDS-PAGE and anti-RLI Western blot for determination of elution pool.

[0164] Purification by hydrophobic chromatography on phenyl-Sepharose resin Each Q-Sepharose elution pool was diluted 1.6-fold online to 750 mM ammonium sulfate in buffer B (25 mM Tris-HCl pH 7.5; 2 M ammonium sulfate) and loaded at 149 cm / h (20 mL / min; residence time 5 min) onto a 100 mL Phenyl-Sepharose column (diameter 32 mm, bed height 12.4 cm) pre-equilibrated in a mixture of 62.5% buffer A (25 mM Tris HCl pH 7.5) and 37.5% buffer B (25 mM Tris-HCl pH 7.5; 2 M ammonium sulfate). After loading, the column was washed with 5 CV of the mixture 62.5% buffer A / 37.5% buffer B at the same flow rate. Proteins were eluted from the column by decreasing the salt concentration. A linear gradient of 20 CV was applied from 37.5% to 0% buffer B followed by a 5 CV step at 100% A (step 2). Finally, a 5 CV step was applied with buffer C (isopropanol 30%, step 3) for stripping. Purification was followed using the UV signal at 280 nm. The elution with a linear gradient was fractionated and collected in 40 mL fractions. Purified fractions were analyzed by SDS-PAGE and anti-RLI Western blot for determination of the elution pool.

[0165] Formulation process: Concentration and diafiltration by TFF The phenyl-Sepharose elution pool was concentrated by a factor of 2.6-4.4 and diafiltered for buffer exchange (using formulation buffer 20 mM L-histidine, 6% D-sorbitol, pH 6.5) with at least 7 volumes of diafiltration buffer. This material was then immediately concentrated on a Vivaspin unit with a 10 kDa cutoff to reach the final target concentration.

[0166] concentrated The diafiltered sample was concentrated using a Vivaspin unit with a 10 kDa cutoff until a theoretical concentration of 1 mg / mL was reached.

[0167] Potency assay with kit225 The activity of both IL-2 and IL-15 can be determined by induction of proliferation of kit225 cells as described by Hori et al. (1987). kit225 cells (Hori, Uchiyama et al. 1987) were passaged in kit225 basal medium and used for potency assays at passages 4 to 7. Prior to potency assays, kit225 cells were cultured for 24 hours in kit225 basal medium without IL-2 (starvation period). 1 × 10 4 Kit225 cells were plated in 96-well plates and serial dilutions of RLI-15 and the respective molecule PEM-RLI-15 were added to the cells. The cells were incubated at 37° C., 5% CO2 for 72±3 hours. After incubation, 10 μl (10% of the volume in the well) of Alamar Blue was added to each well and after 6 hours, absorbance was measured at 560 nm with a 620 nm reference (mixing was set for 15 seconds before detection) using a Tecan Spark absorbance microplate reader. In some cases, when lower potency RLI2 variants were tested, incubation with kit225 cells was extended from 3 days (72 hours±3 hours) to 5 days.

[0168] Preferably, methods such as colorimetry or fluorescence are used to determine proliferation activation by IL-2 or IL-15 stimulation, for example as described by Soman et al. (Soman, Yang et al. 2009) using CTLL-2 cells. As an alternative to cell lines such as kit225 cells, human peripheral blood mononuclear cells (PBMCs) or buffy coats can be used. A preferred bioassay for determining IL-2 or IL-15 activity is the IL-2 / IL-15 Bioassay Kit (Promega Catalogue No. CS2018B03 / B07 / B05) using STAT5-RE CTLL-2 cells.

[0169] The concentrations of the RLI variants analyzed were as follows: RLI2 supernatant: 0.133mg / ml (ELISA, average from two experiments) RLI2AQ Supernatant: 0.0297mg / ml (ELISA, average from two experiments)

[0170] RLI2 Characteristics Purity (RP-UPLC) 99.8% Preparation 20 mM histidine, 6% (w / v) sorbitol, pH 6.5 Storage temperature -20℃

[0171] kit225 basal medium RPMI (460 mL) + FBS (30 mL) + Glutamax (5 mL) + penicillin-streptomycin (5 mL) + cytokines added to flask (75 cm2); IL-2 (5 ng / mL). Cytokines were added to the medium immediately before culture.

[0172] hPBMC potency assay Buffy coats were obtained from healthy donors. PBMCs were isolated by Ficoll Paque gradient, washed three times, and resuspended in T cell complete medium in 96-well plates. Immune cytokines were added at the indicated concentrations, and plates were incubated at 37°C, 5% CO2 for 7 days. Proliferation of immune cell populations was detected by flow cytometry.

[0173] T cell complete medium RPMI1640 medium, CTS GlutaMAX-I 1x, 100U / mL penicillin-streptomycin, 1mM sodium pyruvate, NEAA 1x (non-essential amino acid mixture), 2-mercaptoethanol 0.05mM and 10% AB human serum (heat inactivated).

[0174] [Table 3]

[0175] Isolation of human NK cells (hNK): Fresh blood from healthy donors was diluted in a 1:1 ratio with cold PBS-EDTA, ph 7.4, and PBMCs were isolated by Ficoll-Paque gradient isolation. Isolated PBMCs were resuspended in complete culture medium. hNK cells were isolated from PBMCs using the EasySep Human NK Cell Isolation kit (Stem Cell Technologies, USA) according to the manufacturer's instructions. The isolated hNK cells from each donor were diluted at 3 × 10 6 The cells were resuspended in NK medium containing 10% serum at a concentration of cells / ml.

[0176] PD-1 / PD-L1 blockade bioassay The assay was performed according to the manufacturer's instructions (Promega PD-1 / PD-L1 Blockade Bioassay J1250). Briefly, PD-L1 aAPC / CHO-K1 cells were plated in a 96-well plate and incubated for 16-20 hours in a 37°C, 5% CO2 incubator. The indicated concentrations of PEM-RLI immunocytokines and PD-1 effector cells were then added to the cells and incubated for 6 hours in a 37°C, 5% CO2 incubator. After the incubation period, Bio-Glo™ reagent was added to the wells, incubated for 15 minutes at room temperature, and luminescence measurements were performed.

[0177] Cynomolgus monkey research The pharmacokinetics of the indicated PEM-RLI molecules were studied in cynomolgus monkeys (n=2-3) on days 1 or 15 after administration of the indicated doses. Blood for serum isolation was collected 1, 4, 8, 24, 48, 60, 72, 84, 96, 120 and 168 hours after administration (some time points were omitted in some cases). Serum concentrations of immune cytokines were determined by ELISA using the antibodies in Table 3. Selected immune cell populations (NK cells and CD8 +Blood for flow cytometric assessment of T cells) was collected pre-dose, and on days 5, 8, 12, 15, 19, 22, and 26.

[0178] [Table 4]

[0179] [Table 5]

[0180] Mouse Efficacy Study The purpose of these studies was to obtain female hPD1 single KI HuGEMM mice (C57BL / 6-Pdcd1 em1(hPDCD1) / Smoc The objective of this study was to evaluate the in vivo therapeutic efficacy of PEM-RLI2 NA x1 and pembrolizumab as monotherapy in the treatment of HuCell MC38-hPD-L1 tumor cell line in mice (n=8 mice / group). Each mouse was inoculated with MC38-hPD-L1 tumor cells (1×10 6 ) was inoculated subcutaneously. The average tumor size was 108 mm 3 Randomization began when the mean age of the tumor reached 18.4 mm. Forty mice were enrolled in this study. All animals were randomly assigned to five test groups. Randomization was based on the "Matched distribution" method (StudyDirector™ software, version 3.1.399.19). The date of randomization was designated as day 0 (D0). After tumor cell inoculation, animals were checked daily (or more frequently, as necessary, at the discretion of the Study Director) for morbidity and mortality. Tumor volumes were measured three times a week in two dimensions using calipers, and volumes were calculated in mm using the formula: V = (L x W x W) / 2, where V is the tumor volume, L is the tumor length (longest tumor dimension), and W is the tumor width (longest tumor dimension perpendicular to L). 3PEM-RLI2 NA x1 was administered IV at 20 mg / kg on day 0 and pembrolizumab was administered IP at 5 mg / kg on days 0, 3, 6, and 9. Tumor observation was continued for 18 days. Concurrently, PEM-RLI2 NA x1 (IL-15 with N65A and AQ mutations) was administered IV at 5, 10 on day 0. Tumor observation was continued for 6 days.

[0181] Mixed lymphocyte reaction Buffy coats were obtained from healthy donors. PBMCs were isolated by Ficoll Paque gradient and washed three times. PBMCs were isolated by Ficoll Paque gradient and washed three times. Pairs of hPBMC donors were cultured with equimolar concentrations of 1 nM pembrolizumab and PEM L-RLI NA x1 for 6 days. IFNγ production in cell supernatants was determined using the Human IFN-γ DuoSet ELISA (R&D systems, no. DY258B). Data are presented as relative response [%] of IFNγ production and represent the mean ± SEM from 12 pairs of hPBMC healthy donors.

[0182] 2. SDS-PAGE and anti-RLI Western blot analysis The purified proteins from Example 1 were analyzed by SDS-PAGE and anti-RLI Western blot.

[0183] Coomassie staining: protein bands are visualized according to their molecular weight under denaturing conditions. Briefly, 1 volume of loading buffer (with or without β-mercaptoethanol) was added to 3 volumes of samples to be analyzed (then more or less diluted in 1× loading buffer), homogenized, and denatured at 95° C. for 5 min. The denatured samples were loaded onto Criterion TGX gels and run in 1× TGS buffer at constant voltage (300 V) and limiting current (75 mA or 135 mA per gel depending on gel type) in running buffer for 18 or 21 min depending on gel type. The gels were removed from the cassette, washed 3 times in water for 5 min, stained with Biosafe staining solution (Biorad) for 20 min, washed 3 times in water for 20 min, followed by a final destain wash in water for 3 h. The stained gels were then scanned on a gel scanner.

[0184] Western blot analysis: The gel is then transferred to a nitrocellulose membrane and used for Western blot analysis with different antibodies. At the end of the transfer, the gel is used for protein transfer to a nitrocellulose membrane. For the reference example (Biorad number 170-4155, Trans-BlotR Turbo™ Transfer Starter System), the transfer parameters are 2.5A, 25V, 7 minutes (for Criterion gels) or 2.5A, 25V, 3 minutes (for Mini-PROTEAN gels). After membrane saturation in iBind™ Flex solution, antibody incubation and washing steps are then performed in the iBind system. After exposure and complete drying, the membrane is scanned for analysis. The primary antibody used was anti-RLI2-PR01 antibody (Cytune, dilution 1:25000) and the secondary antibody used was donkey anti-rabbit IgG-AP antibody (Santa Cruz Biotechnology, dilution 1:5000).

[0185] 3. Capillary electrophoresis Protein analysis by capillary electrophoresis relies on the separation of LDS-labeled protein variants through a sieving matrix in a constant electric field. The Labchip GXII instrument uses a single sipper microfluidic chip to characterize protein samples loaded onto a 96-well plate. The microfluidic chip technology allows the separation and analysis of protein samples. After detection and analysis of the laser-induced signal, the data provided are: relative protein concentration, molecular size and percent purity using ladder and marker calibration standards.

[0186] The samples are denatured by mixing 5 μL of sample and 35 μL of HT Protein Sample Buffer with or without DTT at a final concentration of 35 mM. If necessary, the samples are pre-diluted in HT Protein Sample Buffer at 1 mg / mL. Denaturation is performed by heating the mixture at 100° C. for 5 minutes. 70 μL of water is then added and the samples are centrifuged at 2,000 g for 10 minutes. The samples (in a 96-well plate) are then loaded onto the LabChip GXII instrument for chip transfer and analysis.

[0187] [Table 6]

[0188] 4. Glycosylation / Deamidation Mutants [Table 7]

[0189] The RLI2 molecule has a major glycosylation site at N176 (RLI numbering) and a secondary site at N168. No glycosylation is found at N209. The glycans are complex, mostly biantennary, fucosylated, and rarely sialylated G0-G2. In cell culture, about 40-50% of the protein is glycosylated, with N168 at about 5%. After purification as described above, about 14-25% of RLI2 is glycosylated. Different levels of glycosylation do not show any effect on potency, stability, and only a minor effect on pharmacokinetics, and glycosylated RLI2 has a shorter half-life, but the heterogeneity of the active pharmacological moiety remains problematic from a regulatory point of view.

[0190] A potential hotspot for deamidation identified in IL-15 expressed in E. coli (Nellis et al., 2012) is N77 (IL-15 numbering) / N174 (RLI numbering). Although it has been described that N-glycosylation of N79 partially prevents N77 deamidation (Thaysen-Andersen et al., 2016), we have indeed seen in mass spectrometry that N77 is deamidated in CHO-expressed RLI2, identifying deamidation as a real issue for potential heterogeneity of RLI2 and RLI-based products. Therefore, deamidation should be avoided.

[0191] Figure 1A shows that RLI2 wt (no mutations) is indeed a heterogeneous product with two major bands of approximately 20 kDa and 25 kDa and a few smaller bands, all immunoreactive with anti-RLI2 antibodies and thus representing different modifications of the RLI2 protein.

[0192] We wanted to avoid mutating N77 as an obvious way to eliminate the deamidation of N77 and thereby remove the polar amide, since a conservative substitution to glutamine did not resolve the risk of deamidation. To eliminate the potential deamidation at position N77, a single substitution G78A (IL-15 numbering) / G175A (RLI numbering) in RLI2 was instead introduced (RLI2A). Although the loss of deamidation was not seen by Coomassie staining or Western blot, the major acidic peak (pI 6.0) in RP-UPLC was significantly reduced in cIEF, as expected for the loss of deamidation, confirming that the deamidation hotspot N174 was indeed deamidated (data not shown). Mass spectrometry of the PEM-RLI AQ construct also showed the absence of deamidation (data not shown).

[0193] Surprisingly, the G78A mutation resulted in a slight increase in glycosylation (see FIG. 1A, better seen in FIG. 1B), with larger / more glycosylated species compared to RLI2 wt. Additional bands appear, indicating this new glycosylation pattern (see dashed box 3 in FIG. 1B). The RP-UPLC peak was also slightly shifted (data not shown). This altered glycosylation pattern was unexpected, as the effect of the deamination mutation G78A on glycosylation was unpredictable.

[0194] Q (RLI2AQ, RLI2 AQA significant reduction in the larger species of RLI2 was observed with the additional substitution of N79 (IL-15 numbering) / N176 (RLI numbering) with N79 (IL-15 numbering) / N168 (RLI numbering) (see dashed box 1 in FIG. 1B). The remaining larger band (see solid box 2 in FIG. 1B) likely represents glycosylation at N71 (IL-15 numbering) / N168 (RLI numbering) of approximately 20% of the RLI molecules, which appears to be slightly increased compared to RLI2 wt and RLI2. The box 1 band may represent RLI2 glycosylated at N176, whereas the box 3 band may represent RLI2 glycosylated at N176 and N168. However, the box 3 band may also be RLI2 glycosylated with an unfavorable sialic acid glycan structure at N176. Without wishing to be bound by any theory, a possible explanation for this surprising increase in glycosylation at N71 is that glycosylation at the key site N79 sterically interferes with glycosylation at N71 in RLI2 wt, and such interference is alleviated when N79 is mutated.

[0195] In summary, RLI2 with AQ substitutions AQ , and thus IL-15 AQ also represent RLI2, or IL-15, variants with greatly improved homogeneity and reduced risk of deamidation.

[0196] To compare the effect / impact of glycosylation on the biological activity of RLI variants, we specifically inactivated three potential glycosylation sites N71 / N79 / N160 (N168 / N176 / N209 for RLI) of IL-15 by site-directed mutagenesis (Stratagene Site Directed Mutagenesis XL Kit). N168S / N176Q / N209S and RLI1 N168S / N176Q / N209S To confirm the predominant N-glycosylation occupancy at N79 (=N176 in RLI), RLI2 N176QA mutant was generated, and transient expression in CHO cells results in a unique 25 kDa band (see FIG. 2, right panel).

[0197] The RLI protein mutated only at the major glycosylation site (RLI2 N176Q ) also showed a unique 25 kDa band, thus confirming the major glycosylation occupancy on the N176 residue of RLI expressed in CHO (transiently expressed). The secretion yields of the deglycosylated mutants transiently expressed in CHO cells were similar to their glycosylated / original counterparts. Thus, there was no significant effect of deglycosylation on the expression levels. The same was observed in a Pichia Pastoris expression system (data not shown).

[0198] Moreover, these mutations on the N-glycosylation sites do not appear to induce any significant effect on the in vitro proliferation activity of RLI on kit225 or 32Dβ cells. As usual, all RLI versions (RLI1 or RLI2, glycosylated or non-glycosylated, CHO or baculo or pichia) stimulated proliferation of kit225 cell lines similarly.

[0199] 5. RLI2 AQ The Potency of Varieties The activity of both IL-2 and IL-15 can be determined by induction of proliferation of kit225 cells as described by Hori et al. (1987). Kit225 cells (Hori et al., 1987) were passaged in kit225 basal medium and used for potency assays at passages 4 to 7. Prior to potency assays, kit225 cells were cultured in kit225 basal medium without IL-2 for 24 hours (starvation period). 1 × 10 4Kit225 cells were plated in 96-well plates and serial dilutions of RLI-15 and the respective molecule PEM-RLI-15 were added to the cells. The cells were incubated at 37° C., 5% CO2 for 72±3 hours. After incubation, 10 μl (10% of the volume in the well) of Alamar Blue was added to each well and after 6 hours, absorbance was measured at 560 nm with a 620 nm reference (mixing was set for 15 seconds before detection) using a Tecan Spark absorbance microplate reader. In some cases, when lower potency RLI2 variants were tested, incubation with kit225 cells was extended from 3 days (72 hours±3 hours) to 5 days.

[0200] Preferably, methods such as colorimetry or fluorescence are used to determine proliferation activation by IL-2 or IL-15 stimulation, for example as described by Soman et al. (Soman et al., 2009) using CTLL-2 cells. As an alternative to cell lines such as kit225 cells, human peripheral blood mononuclear cells (PBMCs) or buffy coats can be used. A preferred bioassay for determining IL-2 or IL-15 activity is the IL-2 / IL-15 Bioassay Kit (Promega Catalogue No. CS2018B03 / B07 / B05) using STAT5-RE CTLL-2 cells. The concentrations of the RLI variants analyzed were as follows: RLI2 supernatant: 0.133mg / ml (ELISA, average from two experiments) RLI2 AQ Supernatant: 0.0297mg / ml (ELISA, average from two experiments)

[0201] [Table 8]

[0202] [Table 9]

[0203] Therefore, the glycosylated mutant RLI2 as a supernatant AQ showed very similar potency in stimulating kit225 and / or 32Db cells when compared to RLI2 from the supernatant, which was surprising since for many glycoproteins loss of glycosylation results in lower activity.

[0204] In SPR (Biacore) binding experiments to the IL-2 / IL-15βγ receptor, the k between RLI2 and RLI AQ was also on speed, k off Rate and equilibrium constant K d No relevant differences in were observed (data not shown).

[0205] In summary, RLI2 with AQ substitutions AQ , and thus IL-15 AQ represents an RLI2, or IL-15, variant with greatly improved homogeneity, reduced risk of deamidation, and equivalent potency in activating immune cells.

[0206] 6. Cynomolgus monkey PK / PD study of hyperglycosylated and hypoglycosylated RLI2 To compare hyperglycosylated and hypoglycosylated RLI2 with respect to their PK and PD properties, a 200 L scale production campaign was performed, harvested on S0SP and X0SP depth filters, and the protein was captured on a PPA column. Virus was inactivated by solvent detergent treatment and purification continued through a Capto Adhere column and a hydroxyapatite type II column (flow-through mode) followed by a second virus removal step by nanofiltration. The RLI preparation was polished on a Capto Impres Phenyl column (CPI Phenyl HIC) and selected fractions for hyperglycosylated RLI2 were pooled (RLI-15-HG) and selected fractions for hypoglycosylated RLI2 were pooled (RLI-15-LG) (see Figure 5A-C). Finally, UFDF filtration was performed onto a 10 kDa cutoff UF membrane into final formulation buffer (20 mM histidine, 6% sorbitol, pH 6.5). RLI-15-HG shows the majority of the RLI in the upper band of glycosylated RLI isomers, whereas RLI-15-LG contains only a smaller fraction of glycosylated RLI isomers (Figures 5B and C).

[0207] A total of 3 male and 3 female cynomolgus monkeys were included in the PK / PD study. Animals were assigned to two groups receiving RLI2 at 15 μg / kg (nominal dose) as RLI-15-HG and RL1-15-LG by daily subcutaneous administration according to a crossover dosing design. Dosing was over two 4-day periods (2×4) separated by a 10-day washout period (Days 1-4: RLI-125-LG for males and RLI-15-HG for females, Days 15-18: RLI-15-HG for males and RLI-15-LG for females). Pharmacodynamic parameters (NK, CD4 + and CD8 +Expression of IgG4-associated leukemia cells (including Ki67 expression in cells) was analyzed from blood samples collected in the pretreatment period, on days 5, 12 and 19. Blood samples for pharmacokinetic studies were taken from all animals on days 1 and 15 after the first dose in each treatment period at the following time points: pre-dose and 0.5, 1, 2, 6, 12 and 24 hours after dosing. Bioanalysis was performed. In addition, back-up serum samples (D1 (pre-dose)). D15 (pre-dose) and D16 (24 hours) were partially used for immunogenicity evaluation (ADA determination).

[0208] Pharmacokinetic (PK) analysis was performed using noncompartmental analysis with Phoenix™ WinNonlin® software (version 6.4, Certara LP (Satara)).

[0209] Pharmacokinetic Profile: All treated animals were exposed to the test article such that quantifiable amounts of RLI2 were measured over the majority of the post-dose sampling periods on Days 1 and 15. Key pharmacokinetic parameters are summarized in Table 8.

[0210] [Table 10]

[0211] C max and AUC 0-t Exposure to C differed between male and female animals. max and AUC 0-t was approximately 2-fold higher in females than in males. Independent of this gender difference, differences in the pharmacokinetics of RLI-15-HG and RLI-15-LG were also observed. Surprisingly, the exposure with RLI-15-HG was lower than that with RLI-15-LG. The ratio between RLI-15-HG and RLI-15-LG was C, independent of the sex of the animals. max and AUC 0-t The values ​​were 0.606 and 0.453, respectively.

[0212] 7. RLI2 AQProduction of immune cytokines based on Two RLI2AQ fusion proteins were fused to the C-terminus of the heavy chain of an anti-PD-1 antibody / IgG4 without a linker, or one RLI2AQ fusion protein was fused to one heavy chain (knob chain) using knob-in-hole technology (know-in-whole, KIH) with the HC knob mutation T366W and the HC hole chain mutations T366S / L368A / Y407V to generate immunocytokines. The anti-PD-1 antibody was pembrolizumab (PEM) with or without the Fc mutations shown in Table 9.

[0213] [Table 11]

[0214] The immunocytokines and controls in Table 9 were tested for their predicted stability by measuring their melting temperatures (Tm) using differential scanning fluorimetry (DSF). Differential scanning fluorimetry uses a real-time PCR instrument to monitor heat-induced protein denaturation by measuring the change in fluorescence of a dye that preferentially binds to unfolded proteins (e.g., Sypro Orange. Sypro Orange binds to hydrophobic regions of proteins exposed by unfolding, but water strongly quenches its fluorescence). This experiment is also known as a Protein Thermal Shift Assay because a shift in the apparent melting temperature can be measured upon addition of a stabilizing or destabilizing binding partner or buffer component. Briefly, SYPRO 50x pre-diluted in ultrapure water (UPW), protein sample and water are mixed to obtain a 25 μL reaction sample with a final protein concentration of 5-10 μM in SYPRO 5x. A negative control containing SYPRO diluted to a final concentration of 5x in UPW only, and the same mixture with lysozyme at a final concentration of 10 μM for the positive control are performed. 25 μL of each mixture are made in triplicate in a PCR plate and a specific program of thermal cycling is run. This program has been created to obtain the best possible resolution with our thermocycler. Melting curves are drawn from 20.0°C to 95.0°C in increments of 0.2°C every 20 seconds. No fluorescence signal should be measured in the negative control and only one peak should be detected at 70°C ± 1°C in the positive control. To determine buffer suitability, the same control is performed with buffer instead of UPW and the same results are expected. The derivative of the fluorescence vs. temperature curve is used to determine the Tm of the protein, defined as the temperature at which 50% of the protein sample is in the folded state and 50% is in the unfolded state.

number

[0215] A decrease in melting temperature of 1.5 °C was observed when the KIH mutation was present (60.1 °C vs. 61.6 °C for PEM WT). The KIH mutation on the Fc domain of pembrolizumab, without RLI coupling, induced a decrease in the stability of the antibody. A second melting temperature was observed, which was between 69 °C and 71 °C for all constructs. This Tm corresponds to the denaturation of a highly stable domain of the PEM antibody, as it is present in the RLI-uncoupled construct.

[0216] As expected, the IL-15 mutants did not affect the melting temperatures of the immunocytokines tested.

[0217] A significant decrease in Tm was observed as a function of the mutations present in the Fc of PEM. L (LE) mutations induced a decrease of 0.6°C to 1.8°C in Tm compared to the non-mutated construct, whereas Y (YTE) mutations induced a decrease of 5°C to 6.5°C. The double mutant LY seems to combine the effect of the two mutations, since the decrease was able to reach 7°C to 9°C compared to the non-mutated construct. The Tm decreased from 60°C in PEM-RLI N65A x1 to 52°C in PEM LY-RLI N65A x1 and from 61°C in the non-mutated PEM construct to 53°C in PEM LY-RLI N65A x2.

[0218] A rituximab-based immunocytokine was generated, RLI2, which has the same light chain (SEQ ID NO: 34) fused to both heavy chains with (SEQ ID NO: 32) or without (SEQ ID NO: 33) the L40 linker (SEQ ID NO: 31). AQ showed no significant biological differences when compared with those of the control (data not shown).

[0219] 8. PEM L-RLI NAx1 molecule enhances IFN-γ production in mixed lymphocyte reactions more than pembrolizumab PEM L-RLI N65A x1 was evaluated for its potential to enhance T cell activation and IFNγ production using a mixed lymphocyte reaction (MLR). MLR is an in vitro assay in which leukocytes from two genetically distinct individuals of the same species are co-cultured to result in cell blast transformation, DNA synthesis and proliferation. The generation of MLR occurs as a result of incompatibility of allogeneic determinants expressed on the surface of cell populations and encoded by the major histocompatibility complex (MHC). For this reaction, buffy coats from healthy donors were obtained. PBMCs were isolated by Ficoll Paque gradient and washed three times. hPBMC donor pairs were cultured with equimolar concentrations of 1 nM pembrolizumab and PEM L-RLI-NA x1 for 6 days. IFNγ production in cell supernatants was determined using the Human IFN-γ DuoSet ELISA (R&D systems, no. DY258B). Data are presented as relative response [%] of IFNγ production and represent the mean ± SEM from 12 paired hPBMC healthy donors.

[0220] Incubation of mismatched human PBMC donor pairs with PEM L-RLI NA x1 (1 nM) (IL-15 with N65A and AQ mutations) increased IFNγ production compared to equimolar amounts of pembrolizumab (see FIG. 6). Data represent the mean ± SE of 12 donor pairs for pembrolizumab and PEM L-RLI NAx1. These data suggest a mechanistic effect of PEM L-RLI NAx1 superior to pembrolizumab in stimulating IFN-γ from T cells.

[0221] 9. PEM-RLI NAx1 molecule shows antitumor efficacy in mouse tumor models The aim of this study was to preclinically evaluate the in vivo therapeutic efficacy of testing the construct PEM-RLI NAx1 and pembrolizumab as monotherapy in the treatment of HuCell MC38-hPD-L1 cell line implanted in female hPD1 single KI HuGEMM mice (n=8 mice / group). The mean tumor size was 108 mm on randomization day 0. 3 Treatment was initiated when tumors reached 18 days of age. PEM-RLI NA x1 (IL-15 with N65A and AQ mutations) was administered IV at 20 mg / kg on day 0 and pembrolizumab was administered IP at 5 mg / kg on days 0, 3, 6, and 9. Tumor observation was continued for 18 days (Figure 7A). Concurrently, PEM-RLI NA x1 (IL-15 with N65A and AQ mutations) was administered IV at 5, 10 on day 0. Tumor observation was continued for 6 days (Figure 7B).

[0222] A single dose of PEM-RLI NA x1 strongly reduced tumor volume in this model compared to the control untreated group (p-value was <0.05) and similar to multiple doses of pembrolizumab (see FIG. 7A). Tumor reduction with PEM-RLI NA x1 is also observed after a single dose of the lower 5 mg / kg dose compared to multiple doses of pembrolizumab (see FIG. 7B).

[0223] 10. IL-15 Muteins for Reduced In Vitro Potency To reduce the binding of the RLI conjugate to the IL-2Rβ and / or γ receptors and thus the in vitro potency, and to reduce the heterogeneity of RLI2-containing products, mutations were introduced into the IL-15 portion of the RLI2 conjugate. The indicated amino acid substitutions were made in the mature human IL-15 sequence (see Table 10).

[0224] [Table 12]

[0225] [Table 13]

[0226] The IL-15 substitutions tested that affect binding to IL-2 / IL-15Rβ and / or γ significantly reduced the potency of the RLI molecules against kit225 cells. The single mutant N65A caused the most significant reduction, but less than the triple mutant NQD (see Table 11). The other substitutions had only minor effects on potency.

[0227] [Table 14]

[0228] Even for RLI-15 mutant proteins tested without antibody binding, the NA mutation results in an approximately 2 log decrease in activity, measured here as EC50 in kit225 cells.

[0229] 11. IL-15 N65A mutation in PD-1-targeted immune cytokines shows reduced efficacy against kit225 cells Immunocytokines based on the anti-PD-1 antibody pembrolizumab were produced in various formats, including RLI molecules. Pembrolizumab is a humanized IgG4-κ antibody with a stabilizing S228P mutation in the Fc portion of the antibody. To improve this construct for use in immunocytokines, variations of pembrolizumab ("PEM") were tested. Although it is known that the IgG4 antibody class has relatively low ADCC activity, the L235E mutation (Alegre et al., 1992) ("LE" or short "L") was introduced to further reduce ADCC (SEQ ID NO: 28). To limit the potential for immunogenicity / anti-drug antibodies, more complex ADCC-inactivating mutations were avoided. One or two RLI2 molecules were genetically fused to the C-terminus of the PEM antibody. For homodimeric PEM variants ("x2"), one RLI2 molecule was fused to each heavy chain, whereas heterodimeric PEM variants ("x1") were generated using knob-in-hole (KiH) technology (Elliott et al., 2014), where one RLI2 molecule was fused to the knob heavy chain (SEQ ID NO: 26) with a T336W substitution that further had an L235E mutation to reduce ADCC activity, whereas the hole heavy chain (without RLI2 fusion) contained a T366S / L368A / Y407V substitution (SEQ ID NO: 27), which also further had an L235E mutation. When RLI2 was fused to the heavy chain, the terminal lysine (K) was deleted ("dK") to reduce product heterogeneity. Additionally, different RLI2 mutant proteins were used to fused to the heavy chain of the antibody. All RLI2 molecules have AQ (G78A / N79Q) substitutions to reduce product heterogeneity, and the following substitutions that reduce RLI2 binding to IL-2 / IL-15Rβγ were tested in the PEM-RLI immunocytokines: KAQD, DA, NA, ND, and NQD. The PEM-RLI immunocytokines that were made are listed in Table 13, left column. An exemplary PEM-RLI heterodimeric immunocytokine SOT201 is synthesized as shown in SEQ ID NO:22 (HC knob:IgG4 S228P.L235E.T366W.dK-RLI2.N162A.G175A.N176Q - RLI2 AQN162A), SEQ ID NO:23 (HC hole:S228P.L235E.T366S.L368A.Y407V), and SEQ ID NO:24 (LC).

[0230] Several homodimeric and heterodimeric forms of PEM-RLI2 with different IL-15 substitutions were prepared. AQ The potency of the immunocytokines was compared by measuring the in vitro EC50 in kit225 cells using RLI2 as a standard and set at 100% for relative potency (Table 13). EC50 was calculated using GraphPad Prism 8.4.3. The aim was to identify the least potent mutant protein of RLI-15 in kit225 cells. Results shown are the average of 2-5 experiments.

[0231] [Table 15] RLI:RLI2 AQ ND: not detected (limit of assay sensitivity)

[0232] RLI2 in 1x PEM-RLI-NA AQ NA was identified as the least potent RLI mutant protein with a single mutation that reduces IL-2 / IL-15Rβγ, which is still approximately 10-fold more active than the NQD mutant with three amino acid substitutions, thereby carrying a relatively high risk of immunogenicity.

[0233] 12. Evaluation of low potency PEM-RLI mutants bound to HC or LC in kit225 cells in vitro Several low potency IL-15 muteins in PEM-RLI immunocytokine with or without a mutated Fc antibody moiety (LE-YTE, or "LY" for short: LE refers to the Fc mutation L235E according to EU numbering of an IgG4 antibody to reduce the ADCC activity of the Fc domain, and YTE refers to the Fc mutation M252Y / S254T / T256E according to EU numbering, which have been reported to enhance FcRn binding and increase in vivo half-life) were compared for their potency compared to PEM LE / YTE-RLI NA x1 as a reference. In the "Lc" immunocytokine, the RLI conjugate was fused to the C-terminus of the light chain of the antibody without a linker (see SEQ ID NO:30 plus the indicated IL-15 substitutions DA, NA and DANA), while all other constructs have the RLI conjugate fused to one C-terminus of both heavy chains. In vitro potency testing was accomplished using kit225 cell line with a modified protocol (long term cell incubation). Potency of molecules was assessed as EC50 and also calculated as relative potency against PEM LE / YTE-RLI NA x1 molecule. Data represent the average of 2-4 experiments.

[0234] [Table 16]

[0235] Substitution combinations QDQA (Q101D / Q108A), NQD (D30N / E64Q / N65D), DANA (D61A / N65A) and DANAQD (D61A / N65A / Q101D) further reduced the potency of the PEM-RLI immunocytokine construct until it was not measurable for the DANAQD construct. Immunocytokines with an RLI conjugate fused to the light chain of the antibody showed similar potency compared to constructs with only one RLI conjugate with the same IL-15 mutations on one heavy chain of the antibody.

[0236] 13. Analysis of anti-drug antibodies in PEM-RLI immunocytokines in cynomolgus monkeys Cynomolgus monkeys were administered 0.3 mg / kg of the indicated PEM-RLI immunocytokine according to the scheme shown in Table 15.

[0237] [Table 17] For Group C: Administration PEM L-RLI2 NA x1 on day 1, PEM LY-RLI2 NA x1 on days 15, 22, and 29.

[0238] ADA titers were measured from serum collected on day 15 and determined by ELISA. Neutralizing antibodies were determined by FACS analysis of STAT5 phosphorylation by serum samples from tested PEM-RLI immunocytokines in kit225 cells.

[0239] [Table 18]

[0240] Since pembrolizumab is known to induce ADA in cynomolgus monkeys, it was not surprising that all monkeys produced ADA to the immunocytokines tested, and all monkeys expressed ADA that could be shown to be reactive to the antibody portion of the immunocytokine (αPEM column). AQ Both groups with NA monosubstitution (IV administration) had only one monkey expressing ADA to the RLI moiety, whereas RLI2 AQ All monkeys with triple NQD substitution developed ADA to RLI. Subcutaneous administration of PEM LY-RLI NA x1 immunocytokine also produced ADA in all monkeys.

[0241] It was further demonstrated that the generated ADA partially neutralized the immunocytokine and therefore potentially limited its therapeutic efficacy, especially for multiple administrations. Again, NQD triple substitution was inferior to NA single substitution (IV), whereas SC administration generated the highest amount of neutralizing antibodies.

[0242] Monkeys receiving PEM-RLI2 NQD (as well as monkeys receiving SC) showed a marked reduction in lymphocyte restimulation after day 29 compared to stimulation after days 1 and 15, whereas RLI2 AQ Expression of ADA and nAB against the RLI portion of the immunocytokine correlated with the pharmacodynamic response in these monkeys, since IV-administered immunocytokine with NA muteins still induced robust lymphoproliferation (data not shown). The loss of ability to be restimulated may be due to neutralizing antibodies.

[0243] 14. PEM-RLI NA x1 immunocytokine shows antitumor efficacy in mouse tumor models The in vivo therapeutic efficacy of PEM-RLI NA x1 immunocytokine was compared to pembrolizumab as monotherapy in the treatment of HuCell MC38-hPD-L1 tumor cell lines implanted in female hPD-1 single KI HuGEMM mice (n=8 mice / group). The mean tumor size was 108 mm on randomization day 0. 3 Treatment was initiated when tumor size reached 100%. PEM-RLI NA x1 was administered IV at 20 mg / kg on day 0 and pembrolizumab was administered IP at 5 mg / kg on days 0, 3, 6, and 9. Tumor observation was continued for 18 days. Concurrently, PEM-RLI NA x1 (IL-15 with N65A and AQ mutations) was administered IV at 5, 10 on day 0. Tumor observation was continued for 6 days.

[0244] PEM-RLI NA x1 strongly reduced tumor volume in this model compared to the control untreated group (p value was <0.05) and similar to the pembrolizumab treated group (see Figure 7). It should be noted that for immunocytokines, no significant difference with pembrolizumab was observed, but a single injection of immunocytokines achieved similar results as four doses of pembrolizumab. Moreover, a lower dose of 5 mg / kg was similarly efficient. Moreover, since mice are known to be about 10 times less sensitive to RLI, the full functionality of PEM-RLI NA x1 could not be tested in this mouse model, and therefore, the treatment effect in humans is expected to be better.

[0245] 15. ADCC activity of immunocytokines based on anti-Claudin 18.2 hCl1a antibodies with modified effector functions cell line Human cell lines PA-TU-8988S (Creative Bioarray, Cat. No. CSC-C0326) and A549 (ATCC CCL-185) overexpressing Claudin18.2 (A549-Cldn18.2) were grown in DMEM medium (Gibco) supplemented with 10% fetal bovine serum, 2 mM glutamine (GlutaMAX, Gibco), 100 U / ml penicillin, 0.1 mg / ml streptomycin (Invitrogen), and 2 μg / ml puromycin (Gibco). A549 cells were co-transfected by electroporation with a transposase expression construct (pcDNA3.1-hy-mPB), a construct carrying transposable full-length huCLDN18.2 (pPB-Puro-huCLDN18.2) together with a puromycin resistance cassette and a construct carrying EGFP as a transfection control (pEGFP-N3) (Waldmeier et al., 2016). After electroporation, cells were allowed to recover for 2 days in growth medium at 37 °C in a humidified incubator with a 5% CO2 atmosphere. Transfection was verified by FC analysis of EGFP expression. Cells expressing CLDN18.2 were then selected by adding puromycin at 1 μg / ml to the culture and further expanded to generate frozen stocks in FCS containing 10% DMSO. Expression of CLDN18.2 in transfected cells was analyzed by FC. To have a more homogenous PA-TU-8988S cell population, cells were sorted by FACS to select only cells with higher CLDN18.2 expression. Briefly, PA-TU-8988S cells suspended in FACS buffer (PBS, 2% FCS) were incubated with 2 μg / ml zolbetuximab for 30 min on ice. After washing in FACS buffer, cells were incubated with PE-labeled Fcγ-specific IgG goat anti-human secondary antibody (eBioscience) for 30 min on ice. After washing, stained cells were resuspended in FACS buffer and analyzed and sorted by FACSAria™ instrument to separate moderately expressing cells from high expressing cells. After sorting, collected PA-TU-8988S-High cells (PaTu) were resuspended in growth medium, expanded, and frozen aliquots were stored in liquid N2.

[0246] The human NK cell line NK92 (ATCC CRL-2407) exogenously expressing human CD16 (NK92-hCD16, referred to herein as NK92) was generated as described by Clemenceau et al. (2013). The cells were grown in RPMI1640 medium (Gibco) supplemented with 10% AB human serum (One Lambda), 2 mM glutamine (GlutaMAX, Gibco) and 5 ng / ml IL-2 (Peprotech). All cells were maintained at 37°C in a humidified atmosphere containing 5% CO2.

[0247] Human NK cells were isolated from fresh blood from healthy donors and diluted in a 1:1 ratio with cold PBS-EDTA, ph 7.4, and PBMCs were isolated by Ficoll-Paque gradient isolation. Isolated PBMCs were resuspended in complete culture medium. hNK cells were isolated from PBMCs using the EasySep Human NK Cell Isolation kit (Stem Cell Technologies, USA) according to the manufacturer's instructions. The isolated hNK cells from each donor were diluted at 3 × 10 6 The cells were resuspended in NK medium containing 10% serum at a concentration of cells / ml.

[0248] Cell-based ADCC assay A549-Cldn18.2 or PaTu cells were seeded in 96-well plates at the appropriate concentration (A549-Cldn18.2 - 20000 cells, PaTu - 30000 cells) and incubated for 24 hours. NK92 cells or isolated human NK cells were harvested by centrifugation, washed, and resuspended in ADCC assay medium (RPMI1640 (without phenol red) supplemented with 2 mM glutamine and 10% heat-inactivated (20 min at 56 °C) pooled complement human serum (Innovative Research)). Medium was removed from the 96-well plates containing adherent cells (target cells T) and NK92 cells in suspension in ADCC assay medium (effector cells E) were added to the adherent target cells at an E:T ratio of 10 for A549-Cldn18.2 and 5 for PA-TU-8988S cells. Antibodies or immunocytokines (ICKs) to be tested were added at concentrations ranging from 0.001 to 100 nM or 0.0001 to 10 μg / ml. A human IgG1 isotype antibody (Ultra-LEAF™ Purified Human IgG1 Isotype Control Recombinant Antibody, Biolegend, Cat. No. 403502) was included as a non-specific control. The mixture was incubated overnight at 37°C. After 24 hours, cytotoxicity, expressed as the activity of lactate dehydrogenase enzyme released from dead cells, was measured using the LDH Cytotoxicity Assay (Abcam, ab65393) according to the manufacturer's instructions. 10 μl of the supernatant was transferred to a new 96-well plate, mixed with LDH substrate, and the color change developed was measured at OD of 450 nm using a spectrophotometer.Cytotoxicity was calculated according to the following formula: Cytotoxicity (%) = ((Test sample - Effector cell control - Low control) / (High control - Low control)) x 100; "Test sample": effector / target mixture; "Effector cell control": one well with NK92 cells only (determines LDH activity released from effector cells); "Low control": one well with target cells only (determines spontaneous release of LDH activity from untreated target cells); "High control": one well with target cells permeabilized with lysis buffer (determines maximum releasable LDH activity).

[0249] FIG. 8 shows the ADCC activity of immunocytokines based on the hCl1a antibody with altered effector functions. All of the immunocytokines tested have a heterodimeric Fc domain and one RLI2 AQ The conjugate was fused to the C-terminus of one of the heavy chains. An exemplary immunocytokine for Claudin18.2 is SEQ ID NO: 35 (RLI2 AQ The conjugates are constructed from hCl1a heavy chain knob with AAA mutations fused to NA, SEQ ID NO: 36 (hCl1a heavy chain hole) and SEQ ID NO: 37 (hCl1a light chain). When immunocytokines with effector domain mutations that reduce ADCC were tested, the immunocytokine hCl1a LALAPG-RLI DANA showed nearly abolished ADCC activity when tested against A549-CLDN18.2 cells (upper panel) or PA-TU-8988S (lower panel) in the presence of NK92 cells, when compared to the hCl1a-DANA immunocytokine with the hCl1a antibody alone. The hCl1a-LALA antibody also showed reduced ADCC activity compared to the hCl1a antibody, however, ADCC activity was not completely abolished. The addition of the conjugate did not affect the ADCC activity of the immunocytokines, where the ADCC activity was reduced, when compared to the ADCC activity of the antibody alone. Table 17 summarizes the ADCC EC50 values ​​measured for each immunocytokine or antibody tested. EC 50Values ​​were determined using Graphpad Prism Software with the built-in "log(AGONIST) vs. response - variable slope (four parameters)" EC50 determination. When immunocytokines with effector domain mutations that enhance ADCC were tested, all tested immunocytokines based on hCl1a antibodies with DLE, DE, AAA, TE or IE mutations in the Fc domain showed enhanced ADCC activity compared to the same immunocytokines or antibodies alone without those mutations (Figure 9). To enhance ADCC activity, afucosylation was also tested. Figure 9F shows that in A549-Cldn18.2 and PA-TU-8988S cells, the defucosylated immunocytokine hCl1a-DANA afuc has enhanced ADCC activity compared to hCl1a-DANA, and ADCC activity comparable to that of the immunocytokines with the DE and DLE mutations mentioned above. However, when defucosylation was combined with effector domain-enhancing mutations, defucosylation surprisingly had a negative effect on the ADCC enhancement induced by the DE or DLE mutations (see Figures 9B and A). Nevertheless, when defucosylation was combined with AAA mutations, the enhanced ADCC activity was maintained (Figure 9C).

[0250] [Table 19]

[0251] 16. Assessment of antibody Fc binding to the ADCC activating receptors FcγRIIIa V158 and FcγRIIIa F158 and the ADCC inhibitory receptor FcγRIIb by surface plasmon resonance (SPR) The human FcγRIIIa receptor (hFcγRIIIa; CD16a) exists as two polymorphic variants at position 158, hFcγRIIIaV158 and hFcγRIIIaF158. FcγRIIIa activates ADCC activity, whereas FcγRIIb inhibits ADCC. The ADCC activity of immunocytokines can be expressed as the ratio of the EC50 binding affinity to FcγRIIIa to the EC50 binding affinity to FcγRIIb when their affinity to the receptor is measured by SPR. SPR experiments were performed on a Biacore 8K (Cytiva, Chicago, IL, USA) using a CM5 sensor chip (Cytiva) immobilized with THE His-tag antibody (Genscript). FcγRIIIa V158, FcγRIIIa F158 or FcγRIIb protein was used for capture in 1×HBS-EP+ running buffer at a flow rate of 10 μl / min with a contact time of 30 s. The association / dissociation rate was measured for each immunocytokine tested in the appropriate range of concentration serial dilutions using association / dissociation times of 300 s / 300 s at a flow rate of 30 μl / min, except for constructs with DLE and DE with or without defucosylation. For constructs with DLE and DE, association / dissociation times of 120 s / 1200 s were applied. Table 18 below summarizes the results of the SPR measurements.

[0252] [Table 20] A / I ratio=(affinity for FcgRIIIa) / (affinity for FcgRIIb). Affinity=1 / Kd. "afuc" refers to defucosylation.

[0253] The A / I ratio allows one to assess the strength of binding to an ADCC activating receptor ("A"; FcγRIII) compared to an ADCC inhibitory receptor ("B"; FcγRIIb). The higher this ratio, the stronger the binding of the antibody or immunocytokine to the activating receptor.

[0254] Overall, the SPR data confirm that all immunocytokines with ADCC-enhancing mutations exhibit higher A / I ratios than immunocytokines without ADCC-enhancing mutations that are part of the TL mutations. The relatively low A / I ratios of the TL mutations may be due to increased glycosylation of such mutations (see Example 17).

[0255] 17. Stability / developability of hCl1a-based immunocytokines with enhanced ADCC activity Immunocytokines based on hCl1a that have ADCC-enhancing DLE, DE, AAA, TL or IE mutations or are defucosylated are administered to patients with ADCC. H The stability and developability of the two domains were assessed by evaluating their melting temperature, sequence liability and glycosylation (N-glycan) profile.

[0256] C H The melting temperatures of the two domains were measured by differential scanning calorimetry (DSC) using a MicroCal PEAQ-DSC automated system (Malvern Panalytical). Briefly, the immunocytokine samples were diluted to 1 mg / ml in their storage buffer. Heating was performed from 20°C to 100°C at a rate of 1°C / min. The protein solution was then cooled in situ and the same thermal scan was performed to obtain a baseline for subtraction from the first scan. For N-glycan analysis, proteins were first reduced with DTT and then transferred to an HPLC column equipped with a glass insert vial for injection. Proteins were separated by reversed-phase chromatography and detected by a Waters / XEVO G2XS-QTOF online LC-MS coupled with a UV detector. The molecular weights of the detected glycan chains were matched with known N-glycan types, and the relative abundance of N-glycans was calculated and expressed by the intensity of the detected peaks. The amino acid sequences of the immunocytokine constructs carrying ADCC enhancing mutations were analyzed for the presence of the following additional sequence instabilities (not present in constructs without ADCC enhancing mutations), as listed in Table 19.

[0257] [Table 21]

[0258] The TL mutation introduced N-glycosylation sequence instability (mutation K392T adjacent to N390 in the IgG1 sequence). No sequence instability was introduced by the other mutations (see Table 20).

[0259] [Table 22] Score 4: the parameter is within the expected range for a mAb-based drug product; Score 3: careful monitoring / evaluation of required quality attributes during development; Score 2: Likely to have a significant impact on timelines and / or costs; Score 1: High risk of not being adequately controlled.

[0260] Overall, defucosylation does not affect the stability and developability and may therefore be used to enhance the ADCC activity of immunocytokines. HThe TL mutations caused a significant decrease in the melting temperature of the 2 domains (see Table 21), with the potential to affect the stability of the immunocytokines in solution. However, these mutations did not affect the glycosylation of the immunocytokines. The sequence instability introduced by the TL mutations resulted in the introduction of undesirable sialylated and high mannose glycan species (see Table 22). These species could adversely affect the pK of the immunocytokines. Similarly, the immunocytokines with IE mutations had a high proportion of mannose species, which had the potential to affect the properties of those immunocytokines. The immunocytokines with AAA mutations resulted in an increase in mannose species (see Table 22). However, the generation of defucosylated immunocytokines partially restored glycosylation to an acceptable level for developability. Therefore, if enhancement of an hCl1a-based immunocytokine is desired, the AAA mutation, optionally in combination with defucosylation, may be the recommended mutation with the least impact on its stability and developability. Defucosylation did not affect the properties evaluated. The DLE and DE mutations caused a significant decrease in Tm and had the potential to destabilize the molecule. The TL mutation introduced an additional glycosylation site into the Fc. The construct with the IE mutation had a high proportion of mannose species.

[0261] [Table 23]

[0262] [Table 24]

[0263] 18. In vivo efficacy testing of Claudin18.2 immunocytokines in mice The aim of this study is to test the in vivo therapeutic efficacy of hCl1a-RLI immunocytokine in a mouse model. Female NMRI nude mice are implanted with pancreatic human cell line-derived xenografts BXPC3 (ATCC CRL-1687™) exogenously expressing Claudin18.2 (BXPC3-CLDN18.2) at 5-7 weeks of age. Tumors are implanted by unilateral subcutaneous injection. The animals are cultured at approximately 100 mm 3 Mice are randomized based on tumor volume in mm. Mice are assigned to different groups (n=7 per group) and treated according to Table 23 on day 1. Animals are checked twice weekly for weight loss and tumor volume. Tumor volumes are measured by caliper and expressed in mm using the formula: V=(L×W×W) / 2, where V is tumor volume, L is tumor length (longest tumor dimension), and W is tumor width (longest tumor dimension perpendicular to L). 3 It is expressed as 2000mm. 3 Mice are euthanized when they reach a tumor burden of 0.05 mg / kg or experience significant weight loss (>30% overall or >20% on two consecutive days).

[0264] [Table 25]

[0265] 19. Anti-PD-1 antibody and SOT201 inhibit CD8 + Synergize in T cell activation SOT201 binds RLI-15 at the C-terminus of the knob heavy chain. AQASOT201 is a heterodimeric immunocytokine having an antibody derived from humanized IgG4 pembrolizumab with heavy chain T366W-knob / T366S, L368A, Y407V-hole, L235E substitutions and terminal K deleted (see SEQ ID NO:22, SEQ ID NO:38, SEQ ID NO:24) fused to SOT201. SOT201 and Keytruda® (pembrolizumab) were compared in a PD-1 / PD-L1 blocking assay according to Example 1. Figure 10A shows that SOT201 effectively blocks PD-1 / PD-L1 interaction similar to the anti-PD-1 antibody Keytruda. The K determined for SOT201 and pembrolizumab D The values ​​are shown in Table 24.

[0266] [Table 26]

[0267] Human PBMCs from 11 healthy donors were cultured using RLI2AQ N65A (RLI-15 AQA ) variant, or RLI2, which has the same antibody heavy and light chains as SOT201 but does not have reduced binding of the IL-15 moiety to IL-2 / IL-15Rβγ. AQ The control molecule with the variant ("SOT201 wt") was stimulated in vitro for 7 days. Cell proliferation was assessed by flow cytometry analysis using Ki-67 and + NK cells and CD8 + SOT201 was determined to activate NK cells and CD8 T cells at a higher EC50 concentration than a comparable immunocytokine molecule (SOT201 wt) with unreduced receptor binding to the RLI-15 molecule. + Activates T cell proliferation (Figure 10B). RLI-15 AQAThe murine surrogate SOT201 (mSOT201: see SEQ ID NO: 39, SEQ ID NO: 40 and SEQ ID NO: 41) containing the anti-mouse PD-1 antibody RMP1-14 (BioXCell, Lebanon, NH, USA) with similar substitutions for heterodimerization (E356K, N399K / K409E, K439D), ADCC silencing (D265A) and stabilization (dK) fused to the monoclonal anti-mouse PD-1 antibody RMP1-14 itself (mPD1) and RLI-15, which has a similar in vivo half-life as mSOT201. AQA Anti-human PD1 mouse IgG1-RLI-15 as a control AQA The results were compared to a single activity control represented by (hPD1-mSOT201), which does not exert any PD-1 blocking activity in C57BL / 6 mice. Cell proliferation (Ki67) was detected in the spleen by flow cytometry 5 days after IV injection of the compound equimolar to 5 mg / kg mSOT201 in healthy C57BL / 6 mice (n=2 / group). Anti-PD-1 antibody and RLI-15 in mouse surrogate mSOT201 AQA The mutant protein portion is CD8 + It showed a synergistic effect on T cell proliferation (Figure 10C).

[0268] 20. Tumor regression in the MC38 mouse model C57BL / 6 mice (hPD1 transgenic) were implanted with the syngeneic MC38 cell line. The test drugs mSOT201, hPD1-mSOT201, and mPD1 were administered on day 1 (randomization day, tumor volume 80-100 mm 3) were injected IV at equimolar doses with 5 mg / kg mSOT201 (n=10 / group) and compared with control (NaCl). mSOT201 induced tumor regression in 9 out of 10 mice after a single IV administration, whereas in comparison, monoclonal anti-mouse PD-1 antibody (mPD1) and anti-human PD-1 mouse IgG1-RLI-15 mutant protein immunocytokine (hPD1-mSOT201), which exerts no anti-PD-1 effect in mice, only showed a marginal effect on tumor growth compared with control mice (Figure 11A). Similarly, anti-mouse PD-1 antibody alone (mPD1) or RLI-15 AQA Anti-mouse PD-1 antibody and RLI-15 in a fusion protein (mSOT201) compared to anti-human PD1 mouse IgG1-RLI-15 mutant protein immunocytokine (hPD1-mSOT201) as a control for the mutant protein alone AQA The synergistic activity of the mutant proteins is demonstrated in surviving mice over a time course up to 100 days after treatment (FIG. 11B).

[0269] 21. Induction of pathways and genes related to antitumor immunity in MC38 tumors and activation of immune cells in the spleen and lymph nodes RNA isolation: RNA samples were isolated from tumors in syngeneic MC38 tumor-bearing C57BL / 6 mice 7 days after a single IV dose of mSOT201 (5 mg / kg). Day 1 (randomization day, tumor volume 80–100 mm 3 Three mice were treated IV with mSOT201 (5 mg / kg) and four control mice were left untreated. RNA was isolated from tumor tissues using an RNeasy MicroKit. The quality of the RNA samples was checked using an Agilent Bioanalyzer RNA Nano Chip and Qubit HS RNA assay. RNA sequence analysis: Sequencing libraries were prepared from RNA samples by SMARTer® Stranded Total RNA-Seq Kit v3-Pico Input Mammalian Kit (Takara Bio USA, Inc.), library quality control was performed using a capillary gel electrophoresis system (Agilent Bioanalyzer with HS DNA chip) and Qubit HS DNA assay, and sequencing was performed on a NovaSeq 6000 with a 2x151bp run using the NovaSeq 6000 300 cycles Reagent Kit. Data analysis: Raw data were processed according to a standard RNA-seq pipeline including the following steps: quality control (by FastQC and FastqScreen), adapter trimming (8bp trimming in Read2 by using seqtk), mapping to the reference genome GRCm39 (using HISAT2) and transcript counting (using ht-seq). The resulting output, a quantification file containing the number of transcripts for each sample, was further processed via R packages and ggplot2, tydiverse, dplyr. Raw counts were normalized to median ratio normalization by DESeq2. Differential gene expression analysis was performed using DESeq2 (version 1.24.0) in R (abs(log2FC)=1, FDR<0.05). Heatmaps were generated using the ComplexHeatmap package in R. Analysis of the function and enrichment of DEGs was performed using ClusterProfiler and the web-based tool Gene ontology (GO). To calculate TPM values ​​for cell population analysis, the salmon tool was used on the trimmed fastq files. Cell population analysis was performed with TIMER 2.0 and xCell tools. Results: Differential expression analysis (abs(log2FC)=1, FDR<0.05) resulted in the upregulation of 800 mouse genes and the downregulation of 1910 mouse genes in mSOT201-treated tumors compared to control samples. Gene Ontology (GO) term enrichment analysis identified upregulated DEGs mainly related to αβ T cells, γδ T cells, B cells, NK cell activation, cytotoxicity, cell killing, cytokine production, cell chemotaxis and cell adhesion, whereas downregulated genes were related to tumor development and tumor signaling. These data indicate that mSOT201 activates both innate and adaptive immunity in the tumor microenvironment. Next, we employed "metagene" markers to estimate the relative abundance of different immune cell populations in the tumor microenvironment. In accordance with the whole transcriptome findings, mSOT201-treated samples were analyzed using a 3-fold analysis of the CD8 + T cells (p<0.001), CD8 + Naive T cells (p<0.0005), CD8 + Effector memory T cells (p=0.001), CD8 + Central memory T cells (p<0.001), γδ T cells (p=0.0002), NK cells (p<0.001), CD4 + T cells (p=0.0157), CD4 + Naive T cells (p=0.1176), CD4 + Gene sets associated with effector memory T cells (p=0.003), B cells (p=0.0602), and myeloid dendritic cells (p=0.0120) were enriched, whereas gene sets associated with cancer-associated fibroblasts were significantly depleted (p=0.0254) (Figure 12A). mSOT201 induced the proliferation of selected immune cell populations in the spleen and lymph nodes of MC38 tumor-bearing mice (Figure 12B). 3 ) On day 7 after mSOT201 treatment, cell proliferation (Ki67) was detected by flow cytometry (n=2).

[0270] 22. EC50 values ​​of different IL2 / IL-15Rβγ agonists in kit225 cells RLI-15(SOT101), SOT201(PEM-RLI-15 AQA ), hPD-1-IL-2v, and αhPD1-IL-15m M1 EC50 values ​​were determined as described in Example 1. In hPD-1-IL-2v, one IL-2 mutein IL-2v (SEQ ID NO: 43) is fused to the C-terminus of one heavy chain of an anti-human PD-1 antibody as described in WO 2018 / 184964 A1 (sequences of SEQ ID NO: 22, SEQ ID NO: 23, and SEQ ID NO: 25 in WO 2018 / 184964 A1). In αhPD1-IL-15m M1, one IL-15 mutein with mutations N1A-D30N-E46G-V49R (SEQ ID NO: 44) is fused to the C-terminus of one heavy chain of an anti-human PD-1 antibody as described in WO 2019 / 166946 A1 (see FIG. 1D therein, SEQ ID NO: 89, SEQ ID NO: 74 and SEQ ID NO: 65 therein). The EC50 values ​​are shown in Table 25.

[0271] [Table 27]

[0272] A further interesting candidate to be tested is αhPD1-IL-15m M2 in which one IL-15 mutein with the mutations N1G-D30N-E46G-V49R-E64Q (SEQ ID NO: 45) is fused to the C-terminus of one heavy chain of the anti-human PD-1 antibody described in WO 2019 / 166946 A1 (see FIG. 1C therein, SEQ ID NO: 90, SEQ ID NO: 74 and SEQ ID NO: 65 therein). Thus, SOT201 has a substantially lower EC50 than PD1-IL-2v and αhPD1-IL-15m M1 in kit225 cells, allowing for higher doses and a longer in vivo half-life, and is also expected to exert a stronger and longer lasting effect with respect to activity in disrupting the anti-PD-1 / PD-L1 interaction.

[0273] 23. Comparison of mSOT201 and mPD1-IL-2Rβγ agonist in MC38 tumor model mSOT201 (a mouse SOT201 surrogate) was compared with control (NaCl), anti-mouse PD-1 antibody fused to IL-2v IL-2 mutant protein RMP1-14 (mPD1-IL-2Rβγ agonist), and RLI-15 AQA The combination of mSOT201 and mPD1 antibody was compared in the MC38 tumor model with a single IV dose as described in Example 20. Administration of mPD1-IL-2Rβγ significantly increased NK cell and CD8+ counts on day 5 after IV administration of 5 mg / kg mSOT201 in healthy C57 / BL6 mice. + The dose was chosen to match T cell proliferation and was equivalent to 0.25 mg / kg mPD1-IL-2Rβγ. + ) was detected by flow cytometry. + It induced activation of T cells and NK cells that, in contrast to the mPD1-IL-2Rβγ agonist, persisted until day 8 (Figure 13B). mPD1-IL-2Rβγ is an IL-2 / IL-15Rβγ agonist in which the IL-2 mutein IL-2v (SEQ ID NO: 43) contains substitutions F42A, Y45A and L72G to the IL-2 sequence that reduce affinity for IL-2Rα (see WO 2018 / 184964A1, e.g., the paragraph spanning pages 27 and 28), as well as further substitutions, T3A to eliminate O-glycosylation at position 3 (paragraph spanning pages 28 and 29), and C125A to increase expression or stability (page 30, 3rd paragraph). A murine surrogate of SOT201 (mSOT201) induced tumor regression in 9 of 10 MC38 tumor-bearing mice after a single IV administration compared with 5 of 10 for the mPD1-IL-2Rβγ agonist, but not RLI-15. AQA The combination of IgG1 with mPD1 antibody only resulted in a delay in tumor growth compared to control mice (Figure 13A). mSOT201 inhibited NK cells and CD8 +Induced T cell proliferation, which was observed with the mPD1-IL-2Rβγ agonist and equimolar amounts of RLI-15 in combination with mPD1. AQA In contrast to the MC38 tumors, treatment lasted for 7 days after randomization. 3 (n=10 / group). Furthermore, mSOT201 still showed persistent CD8+ proliferation at day 8, in contrast to the mPD1-IL-2Rβγ agonist, which showed a marked reduction in proliferating cells at day 8. + It induced a significant and prolonged activation of T cells and NK cells (Figure 13B). SOT201 also inhibited NK cells and CD8 + Induce T cell proliferation, which is achieved by combining mPD1-IL-2v and an equimolar amount of RLI-15. AQA and mPD1 antibody combination, persisted for 7 days after administration (Figure 13C).

[0274] 24. PK profile of SOT201 in cynomolgus monkeys SOT201 was administered IV at 0.6 mg / kg on day 1 to cynomolgus monkeys, and NK cell and CD8 + T cell proliferation (Ki67 + ) and absolute cell counts were determined over time by flow cytometry and hematology. SOT201 significantly increased NK (approximately 90% at day 5) and CD8 cell counts in the blood of cynomolgus monkeys after IV administration. + It induced high proliferation and expansion of T cells (approximately 80% on day 5) ( FIG. 14A ). Pharmacokinetic parameters are shown in Table 26.

[0275] [Table 28]

[0276] SOT201 increased NK cells and CD8 cells after repeated IV administration in cynomolgus monkeys. + induced T cell activation (Figure 14B).

[0277] 25. PD activity of mouse SOT201 surrogates The primary objective of this study was to determine whether treatment with the murine surrogate molecule mSOT201 (see Example 19) reduced CD8 expression in C57BL / 6 mice compared with treatment with hPD1-mSOT201 or mPD-1. + The aim of the study was to evaluate whether mSOT201 wt murine surrogate molecule has additive / synergistic effects on T cell proliferation. The secondary objective of the study was to compare the pharmacodynamic activity of the murine surrogate molecule mPD1-IL2v in C57BL / 6 mice. A description of the murine surrogate molecules tested is provided in Table 27. PD activity was assessed on days 5 and 8. FACS analysis was performed as described above.

[0278] [Table 29]

[0279] [Table 30]

[0280] Because pembrolizumab does not recognize mouse PD-1, hPD-1-mSOT201 was compared with RLI-15 bound to a non-conjugated antibody with a similar PK profile. AQA and therefore has such a PK profile. AQA The mPD-1 molecule reflects the PD activity of the anti-PD-1 antibody alone. CD8 + With respect to T cell activation, mSOT201 exhibits more than additive (i.e., synergistic) effects on days 5 and even 8 compared to its single component surrogates hPD1-mSOT201 and mPD-1, administered at equimolar amounts. In comparison, both mPD1-IL2v and mSOT201 wt (both with more active IL-2 / IL-15Rβγ agonists), administered at lower doses to account for the anticipated higher activity on day 5, exhibited greater CD8 T cell activation on day 5. + Although it shows slightly higher activation of T cells, such effects are short-lived and do not affect CD8 +Activation of T cells is much stronger for mSOT201 on day 8. Looking at activated NK cells, the difference is less pronounced. As expected, mPD-1 does not activate NK cells, but hPD1-mSOT201, mPD1-IL2v, mSOT201 and mSOT201 wt strongly activate NK cells on day 5, with mSOT201 being somewhat weaker than the others. On day 8, mSOT201 again shows stronger activation of NK cells compared to mPD1-IL2v and mSOT201 wt (Figure 15A). Although mSOT201, hPD1-mSOT201, and mPD-1 were administered at twice the doses as in A, similar behavior was observed when mSOT201 wt and mPD1-IL2v were administered at lower doses, likely because the cells had already reached maximum activation in experiment A (see FIG. 15B). As expected, mSOT201 wt and mPD1-IL2v suppressed the activation of CD8 + We showed a decrease in activation of both T cells and NK cells, which was mediated by CD8 + For T cells, there was again a decline towards control levels by day 8. These data suggest that SOT201, which has significantly reduced binding to IL-2 / IL-15Rβγ, together with its anti-PD-1 moiety, inhibits NK and CD8 + It is a potent and long-lasting activator of T cells, whereas molecules with higher IL-2 / IL-15Rβγ agonistic activity are specifically targeted to CD8 + This shows that the activation of T cells is much shorter than that of the control. + T cells) or in trans (i.e., in close proximity to different CD8 + PD-1 expression on CD8 T cells + The avidity effect of simultaneous binding of PD-1 and IL-2 / IL-15Rβγ on T cells is that CD8 + It is hypothesized that this results in such preferential activation of T cells.

[0281] 26. Antitumor efficacy activity of mSOT201 in PD-1 sensitive and PD-1 treatment resistant mouse models The purpose of this study was to evaluate the antitumor activity of mSOT201 in anti-PD-1 treatment-sensitive (CT26, MC38) and anti-PD-1 treatment-resistant (B16F10, CT26 STK11 ko) mouse models. Descriptions of the mouse surrogate molecules tested are listed in Table 29.

[0282] [Table 31]

[0283] The murine surrogate molecule of SOT201, mSOT201, showed synergistic effects compared to its single component surrogates mPD-1 and hPD1-mSOT201 in the PD-1-sensitive tumor models tested, CT26 and MC38, showing complete responses in 5 of 10 and 9 of 10 mice, respectively (Figure 16A). Even in tumor models known to be resistant to anti-PD-1 therapy, mSOT201 showed synergistic effects compared with its single components, but the therapeutic effect was not as strong as in the sensitive models, with only 1 out of 10 mice showing a complete response for the B16F10 model (Figure 16B).

[0284] 27. mSOT201 vs. RLI-15 AQA Antitumor efficacy of mutant protein + anti-PD-1 antibody The objective of this study was to evaluate the antitumor activity of mSOT201 versus RLI-15 AQA mutant protein + anti-PD-1 treatment in the MC38 mouse model. A description of the mouse surrogate molecules tested is provided in Table 30.

[0285] [Table 32]

[0286] Anti-PD-1 moiety and IL-2 / IL-15βγ agonist RLI-15 AQA (2 doses, G2 and G3) were fused with a combination of equimolar amounts of each component (G4: RLI-15 AQAThe results showed a strong synergistic effect compared to G11:hPD1-mSOT201+mPD1, or G11:hPD1-mSOT201+mPD1. See Figure 17. It is hypothesized that the temporal and spatial linkage of activation of PD-1 positive immune cells is mechanistically stronger than the activation of immune cells by individual components.

[0287] 28. Antitumor efficacy activity of mSOT201 vs. SOT101 + anti-PD-1 antibody The objective of this study was to evaluate the antitumor activity of mSOT201 versus SOT101 + anti-PD-1 treatment in the MC38 mouse model. Descriptions of the mouse surrogate molecules tested are listed in Table 31.

[0288] [Table 33]

[0289] A single dose of 2 mg / kg mSOT201 (G3) was associated with an RLI of 1 mg / kg AQ + 4 doses of mPD1 at 5 mg / kg (G8) or 1 mg / kg of RLI2 AQ The combination therapy showed similar therapeutic efficacy as four doses of mSOT201 at 5 mg / kg + four doses of mPD1 at 5 mg / kg (G9). However, a single dose of mSOT201 at 5 mg / kg (G2) outperformed multiple doses of the individual components (G8 and G9) (see Figure 18).

[0290] 29. Mechanistic study of differences in immune cell activation under mSOT201 vs. SOT101 + anti-PD-1 antibody treatment The purpose of this study was to evaluate the antitumor activity of similar effective doses of mSOT201 versus SOT101 + anti-PD-1 treatment in the MC38 mouse model. Descriptions of the mouse surrogate molecules tested are listed in Table 31.

[0291] Differences in the relative numbers of various immune cell populations in both treatments were detected in the tumor, spleen and lymph nodes. + T cells and αβTCR-bearing CD3 +The relative proliferation of cells did not change between both treatments in the spleen and lymph nodes. However, in the tumor, mSOT201 significantly increased the proliferation of CD8 + Induced a higher relative increase in T cells, but RLI2 AQ + anti-PD-1 treatment increased more NK cells. Interestingly, mSOT201 increased the percentage of γδTCR-bearing CD3 + Cells were induced but not RLI2 AQ + Anti-PD-1 combination treatment primarily stimulated a higher proportion of γδTCR-bearing CD3 + The cells were induced (see FIG. 19).

[0292] 30. DC-T cell-based assays and Fluorospot assays to determine immunogenicity DC-T cell-based assays to determine immunogenicity Buffy coats were obtained from healthy donors. The blood was diluted with PBS-EDTA (to give 175 mL of diluted blood) and PBMCs were isolated by Ficoll Paque gradient (15 mL of Ficoll + 35 mL of diluted blood). CD14 + Monocytes were isolated using the EasySep™ Human CD14 Positive Selection Kit II (17858, StemCell) according to the manufacturer's instructions. - The fraction was pipetted into a new Falcon tube and the remainder was centrifuged at 1200 rpm for 10 min, then resuspended in CryoStore medium, frozen and temporarily stored at -80 °C. + Monocytes were resuspended in DC medium (CellGro supplemented with IL-4 and GM-CSF). Cells were incubated at 37°C with 5% CO2 for 5 days, harvested and seeded in 48-well plates. iDCs were loaded with protein for 4 hours and matured overnight with a cytokine cocktail (TNF-α, IL-1β + IL-4 and GM-CSF). They were then washed 4 times with PBS and T cell medium. Cells were stained with autologous CFSE stained CD4 +They were co-cultured with T cells at a ratio of 1:10 (negative magnetic separation) and cultured for 7 days. CFSE dilution was detected by flow cytometry.

[0293] [Table 34]

[0294] The aim of this study was to evaluate in vitro the immunogenic risk of a pembrolizumab-based immunocytokine carrying one RLI-15 mutant protein (PEM-RLI-15 candidate molecule). A DC-T cell assay was used for this purpose, in which the test product was first incubated with immature dendritic cells (iDCs) and later presented to autologous T cells as processed peptides of the candidate molecule loaded on MHC molecules of mature DCs (mDCs). After a 7-day co-incubation period, T cell proliferation was measured as a surrogate marker of anti-drug antibody formation. The detection of DC-induced T cell proliferation was used to mitigate the stimulatory activity of the RLI-15 components in the test system, which could have a strong influence on the results, but this strong influence is not due to immunogenicity. Keyhole limpet hemocyanin (KLH) was used as a positive control, since KLH is known to induce a strong immune response induction. Pembrolizumab was used as a negative control. Control DCs that were not loaded with protein were used as a control for the assessment of non-specific T cell proliferation.

[0295] [Table 35]

[0296] PEM-RLI-15 candidate molecules from Table 32 were used at two concentrations each for stimulation of iDCs. Maturation of DCs was induced by proinflammatory cytokines. After 24 hours, mDCs were washed and incubated with autologous CD4+ pre-stained with CFSE. + T cells were incubated with 100% T cells, and T cell proliferation was assessed after 7 days based on CFSE detection by flow cytometry.

[0297] This assay could not be performed with SOT201 (PEM L-RLI N65A x1) because the still too high activity of the RLI N65A mutant protein would result in direct T cell activation and spill over to RLI-15 activity.

[0298] Human CD14 + DCs generated from monocytes (11 healthy donors from three separate experiments) were incubated with 10 μg / ml (not shown) or 50 μg / ml of PEM-RLI-15 candidate molecules, pembrolizumab or KLH for 24 h in the presence of maturation signals (proinflammatory cytokines TNFα and IL-1β). Protein-loaded washed mDCs were subsequently cultured with autologous CFSE-stained CD4 + T cells were cultured together. T cell proliferation was measured by flow cytometry after 7 days. CFSE low Considering cells as going through the cell cycle, proliferating CD4 + The percentage of T cells was evaluated based on the CFSE signal. KLH was used as a positive control and pembrolizumab was used as a negative control (see FIG. 20A). The PEM-RLI-15 candidate molecule PEM L-RLI DANA x1 / PEM LY-RLI DANA x1 did not induce significant proliferation of T cells compared to the negative control, reflecting a low immunogenicity risk (positive response detected in 1 of 11 donors). The candidate molecule PEM LY-RLI DANAQD x1 induced significant proliferation of T cells compared to the negative control (p=0.0208, paired t-test), pointing out a potential immunogenicity risk for this RLI-15 mutant protein with three mutations to reduce binding to IL-2 / IL-15Rβγ (positive response detected in 4 of 11 donors).

[0299] FluoroSpot assay for determining immunogenicity Since the overactive RLI-15 mutant protein stimulates an immune response, DC-T cell assays showed that RLI-15 (wild-type sequence) is more potent than RLI-15. AQATherefore, pairs of peptides in which substitutions were introduced were generated across the substitutions and tested in the Fluorospot assay.

[0300] [Table 36]

[0301] PBMCs isolated from each of 40 healthy donors were retrieved from cryogenic storage and thawed in culture medium. CD8 + Cells were depleted from PBMCs using negative bead selection. CD8-depleted PBMCs were seeded on cell culture plates in RPMI + 10% huAB serum, then pulsed with pooled test peptides and further cultured in cytokine-supplemented medium. After overnight culture and on day 4 of culture, medium was refreshed with the supportive cytokines IL-7 and IL-2. After 7 days of culture, enriched CD8-depleted PBMCs were harvested and left overnight. On day 8, PBMCs were seeded and restimulated on IFN-γ / TNF-α FluoroSpot plates in the presence or absence of peptide pools and control molecules. After overnight incubation, T cell activation was assessed by measuring IFN-γ and TNF-α using a Mabtech IRIS FluoroSpot Reader. Figure 20B shows that the confidence intervals for all tested conditions overlap 0, meaning that there is no evidence of a shift in the average dSFU when comparing the mutant peptides with the paired wild-type sequences, and therefore no associated increase in immunogenicity is seen for both the N65A and G175A / N176Q substitution pairs.

[0302] 31. Efficacy of different anti-PD-1 IL-2 / IL-15Rβγ agonist immunocytokines The following anti-PD-1 IL-2 / IL-15Rβγ agonist immunocytokines (Table 34) were generated and their activities compared.

[0303] [Table 37]

[0304] The efficacy of anti-PD-1 IL-2 / IL-15Rβγ agonist immunocytokines was determined in kit225 cells (see Table 35) and hPBMCs (see Table 36).

[0305] [Table 38]

[0306] [Table 39]

[0307] 32. PD-1 / PD-L1 Blocking Activity of Anti-PD-1 IL-2 / IL-15Rβγ Agonist Immunocytokines To evaluate the blocking activity of the PD-1 / PD-L1 axis, the anti-PD-1 IL-2 / IL-15Rβγ agonist immunocytokines were tested using the PD-1 / PD-L1 Blockade Bioassay (Promega, No. J1250) as described above. The results are shown in Table 37.

[0308] [Table 40]

[0309] SOT201 exhibits the highest PD-1 / PD-L1 blocking activity among the three tested anti-PD-1 IL-2 / IL-15Rβγ agonist immunocytokines.

[0310] 33. Efficacy of human and mouse surrogate SOT202 molecules with altered effector functions on kit225 cells SOT202 binds RLI-15 at the C-terminus of the knob heavy chain. AQAIt is a heterodimeric immunocytokine with an antibody derived from humanized IgG1 hCl1a with T366W-knob / T366S, L368A, Y407V-hole substitutions in the heavy chain and with the terminal K deleted (see SEQ ID NO:50, SEQ ID NO:49 and SEQ ID NO:37) fused to IgG1. In the following examples, the term SOT202-XXX denotes a molecule in which further modifying mutations have been made to SOT202, such as the DANA mutation in RLI2 shown in Table 11. For clarity, SOT202-DANA differs from SOT202 only by the additional DA (D61A) mutation, since SOT202 already contains the NA (N65A) mutation (numbers refer to IL-15 numbering). The indicated mutations in the effector domain of the IgG1 molecule that modify the ADCC properties of the antibody, such as AAA, DE and DLE mutations, are listed in Table 2.

[0311] Table 2. The term "afuc" refers to a defucosylated IgG1 molecule. Defucosylated antibodies also have altered ADCC properties.

[0312] The activity of human and mouse surrogate SOT202 ADCC modified molecules for inducing proliferation of kit225 cells was evaluated as described in Example 1, and the EC50 and relative potency compared to SOT101 are shown in Tables 38 and 39. Mouse SOT202 was generated by replacing the human hIgG1 constant domain of SOT202 with its mouse equivalent mIgG2a (mSOT202: SEQ ID NO:51, SEQ ID NO:67 and SEQ ID NO:68; mSOT2020 LALAPG: SEQ ID NO:69, SEQ ID NO:70 and SEQ ID NO:68; mSOT202 isotype: mSOT202 isotype HC knob, SEQ ID NO:72 and SEQ ID NO:73; mSOT202 LALAPG isotype: SEQ ID NO:74, SEQ ID NO:75, SEQ ID NO:73).

[0313] [Table 41]

[0314] This potency assay shows that SOT202 shows the same potency as SOT201 on kit225 cells (see Table 28) and that the ADCC modification did not affect the potency of the immunocytokine. Therefore, this toolbox allows to adjust the ADCC activity of the antibody without affecting the potency of the immunocytokine on the activation of kit225 cells.

[0315] [Table 42]

[0316] For human SOT202, ADCC modification (LALAPG mutation) did not affect the potency of mouse SOT202 surrogates in activating kit225 cells. However, mouse SOT202 surrogates were less potent than their human counterparts. This is likely due to kit225 cells not expressing CD16, which is required for co-signaling with IL-15Rβγ in human and mouse NK cells.

[0317] 34. Human NK cells and CD8 + Potency of human SOT202 ADCC engineered molecules on T cells Human NK cells and CD8 + The activity of human SOT202 ADCC modified molecules for inducing T cell proliferation was assessed as described in Example 1 (hPBMC potency assay) and the EC50 and relative potency compared to SOT202 are shown in FIG. 21 and Table 40.

[0318] [Table 43]

[0319] SOT202-DANA with DLE and DE mutations that enhance ADCC significantly increased human NK cell activity compared to SOT202-DANA without ADCC modification. Defucosylated SOT202 also increased ADCC activity, but to a lesser extent than the DE and DLE mutations. On the other hand, mutations that reduce ADCC, such as the LALAPG mutation, almost abolished NK cell activation. These mutations suppressed the activation of CD8 + It had only a minor effect on T cell activation. Without wishing to be bound by theory, it is hypothesized that the higher binding to the CD16 receptor due to the enhancing mutations synergizes with IL-15Rβγ signaling.

[0320] 35. Human NK cells and CD8 compared with SOT201 molecule + Potency of human SOT202 molecules on T cells Human NK cells and CD8 + The activity of the human SOT202 molecule in inducing T cell proliferation was compared to that of SOT201. The EC50 and relative potency compared to SOT202 and SOT201 are shown in FIG. 22 and Table 41.

[0321] [Table 44]

[0322] Human NK cells and CD8 + The activity of the human SOT202 molecule was compared with that of SOT201-DANA for inducing T cell proliferation. The EC50 and the relative potency compared to SOT202 and SOT201 are shown in Figure 23 and Table 41. The reduced stimulatory activity of the molecule with the DANA mutation compared to the molecule with only the NA mutation confirms the lower stimulatory activity of this mutation, as already described in the previous examples. The SOT202 molecule with enhanced ADCC activity via defucosylation (SOT202 with NA mutation) increases NK cell activity, but does not increase CD8 +It does not increase T cell activity, confirming the results shown in Example 34. SOT201 is based on an IgG4 antibody, which therefore has inherently low ADCC activity. Again, without being bound by theory, it is hypothesized that the higher binding of the defucosylated molecule to the CD16 receptor synergizes with IL-15Rβγ signaling.

[0323] [Table 45]

[0324] The SOT202 and SOT201 molecules bind to human CD8 + It has the same potency on T cells, but not on NK cells. Defucosylation increased human NK cell activity.

[0325] 36. mSOT202 activates immune cells in the spleens of healthy C57BL / 6 mice Murine SOT202 was generated by replacing the human hIgG1 constant domain of SOT202 with its murine equivalent mIgG2a (SEQ ID NO:66, SEQ ID NO:67 and SEQ ID NO:68). Cell proliferation (Ki67) was detected in the spleen by flow cytometry 5 days after IV injection of 5, 10 or 20 mg / kg of mSOT202 compound in healthy C57BL / 6 mice. mSOT202 increased the proliferation of NK cells and CD8 + Dose-dependent stimulation of T cells was observed (Figures 24(A) and (B)).

[0326] 37. mSOT202 induces synergy between RLI2 stimulation of ADCC activity and NK cell proliferation Cell proliferation (Ki67) was detected in the spleen by flow cytometry 5 and 10 days after IV injection of 5 mg / kg of mSOT202 molecule in healthy C57BL / 6 mice. The proliferation activity of mSOT202 (hCl1a-mIgG2a-NA 1x) on NK cells was higher than the effect of hCl1a-mIgG2a (molecule without RLI2) in addition to the effect of mSOT202-LALAPG (hCl1a-mIgG2a-LALAPG-NA 1x without ADCC activity), indicating synergy between the ADCC activity of the antibody in mSOT202 and the proliferation activity of RLI2, which is likely due to CD16 signaling (Figure 25(A)). Therefore, ADCC may contribute to the increase in the activity of NK cells. CD8 + For T cell stimulation, no synergy could be measured in this experimental model (Figure 25(B)).

[0327] References Abramson, HN (2018). “Monoclonal Antibodies for the Treatment of Multiple Myeloma:An Update”. Int J Mol Sci 19(12). Ahmed, AA et al. (2016). “Structural characterization of GASDALIE Fc bound to the activating Fc receptor FcgammaRIIIa”. J Struct Biol 194(1):78-89. Alegre, ML et al. (1992). “Effect of a single amino acid mutation on the activating and immunosuppressive properties of a “humanized” OKT3 monoclonal antibody”. J Immunol 148(11):3461-3468. An,Z.ら、(2009). 「IgG2m4, an engineered antibody isotype with reduced Fc function」. MAbs 1(6):572-579. Bernett,M.J.ら、(2018). 「Abstract 5565:Potency-reduced IL15 / IL15Rα heterodimeric Fc-fusions display enhanced in vivo activity through increased exposure」. Cancer Research 78(13 Supplement):5565-5565. Bolt,S.ら、(1993). 「The generation of a humanized, non-mitogenic CD3 monoclonal antibody which retains in vitro immunosuppressive properties」. Eur J Immunol 23(2):403-411. Chester,C.ら、(2018). 「Immunotherapy targeting 4-1BB:mechanistic rationale, clinical results, and future strategies」. Blood 131(1):49-57. Choi,H.J.ら、(2013). 「A heterodimeric Fc-based bispecific antibody simultaneously targeting VEGFR-2 and Met exhibits potent antitumor activity」. Mol Cancer Ther 12(12):2748-2759. Choi,H.J.ら、(2015). 「Crystal structures of immunoglobulin Fc heterodimers reveal the molecular basis for heterodimer formation」. Mol Immunol 65(2):377-383. Chu,S.Y.ら、(2008). 「Inhibition of B cell receptor-mediated activation of primary human B cells by coengagement of CD19 and FcgammaRIIb with Fc-engineered antibodies」. Mol Immunol 45(15):3926-3933. Clemenceau,B.ら、(2013). 「The human natural killer cytotoxic cell line NK-92, once armed with a murine CD16 receptor, represents a convenient cellular tool for the screening of mouse mAbs according to their ADCC potential」. MAbs 5(4):587-594. Conlon,K.C.ら、(2019). 「Cytokines in the Treatment of Cancer」. J Interferon Cytokine Res 39(1):6-21. Dall’Acqua,W.F.ら、(2002). 「Increasing the affinity of a human IgG1 for the neonatal Fc receptor:biological consequences」. J Immunol 169(9):5171-5180. Darvin,P.ら、(2018). 「Immune checkpoint inhibitors:recent progress and potential biomarkers」. Exp Mol Med 50(12):165. Davis,J.H.ら、(2010). 「SEEDbodies:fusion proteins based on strand-exchange engineered domain (SEED) CH3 heterodimers in an Fc analogue platform for asymmetric binders or immunofusions and bispecific antibodies」. Protein Eng Des Sel 23(4):195-202. De Sousa Linhares,A.ら、(2018). 「Not All Immune Checkpoints Are Created Equal」. Frontiers in Immunology 9(1909). Diebolder,C.A.ら、(2014). 「Complement is activated by IgG hexamers assembled at the cell surface」. Science 343(6176):1260-1263. Dolgin,E. (2020). 「Antibody engineers seek optimal drug targeting TIGIT checkpoint」. Nat Biotechnol 38(9):1007-1009. Edgar,R.C. (2004). 「MUSCLE:multiple sequence alignment with high accuracy and high throughput」. Nucleic Acids Res 32(5):1792-1797. Elliott,J.M.ら、(2014). 「Antiparallel conformation of knob and hole aglycosylated half-antibody homodimers is mediated by a CH2-CH3 hydrophobic interaction」. J Mol Biol 426(9):1947-1957. Forero-Torres, A., et al. (2012). "Results of a phase 1 study of AME-133v (LY2469298), an Fc-engineered humanized monoclonal anti-CD20 antibody, in FcgammaRIIIa-genotyped patients with previously treated follicular lymphoma". Clin Cancer Res 18(5):1395-1403. Fu, Y., et al. (2020). "Therapeutic strategies for the costimulatory molecule OX40 in T-cell-mediated immunity". Acta Pharmaceutica Sinica B 10(3):414-433. Giron-Michel, J., et al. (2005). "Membrane-bound and soluble IL-15 / IL-15Ralpha complexes display differential signaling and functions on human hematopoietic progenitors". Blood 106(7):2302-2310. Godar, M., et al. (2018). "Therapeutic bispecific antibody formats: a patent applications review (1994-2017)". Expert Opinion on Therapeutic Patents 28(3):251-276. Goujon, M., et al. (2010). "A new bioinformatics analysis tools framework at EMBL-EBI". Nucleic Acids Res 38(Web Server Edition):W695-699. Gunasekaran,K.ら、(2010). 「Enhancing antibody Fc heterodimer formation through electrostatic steering effects:applications to bispecific molecules and monovalent IgG」. J Biol Chem 285(25):19637-19646. Hangasky,J.A.ら、(2020). 「Interleukin 15 Pharmacokinetics and Consumption by a Dynamic Cytokine Sink」. Frontiers in Immunology 11(1813). Hori,T.ら、(1987). 「Establishment of an interleukin 2-dependent human T cell line from a patient with T cell chronic lymphocytic leukemia who is not infected with human T cell leukemia / lymphoma virus」. Blood 70(4):1069-1072. Idusogie,E.E.ら、(2001). 「Engineered antibodies with increased activity to recruit complement」. J Immunol 166(4):2571-2575. Lazar,G.A.ら、(2006). 「Engineered antibody Fc variants with enhanced effector function」. Proc Natl Acad Sci USA 103(11):4005-4010. Leabman,M.K.ら、(2013). 「Effects of altered FcgammaR binding on antibody pharmacokinetics in cynomolgus monkeys」. MAbs 5(6):896-903. Leaver-Fay,A.ら、(2016). 「Computationally Designed Bispecific Antibodies using Negative State Repertoires」. Structure 24(4):641-651. Lo,M.ら、(2017). 「Effector-attenuating Substitutions That Maintain Antibody Stability and Reduce Toxicity in Mice」. J Biol Chem 292(9):3900-3908. Lu,X.ら、(2019). 「Deamidation and isomerization liability analysis of 131 clinical-stage antibodies」. MAbs 11(1):45-57. Merchant,A.M.ら、(1998). 「An efficient route to human bispecific IgG」. Nat Biotechnol 16(7):677-681. Mimoto,F.ら、(2013). 「Novel asymmetrically engineered antibody Fc variant with superior FcgammaR binding affinity and specificity compared with afucosylated Fc variant」. MAbs 5(2):229-236. Moore,G.L.ら、(2011). 「A novel bispecific antibody format enables simultaneous bivalent and monovalent co-engagement of distinct target antigens」. MAbs 3(6):546-557. Moore, G. L., et al. (2010). "Engineered Fc variant antibodies with enhanced ability to recruit complement and mediate effector functions". MAbs 2(2):181 - 189. Natsume, A., et al. (2008). "Engineered antibodies of IgG1 / IgG3 mixed isotype with enhanced cytotoxic activities". Cancer Res 68(10):3863 - 3872. Needleman, S. B. and C. D. Wunsch (1970). "A general method applicable to the search for similarities in the amino acid sequence of two proteins". J Mol Biol 48(3):443 - 453. Nellis, D. F., et al. (2012). "Characterization of recombinant human IL - 15 deamidation and its practical elimination through substitution of asparagine 77". Pharm Res 29(3):722 - 738. Nordstrom, J. L., et al. (2011). "Anti - tumor activity and toxicokinetics analysis of MGAH22, an anti - HER2 monoclonal antibody with enhanced Fcgamma receptor binding properties". Breast Cancer Res 13(6):R123. Pearson, W. R. and D. J. Lipman (1988). "Improved tools for biological sequence comparison". Proc Natl Acad Sci USA 85(8):2444 - 2448. Pini, A. et al., (1998). "Design and use of a phage display library. Human antibodies with subnanomolar affinity against a marker of angiogenesis eluted from a two - dimensional gel". J Biol Chem 273(34):21769 - 21776. Richards, J. O. et al., (2008). "Optimization of antibody binding to FcgammaRIIa enhances macrophage phagocytosis of tumor cells". Mol Cancer Ther 7(8):2517 - 2527. Ring, A. M. et al., (2012). "Mechanistic and structural insight into the functional dichotomy between IL - 2 and IL - 15". Nat Immunol 13(12):1187 - 1195. Robinson, N. E. and A. B. Robinson (2004). "Prediction of primary structure deamidation rates of asparaginyl and glutaminyl peptides through steric and catalytic effects". J Pept Res 63(5):437 - 448. Robinson, T.O. and K.S. Schluns (2017). "The potential and promise of IL-15 in immuno-oncogenic therapies". Immunol Lett 190:159-168. Rother, R.P. et al., (2007). "Discovery and development of the complement inhibitor eculizumab for the treatment of paroxysmal nocturnal hemoglobinuria". Nat Biotechnol 25(11):1256-1264. Shang, L. et al., (2014). "Selective antibody intervention of Toll-like receptor 4 activation through Fc gamma receptor tethering". J Biol Chem 289(22):15309-15318. Shields, R.L. et al., (2001). "High resolution mapping of the binding site on human IgG1 for Fc gamma RI, Fc gamma RII, Fc gamma RIII, and FcRn and design of IgG1 variants with improved binding to the Fc gamma R". J Biol Chem 276(9):6591-6604. Smith, T.F. and M.S. Waterman (1981). "Comparison of biosequences". Advances in Applied Mathematics 2(4):482-489. Sola, R.J. and K. Griebenow (2009). "Effects of glycosylation on the stability of protein pharmaceuticals". J Pharm Sci 98(4):1223-1245. Solomon, B. L., & Garrido-Laguna, I. (2018). "TIGIT: a novel immunotherapy target moving from bench to bedside." Cancer Immunol Immunother, 67(11), 1659-1667. Soman, G., et al. (2009). "MTS dye based colorimetric CTLL-2 cell proliferation assay for product release and stability monitoring of interleukin-15: assay qualification, standardization and statistical analysis." J Immunol Methods, 348(1-2), 83-94. Spiess, C., et al. (2015). "Alternative molecular formats and therapeutic applications for bispecific antibodies." Mol Immunol, 67(2 Pt A), 95-106. Stavenhagen, J. B., et al. (2007). "Fc optimization of therapeutic antibodies enhances their ability to kill tumor cells in vitro and controls tumor expansion in vivo via low-affinity activating Fcgamma receptors." Cancer Res, 67(18), 8882-8890. Tao, M. H., & Morrison, S. L. (1989). Studies of aglycosylated chimeric mouse-human IgG. Role of carbohydrate in the structure and effector functions mediated by the human IgG constant region. J Immunol, 143(8), 2595-2601. Thaysen-Andersen, M., et al. (2016). Recombinant human heterodimeric IL-15 complex displays extensive and reproducible N- and O-linked glycosylation. Glycoconj J, 33(3), 417-433. Toutain, P. L., & Bousquet-Melou, A. (2004). Plasma terminal half-life. J Vet Pharmacol Ther, 27(6), 427-439. Vafa, O., et al. (2014). An engineered Fc variant of an IgG eliminates all immune effector functions via structural perturbations. Methods, 65(1), 114-126. Villa, A., et al. (2008). A high-affinity human monoclonal antibody specific to the alternatively spliced EDA domain of fibronectin efficiently targets tumor neo-vasculature in vivo. Int J Cancer, 122(11), 2405-2413. Von Kreudenstein,T.S.ら、(2013). 「Improving biophysical properties of a bispecific antibody scaffold to aid developability:quality by molecular design」. MAbs 5(5):646-654. Vonderheide,R.H. (2020). 「CD40 Agonist Antibodies in Cancer Immunotherapy」. Annual Review of Medicine 71(1):47-58. Waldmann,T.A. (2015). 「The shared and contrasting roles of IL2 and IL15 in the life and death of normal and neoplastic lymphocytes:implications for cancer therapy」. Cancer Immunol Res 3(3):219-227. Waldmann,T.A.ら、(2020). 「IL-15 in the Combination Immunotherapy of Cancer」. Frontiers in Immunology 11(868). Waldmeier,L.ら、(2016). 「Transpo-mAb display:Transposition-mediated B cell display and functional screening of full-length IgG antibody libraries」. MAbs 8(4):726-740. Walker,M.R.ら、(1989). 「Aglycosylation of human IgG1 and IgG3 monoclonal antibodies can eliminate recognition by human cells expressing Fc gamma RI and / or Fc gamma RII receptors」. Biochem J 259(2):347-353. Wang,X.ら、(2018). 「IgG Fc engineering to modulate antibody effector functions」. Protein Cell 9(1):63-73. Wei,X.ら、(2001). 「The Sushi domain of soluble IL-15 receptor alpha is essential for binding IL-15 and inhibiting inflammatory and allogenic responses in vitro and in vivo」. J Immunol 167(1):277-282. Xu,D.ら、(2000). 「In vitro characterization of five humanized OKT3 effector function variant antibodies」. Cell Immunol 200(1):16-26. Yu,X.ら、(2017). 「Beyond Antibodies as Binding Partners:The Role of Antibody Mimetics in Bioanalysis」. Annu Rev Anal Chem(Palo Alto Calif) 10(1):293-320. Zalevsky,J.ら、(2010). 「Enhanced antibody half-life improves in vivo activity」. Nat Biotechnol 28(2):157-159. Zhang,T.ら、(2018). 「The binding of an anti-PD-1 antibody to FcγRΙ has a profound impact on its biological functions」. Cancer Immunology, Immunotherapy 67(7):1079-1090.

Claims

1. An IL-15 variant comprising amino acid substitutions at positions G78 and N79 of mature human interleukin-15 (IL-15).

2. The IL-15 variant according to claim 1, wherein the IL-15 variant comprises the amino acid substitutions G78A, G78V, G78L or G78I, and N79Q, N79H or N79M.

3. The IL-15 variant according to claim 1, wherein the IL-15 variant is expressed in a mammalian cell line.

4. The amino acid substitutions are (a) reducing deamidation at N77 and glycosylation at N79 of the IL-15 variant as compared to mature human IL-15, (b) resulting in less than 30% glycosylated IL-15 variant, and / or (c) increasing glycosylation at N71 of the IL-15 variant as compared to mature human IL-15 The IL-15 variant according to claim 1.

5. The IL-15 variant according to claim 1, wherein the amino acid substitutions do not substantially reduce the IL-15 activity of the IL-15 variant in inducing the proliferation of kit225 cells, 32Db cells, human PBMCs, or in the Promega IL-15 - bioassay.

6. The IL-15 variant according to claim 1, wherein the IL-15 variant does not have substitutions at position N71 and / or position N77.

7. The IL-15 variant comprises at least one additional substitution that reduces binding to IL-2 / IL-15Rβ and / or γ ｃ receptor and / or IL-15Rα, (a) said IL-2 / IL-15Rβ and / or said γ ｃ The site of said further substitution that reduces binding to the receptor is selected from the list consisting of N1, N4, S7, D8, K10, K11, D30, D61, E64, N65, L69, N72, E92, Q101, Q108, and I111, or (b) said further substitution that reduces the binding to said IL-2 / IL-15Rβ and / or said γ ｃ receptor is selected from the list consisting of N1D, N1A, N1G, N4D, S7Y, S7A, D8A, D8N, K10A, K11A, D30N, D61A, D61N, E64Q, N65D, N65A, N65E, N65R, N65K, L69R, N72R, Q101D, Q101E, Q108D, Q108A, Q108E and Q108R, or (c) said further substitution that reduces binding to said IL-2 / IL-15Rβ and / or said γ ｃ receptor is a combined substitution and is selected from the list consisting of D8N / N65A, D61A / N65A, and D61A / N65A / Q101D, and / or (a) the site of the further substitution that reduces the binding to IL-15Rα is selected from the list consisting of L44, L45, E46, L47, V49, I50, S51, E64, L66, I67, I68 and L69, (b) the further substitution that reduces the binding to IL-15Rα is selected from the list consisting of L44D, E46K, E46G, L47D, V49D, V49R, I50D, L66D, L66E, I67D, and I67E, or (c) the further substitution that reduces the binding to IL-15Rα is a combined substitution selected from the list consisting of E46G / V49R, N1A / D30N / E46G / V49R, N1G / D30N / E46G / V49R / E64Q, V49R / E46G / N1A / D30N and V49R / E46G / N1G / E64Q / D30N The IL-15 variant according to claim 1.

8. A conjugate comprising the IL-15 variant according to claim 1.

9. A fusion protein comprising the IL-15 variant according to claim 1.

10. The fusion protein according to claim 9, comprising a human IL-15Rα sushi domain, a linker, and the IL-15 variant according to claim 1.

11. The fusion protein according to claim 9, wherein the fusion protein comprises a targeting moiety, and the targeting moiety is an antibody or a functional variant thereof that binds to a tumor antigen, a tumor extracellular matrix antigen, or a tumor angiogenesis antigen, or an immunomodulatory antibody.

12. A nucleic acid encoding the IL-15 variant according to claim 1, the conjugate according to claim 8, or the fusion protein according to claim 9.

13. A vector comprising the nucleic acid according to claim 12.

14. A host cell comprising the nucleic acid according to claim 12.

15. The IL-15 variant according to claim 1 for use in therapy.

16. The conjugate according to claim 8 for use in therapy.

17. The fusion protein according to claim 9 for use in therapy.

18. The nucleic acid according to claim 12 for use in therapy.

19. A pharmaceutical composition comprising the IL-15 variant according to claim 1 and a pharmaceutically acceptable carrier.

20. A pharmaceutical composition comprising the conjugate according to claim 8 and a pharmaceutically acceptable carrier.

21. A pharmaceutical composition comprising the fusion protein according to claim 9 and a pharmaceutically acceptable carrier.

22. A pharmaceutical composition comprising the nucleic acid according to claim 12 and a pharmaceutically acceptable carrier.

23. The IL-15 variant according to claim 1 for use in the treatment of a subject suffering from a neoplastic disease or an infectious disease, at risk of developing a neoplastic disease or an infectious disease, and / or diagnosed with a neoplastic disease or an infectious disease.

24. The conjugate according to claim 8 for use in the treatment of a subject suffering from a neoplastic disease or an infectious disease, at risk of developing a neoplastic disease or an infectious disease, and / or diagnosed with a neoplastic disease or an infectious disease.

25. The fusion protein according to claim 9 for use in the treatment of a subject suffering from a neoplastic disease or an infectious disease, at risk of developing a neoplastic disease or an infectious disease, and / or diagnosed with a neoplastic disease or an infectious disease. The nucleic acid according to claim 12 for use in the treatment of a subject suffering from a neoplastic disease or an infectious disease, at risk of developing a neoplastic disease or an infectious disease, and / or diagnosed with a neoplastic disease or an infectious disease.

Images