Pegylated il- 15 receptor alpha cytokines

Abstract

Provided herein are IL-15 cytokine therapeutics showing improved proliferation and maintenance of CD8+ T cells preferentially over NK cells and which may show reduced injection site sensitivity.

Description

PEGYLATED IL- 15 RECEPTOR ALPHA CYTOKINESCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 632,231 filed April 10, 2024, the entire contents of which are incorporated herein by reference.REFERENCE TO AN ELECTRONIC SEQUENCE LISTING

[0002] The contents of the electronic sequence listing (670572002940SEQLIST.xml; Size: 46,023 bytes; and Date of Creation: April 7, 2025) is herein incorporated by reference in its entirety.FIELD

[0003] The present disclosure relates generally to IL-15 cytokine therapeutics, and more specifically to IL- 15 cytokine therapeutics showing improved proliferation and maintenance of CD8+T cells preferentially over NK cells and which may show reduced injection site sensitivity.BACKGROUND

[0004] IL- 15 is a pleiotropic cytokine important for both adaptive and innate immunity. IL-15 promotes the activation and maintenance of natural killer (NK), NK T cells, CD8+T cells, and CD8aa intraepithelial lymphocytes, and is of interest as an immunotherapeutic agent for the treatment of cancers and immunodeficiencies. NK cells are innate immune cells that autonomously eliminate tumor and virus-infected cells. They recognize infected or damaged cells and release proteins (e.g. perforins / granzymes) that cause cell lysis. CD8+T cells constitute a broad class of cytotoxic T cells; they are MHC class I-restricted T cells and are mediators of adaptive immunity. They directly kill cancer cells presenting non- selfepitopes. CD44hlCD8+T cells are antigen- specific T cells that can remain long-term after an infection / tumor has been eliminated. Memory T cells quickly convert into large numbers of effector T cells upon re-exposure to the specific antigen, thus providing a rapid response to past illness.

[0005] Exogenous IL- 15 has been shown to stimulate proliferation of CD8+T cells both in vivo and in vitro. Low-dose therapy with IL-15 is hypothesized to promote themaintenance and function of tumor- specific CD8+T cells and thus delay or prevent tumor relapse in failed adoptive immunotherapy (Roychowdhury et al., Cancer Research 64: 8062-7 (2004)). Low-dose therapy by continuous infusion to monkeys over 10 days resulted in a 100-x expansion of CD8+effector T cells in the peripheral blood, which was more effective than a daily bolus dosing regimen (Sneller et al., Blood 118: 6845-8 (2011)). Stabilized muteins of IL-15 have been reported (Nellis et al., Pharm. Res. 29: 722-38 (2012)). Certain permanent and releasable conjugates of IL- 15 with water-soluble polymers have been disclosed (PCT Publication WO2015 / 153753A2; WO2020 / 219943A2). Muteins of IL- 15 showing improved receptor agonism have been disclosed (Zhu et al., J. Immunology 2009, 183(6): 3598). IL-15[N72D] showed a 4-5 fold increase in biological activity over native IL- 15 in cell proliferation assays.

[0006] IL- 15 is unique in being a secreted cytokine that acts in trans; endogenous IL- 15 exists primarily in a tightly bound complex (Ka - 100 pM) with the alpha- subunit of the ternary IL- 15 receptor (IL-15Ra). Signal transduction occurs when this IL-15Ra / IL-15 complex on the surface of an antigen-presenting cell binds to the remaining Rb and Rgcchains present on the surface of another immune cell (Stonier & Schluns 2010, Immunol. Lett. 127(2): 85-92). Unlike IL-15, expression of IL-15Ra is ubiquitous, and numerous isoforms have been described. IL-15Ra is a member of a subfamily of cytokine receptor subunits comprising N-terminal extracellular cytokine-binding domains called “sushi domains.” Isoforms of IL-15Ra lacking the sushi domain are unable to stimulate the IL- 15 response.

[0007] IL- 15 receptor agonists comprising IL- 15 together with the sushi domain of the IL-15Ra (IL-15RaSu) have also been reported, both as complexes and as fusion proteins (Han et al., Cytokine 2011, 56(3):804-10; Mortier et al., J. Biological Chem. 2006, 281: 1612-9; Xu et al., Microb Cell Fact 2021, 20:115; Shen et al. J Exp Med (2022) 219 (12): e20220745; PCT Publication W02007 / 046006; PCT Publication WO2012 / 175222A1; US Patent 10,358,477). A multimeric complex of IL-15[N72D] and IL-15RaSuFc wherein IL- 15Ra is fused to the Fc domain of IgGl (ALT-803 / N-803) is currently in clinical trials (Romee et al., 2018, Blood 131(23): 2515-27). A receptor-linked fusion protein, RLI (SO- C101; SOT101) is also in clinical trials as monotherapy and in combination with pembrolizumab (Keytruda®) (Champiat et al. Cancer Res. 2022 82(12S):CT040). Injection site reactions have been reported for both agents.

[0008] Conjugation of proteins is a common technique for half-life extension, however such conjugation has unpredictable effects on the biological activity of the resulting molecule. Conjugation to polyethylene glycol (PEGylation) of IL-15 has been described. Random PEGylation of IL- 15 using 5-kDa PEG-succinimidyl carbonate produced a protein having an extended half-life but with greatly altered biological activity, losing the ability to stimulate the proliferation of CTLL-2 cells but being able to inhibit the activity of native IL- 15 and thus being considered an IL-15 antagonist (Pettit et al., J. Biol. Chem. 1997 272(4): 2312-8). In contrast, other PEGylated IL- 15 conjugates have been reported to maintain IL- 15 agonist activity (PCT Publication WO2018 / 032817). N-terminal PEGylation of IL-15 has been reported to reduce the biological activity (Chinese Patent CN102145178B). Thus, the effects of PEGylation on IL- 15 biological activity appears unpredictable, and the specific location of PEG attachment to the protein is likely to be a determining factor.

[0009] Other means of extending the half-life of IL- 15 have been described. Slow-release conjugates of IL- 15 cytokines have been disclosed (PCT Publication 2020 / 219943). These complexes, fusion proteins, and conjugates generally show enhanced ability to stimulate T cell expansion relative to IL-15 alone. Continuous i.v. infusion of IL-15 at 20 ug / kg / day for 10 days in monkeys resulted in lOx increases in circulation NK cells, 15x increases in monocytes, and 80-100x increases in circulating CD8+effector memory T cells (TEM cells), compared with a more modest lOx increase in TEM cells when given as daily s.c. injections suggesting that extended exposure may account for certain pharmacodynamic differences (Sneller et al., Blood 2011, 118:6845-8). A short-course treatment using continuous infusion of 4 ug / kg / day of IL- 15 in humans resulted in an average 34x increase in total NK cells, a 144x increase in CD56bnghtNK cells, but only a 3.38x increase in CD8+T cells (Dubois et al. 2021 J Immunother Cancer 9(4): e002193). Continuous infusion of IL-15 in humans using 2 ug / kg / day for 10 days similarly shows a large increase in NK cells (38x total NK cells, 358x CD56bnghtNK cells) but a modest 5.8x increase in CD8+T cells (Conlon et al. 2019 Clin.Cancer Res. 25(16): 4945-54). It has been suggested that the antitumor activity of these agents results from the expansion and activation of tumor-resident CD8+T cells, while systemic expansion of NK cells may be the cause of immunotoxicity induced by administration of such agents, with the NK cells producing the pro-inflammatory cytokine ILN-g (Guo et al., 2022). Eactors governing the selectivity in stimulation of CD8+T cells and NK cells between different analogs remain unclear.

[0010] Subcutaneous administration of these agents may also be accompanied by injection site reactions that may limit their utility. It has been suggested that such injection site reactions in the case of the IL-15 superagonist complex ALT-803 are mediated by gamma / delta T cell infiltration at the injection site (Romee et al., Blood 2018 131(23): 2515- 27).

[0011] There thus remains an unmet need for IL- 15 cytokine therapeutics showing improved proliferation and maintenance of CD8+T cells preferentially over NK cells and which may show reduced injection site sensitivity.BRIEF SUMMARY

[0012] In some aspects, provided are conjugates of a water-soluble polymer and the sushi domain of the IL-15Ra (IL15RaSu). These conjugates may further comprise an associated IL-15, either as a noncovalent complex with or covalently linked to the Polymer-IL-15RaSu. While various methods of extending the half-life of IL- 15, receptor IL- 15 complexes, and receptor-IL-15 fusion proteins have been reported, the conjugates of the present invention provide unexpectedly potent expansion of CD8+T-cells relative to NK cells.

[0013] Pegylated conjugates of the disclosure promote sustained memory CD8+ T-cell responses and modulating innate cytotoxicity. Accordingly, pegylated conjugated of the disclosure offer advantages over existing IL- 15 agonists in enhancing tumor- specific responses while reducing toxicity risks across diverse cancer contexts.

[0014] In one aspect the invention provides conjugates of formula (I)[P-L]x-S (I) wherein P is a water-soluble polymer; L is absent or is a linker; x = 1-5; and S is a polypeptide comprising an IL-15 receptor alpha sushi domain (IL15RaSu). In some embodiments, x = 1. In certain embodiments, x = 1 and P-L is attached to the N-terminal amine of S. In further embodiments, the conjugates further comprise an IL- 15 polypeptide. In certain of these further embodiments, the IL- 15 polypeptide is non-covalently complexed with polypeptide S. In certain other of these further embodiments, the IL- 15 polypeptide is covalently linked to the sushi domain through a peptidic linker such that S is a fusion protein. The conjugates of the invention are unexpectedly potent stimulators of immune T-cell andNK cell proliferation showing marked preference for expansion of T cells over NK cells and showing reduced injection site toxicity over the non-conjugated proteins.

[0015] In other aspects, provided are methods for the preparation and use of these conjugates.DESCRIPTION OF THE FIGURES

[0016] The present application can be understood by reference to the following description taken in conjunction with the accompanying figures.

[0017] Figure 1 illustrates one embodiment of the invention wherein a conjugate [P-L]x-S of formula (I) is prepared using a two-step process wherein in a first step a linker of formula (II) is used to reductively alkylate one or more amine groups of polypeptide S to give a linker- polypeptide Lx-S comprising a reactive functional group Z’ , and in a second step the Lx-S is reacted with an activated water-soluble polymer P-Z such that Z and Z’ react to give a connecting group Z*, thus forming conjugate [P-L]x-S. Further illustrated is isolation of singly- conjugated S, [P-E]i-S.

[0018] Figure 2 illustrates one embodiment of the invention wherein a conjugate [P-E]x-S of formula (I) is prepared by using a water-soluble polymer- aldehyde to reductively alkylate one or more amine groups of polypeptide S.

[0019] Figure 3 illustrates SDS-PAGE gel-shift analysis of a reaction mixture of Example 1 comprising 51% [E]i-S, 38% [E]2-S, and 11% [E]a-S after reaction with 5kDa PEG-DBCO (lane 1); the same reaction mixture without added 5kDa PEG-DBCO; free S (lane 3); and molecular weight markers (lane 4).

[0020] Figure 4 illustrates an example of ion exchange chromatography of [P-E]x-S from Example 2 using Sepharose QFF and SDS-PAGE analysis of fractions. Panel (a) shows the elution profile of free S. Panel (b) shows the elution profile of the reaction mixture producing [P-L]x-S over the range where x = 0-2.

[0021] Figure 5 shows a comparison of the in vitro activity of [P-L]i-S from Example 3, free S, and IL- 15. Panel A shows the results of a U2OS cell-based assay measuring binding to IL-2RPy. [P-L]i-S (filled squares, ECso=1.2 nM) was 7x less potent than free S (open circles, ECso=O.17 nM). Panel B shows the results of a CTLL-2 cell proliferation assay. [P-L] i-S (filledsquares, ECso=L5 nM) was ~3.5 fold less potent than free S (open circles, EC50 = 0.42 nM) and -500 fold less potent than IL- 15 (open squares, EC50 = 2.9 pM). In both panels, data were fit to a three- parameter logistic model. Points represent the mean ± SD.

[0022] Figure 6 shows the mouse pharmacokinetics of the [P-L]i-S of Example 2. C57BL / 6J mice (n=6) were given a single IV (slow bolus) or SC injection of [P-L]i-S (10 pg of protein, 0.43 nmol). ELISA was used to quantify the plasma concentration of [P-L]i-S. Data was fit to a one phase decay equation (IV, ti / 2 = 15 h) or biexponential equation (SC, ti / 2 = 15 h) using Graphpad Prism 9 software. Points are mean ± SD (n=6 / group).

[0023] Figure 7 shows the tolerability of the [P-L]i-S of Example 2 in mice. C57BL / 6J mice (n = 5 / group) were given a single s.c. injection of the conjugate and the body weights were followed over a 3-week period. Doses of 4 pg (0.17 nmol; diamonds), 10 pg (0.43 nmol; open circles), and 20 pg (0.86 nmol; triangles) ug were well-tolerated. The highest dose of 40 pg (1.72 nmol; squares) showed -15% body weight loss after 5 days. All dose weights are expressed in terms of protein content.

[0024] Figure 8 shows the results of total lymphocyte counts upon treatment with the [P- L]i-S of Example 2 compared with other IL-15 agonists. Mice were treated with either 2x2 pg (2x 0.09 nmol) of free S separated by 48 h (A); 10 pg (0.43 nmol) of free S; 10 pg (0.43 nmol) of a slow-release preparation of S (C); or 20 pg (0.86 nmol; D), 10 pg (0.43 nmol; E), or 4 pg (0.17 nmol; F) of [P-L]i-S. Total lymphocytes were counted by FACS analysis of PBMCs 5 days after dosing. All dose weights are expressed in terms of protein content.

[0025] Figure 9 shows the results of CD8+, CD44hlCD8+, and NK lymphocyte counts 5 days after treatment with the [P-L]i-S of Example 2 compared with other IL- 15 agonists. Mice were treated with either 2x2 pg (2x 0.09 nmol) of free S separated by 48 h (A); 10 pg (0.43 nmol) of free S; 10 pg (0.43 nmol) of a slow-release preparation of S (C); or 40 pg (1.72 nmol; D), 20 pg (0.86 nmol; E), 10 pg (0.43 nmol; F) or 4 pg (0.17 nmol; G) of [P-L]i-S. Lymphocytes were counted by FACS analysis of PBMCs 5 days after dosing. Panel A shows the percentage of the cell type in the PBMCs; Panel B shows the absolute cell numbers; and Panel C shows the fold-change in absolute cell numbers. All dose weights are expressed in terms of protein content.

[0026] Figure 10 shows the expansion of B cells and CD4+T cells 5 days after treatment with the [P-L]i-S of Example 2 compared with other IL- 15 agonists. Mice were treated witheither 2x2 pg of free S (2x 0.09 nmol) separated by 48 h (A); 10 pg (0.43 nmol) of free S; 10 pg (0.43 nmol) of a slow-release preparation of S (C); or 40 pg (1.72 nmol; D), 20 pg (0.86 nmol; E), 10 pg (0.43 nmol; F) or 4 pg (0.17 nmol; G) of [P-L]i-S. All dose weights are expressed in terms of protein content.

[0027] Figure 11 shows the time courses of NK, CD8+, and CD44hlCD8+lymphocyte expansion after treatment with the [P-E]i-S of Example 2 compared with other IE- 15 agonists. Mice were treated either with 10 pg (0.43 nmol) of the [P-L]i-S of Example 2 (solid squares); 10 pg (0.43 nmol) of a slow-release preparation of S (solid triangles); or 10 pg (0.43 nmol; inverted open triangles) or 2 doses of 2 pg (2x 0.09 nmol; days 0 and 2) (open circles) of free S. PBMCs were sampled and analyzed by FACS at 2, 5, 7, 14, 21, and 28 days post-dosing. Panel A shows absolute cell counts; Panel B shows fold change in cell counts; Panel C shows % of the cell type expressing the Ki67 proliferation marker. All dose weights are expressed in terms of protein content.

[0028] Figure 12 shows the ratios of relevant target immune cells in the periphery following treatment with [P-L]i-S of Example 2, free S or a slow-release preparation of S as described in Example 6. Panel (A) CD8+ / CD4+cell ratio, Panel (B) CD8+ / NK cell ratio and Panel (C) CD44hlCD8+ / NK cell ratio. Panels (D, E) show the CD8+ / NK cell ratio and CD44hlCD8+ / NK cell ratio, respectively of IL- 15 agonists reported in the literature. C57BL / 6J mice were treated with either 22 nmol / kg [P-L]i-S (squares), 22 nmol / kg slow-release S (triangles, dash line), 22 nmol / kg free S (inverted triangles) or two 4.4 nmol / kg doses of free S separated by 48 h (circles). All dose weights are expressed in terms of protein content.

[0029] Figure 13 shows the immune cell expansion of NK, CD8+, CD44hlCD8+, and Tregs in the PBMCs following treatment with the [P-L]i-S of Example 2 or free S as described in Example 7. Panel A shows absolute cell counts; Panel B shows fold change in cell counts; Panel C shows % of the cell type expressing the Ki67 proliferation marker. C57BL / 6J (n=3- 4 / point / group) mice were treated with either 1 pg [P-L]i-S (triangles), 10 pg [P-L]i-S (squares) or 10 pg free S (inverted triangles). All dose weights are expressed in terms of protein content. Error bars represent plus or minus one standard deviation of the mean.

[0030] Figure 14 shows the immune cell expansion of NK, CD8+, CD44hlCD8+, and Tregs in the spleen following treatment with the [P-L]i-S of Example 2 or free S as described in Example 7. Panel A shows absolute cell counts; Panel B shows fold change in cell counts;Panel C shows % of the cell type expressing the Ki67 proliferation marker. C57BL / 6J mice (n=3-4 / point / group) were treated with either 1 pg [P-L]i-S (triangles), 10 pg [P-L]i-S (squares) or 10 pg free S (inverted triangles). All dose weights are expressed in terms of protein content. Error bars represent plus or minus one standard deviation of the mean.

[0031] Figure 15 shows the ratios of relevant target immune cells in the spleen (Panels A- E) and the periphery (Panels F-J) following treatment with the [P-L]i-S of Example 2 or free S as described in Example 7. Panel (A,F) CD8+ / NK cell ratio, Panel (B,G) CD44hlCD8+ / NK cell ratio, Panel (C,H) NK cell / Treg ratio, Panel (D,I) CD8+ / Treg ratio, and Panel (E,J) CD44hlCD8+ / Tregratio. C57BL / 6J mice were treated with either 1 pg [P-L]i-S (triangles), 10 pg [P-L]i-S (squares) or 10 pg free S (inverted triangles). All dose weights are expressed in terms of protein content.

[0032] Figure 16 shows the combination of [P-L]i-S with anti-PD-1 provides enhanced therapeutic efficacy in the CT26 model. Panel A shows the mean tumor volume and panel B shows the overall survival of Balb / c mice bearing a single CT26 tumor (n=6-8 / group). Mice were treated with either 6 doses of 0.2 mg anti-PD-1, a single dose of 0.5 mg / kg [P-L]i-S or a combination of both anti-PD-1 and [P-L]i-S.DETAILED DESCRIPTION

[0033] The following description sets forth exemplary compositions, methods, parameters and the like. It should be recognized, however, that such description is not intended as a limitation on the scope of the present disclosure but is instead provided as a description of exemplary embodiments.

[0034] In one aspect, provided are conjugates of formula (I)[P-L]x-S (I) wherein P is a water-soluble polymer; L is absent or is a linker; x = 1-5; and S is a polypeptide comprising an IL-15 receptor alpha sushi domain (IL15RaSu). In some embodiments, x = 1. In some embodiments, x = 1 and P-L is attached to the N-terminal amine of S. In further embodiments, the conjugates further comprise an IL- 15 polypeptide. In certain of these further embodiments, the IL- 15 polypeptide is non-covalently complexed with polypeptide S. In certain other of these further embodiments, the IL- 15 polypeptide iscovalently linked to the sushi domain through a peptidic linker such that S is a fusion protein. The conjugates of the invention are unexpectedly potent stimulators of immune T-cell and NK cell proliferation showing marked preference for expansion of T cells over NK cells and showing reduced injection site toxicity over the non-conjugated proteins.

[0035] The sequence of the human IL- 15 receptor alpha subunit is given in SEQ ID No: 1. As used herein, the term “IL- 15 receptor alpha sushi domain” or “sushi domain” (IL15RaSu) refers to a polypeptide sequence beginning at the first cysteine residue and ending after the fourth cysteine residue after the signal peptide of the IL- 15 receptor alpha subunit. The sushi domain from the human IL- 15 receptor alpha subunit is given in SEQ ID No: 2.SEQ ID No: 1 Human IL- 15 Receptor alpha subunitMAPRRARGCR TLGLPALLLL LLLRPPATRG ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSTV TTAGVTPQPE SLSPSGKEPA ASSPSSNNTA ATTAAIVPGS QLMPSKSPST GTTEISSHES SHGTPSQTTA KNWELTASAS HQPPGVYPQG HSDTTVAIST STVLLCGLSA VSLLACYLKS RQTPPLASVE MEAMEALPVT WGTSSRDEDL ENCSHHLSEQ ID No: 2 Human IL- 15 Receptor alpha subunit sushi domainCPPPMSVEHA DIWVKSYSLY SRERYICNSG EKRKAGTSSL TECVLNKATN VAHWTTPSLK C

[0036] Variants of these sequences are known and useful for the present invention. The sequence of the sushi domain from any mammalian IL- 15 receptor alpha subunit can be readily identified by one of skill in the art (Wei et al., 2001 J. Immunol. 167(1): 277-282). In embodiments of the present invention, the term “IL- 15 receptor alpha sushi domain” refers to a sushi domain from a mammal (consensus sequence SEQ ID No: 3), preferably from a primate (consensus sequence SEQ ID No:4), more preferably from human (SEQ ID No: 2).SEQ ID No: 3 Mammalian consensus sequenceCPxPxSxEHA DIxVKxYSxx SRERYxCNSG FKRKAGTSxL xECVxNKxTN xAxWTTPSLK CSEQ ID No: 4 Primate consensus sequenceCPxPxSVEHA DIxVKSYSLx SRERYxCNSG FKRKAGTSSL TECVLNKATN xAxWTTPSLK C

[0037] As used herein, the term “IL- 15” refers to a cytokine that binds to the IL- 15 receptor complex and stimulates the proliferation of T and natural killer (NK) cells and maintains survival of memory T cells in the absence of antigen. In some embodiments, IL- 15 is a polypeptide having at least 90% sequence identity to human interleukin- 15 (SEQ ID No: 5). In certain embodiments, various changes in sequence are made to promote activity and / or increase stability of the protein. Some sequence changes may include N77A (SEQ ID No: 7) to promote stability against deamidation, N72D (SEQ ID No: 8) to enhance agonist activity, and a combination of the two (SEQ ID No: 9). Further sequences changes include L45D, L45E, S51D, L52D, N72A, N72E, N72S, N72Y, and N72P (Zhu et al., J. Immunol. 2009 183(6): 3598- ; Nellis et al., Pharm Res 201229:722-38).SEQ ID No: 5 human interleukin- 15NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVI SLESGDASIH DTVENLIILA NNSLSSNGNV TESGCKECEE LEEKNIKEFL QSFVHIVQMF INTSSEQ ID No: 6 mammalian interleukin- 15 consensus sequence xWxxVxxDLx xIxxLxxxxH xDxTLYTxSx xHPxCKxTxM xCFLLELxVI xxExxxxxxx xxxxNxxxLA NxxLxxxxxx xExGCKxCEE LExxxxxEFL xSFxxIVQMF IxxxSEQ ID No: 7 human IL-15 [N77A]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVI SLESGDASIH DTVENLIILA NNSLSSAGNV TESGCKECEE LEEKNIKEFL QSFVHIVQMF INTSSEQ ID No: 8 human IL-15 [N72D]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NDSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 9 human IL-15 [N72D,N77A]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NDSLSSAGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 10 human IL-15 [L45D]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLDELQVISLESGDASIH DTVENLIILA NNSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 11 human IL-15 [L45E]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLEELQVISLESGDASIH DTVENLIILA NNSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 12 human IL-15 [S51D]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVIDLESGDASIH DTVENLIILA NNSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 13 human IL-15 [L52D]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISDESGDASIH DTVENLIILA NNSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 14 human IL-15 [N72E]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NESLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 15 human IL-15 [N72A]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NASLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 16 human IL-15 [N72S]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NSSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 17 human IL-15 [N72Y]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NYSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 18 human IL-15 [N72P]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA NPSLSSNGNV TESGCKECEE LEEKNIKEFLQSFVHIVQMF INTSSEQ ID No: 19 human IL-15 [N71S,N72A,N77A]NWVNVISDLK KIEDLIQSMH IDATLYTESD VHPSCKVTAM KCFLLELQVISLESGDASIH DTVENLIILA SASLSSAGNV TESGCKECEE LEEKNIKEFL QSFVHIVQMF INTS

[0038] As used herein, the term “peptidic linker” refers to a polypeptide sequence used to connect a sushi domain with an IL- 15 so as to create a fusion protein. These peptidic linkersare advantageously hydrophilic polypeptides having no defined secondary structure such that they can flexibly connect the sushi domain and the IL- 15. Examples of linker sequences are given in US Patent 5,073,627; US Patent 5,108,910; and US Patent 10,358,477. These are typically polypeptides of 30 amino acids or less, comprising primarily small hydrophilic residues such as glycine and serine but also potentially including threonine and alanine for flexibility and / or lysine and glutamate for solubility (Chen et al. 2013 Adv Drug Deliv Rev 65(10): 1357-69). In certain embodiments, the peptidic linker is a “GS linker” comprising repeats of (GGS), (GGGS), and (GGGGS)nunits. Particular peptidic linker sequences are given in SEQ ID No: 20-24.SEQ ID No: 20 peptidic linkerSSGGSGGGGS GGGSGGGGSLQSEQ ID No: 21 UCSF linker 1SGSSGSSGSSGSSGSSGSSINSEQ ID No: 22 UCSF linker 2SGSSSGSSSGSSSGSSSGSSNSEQ ID No: 23 UCSF linker 3SGGGSSGGGSSGGGSSGGGSNSEQ ID No: 24 UCSF linker 4SGGGGSGGSGGSGGSGSSGSN

[0039] As used herein, the term “fusion protein” means a single polypeptide comprising the sequences of a sushi domain, a peptidic linker, and an IL- 15. Such fusion proteins may further comprise additional sequences for leader, signaling, and connecting functions such as the hinge domain of the IL- 15 receptor alpha subunit. The “hinge domain of the IL- 15 receptor alpha subunit” is defined as the polypeptide sequence beginning with the first residue after the sushi domain and ending with the last residue prior to the first potential glycosylation site, and the human sequence is given as SEQ ID No: 25.SEQ ID No: 25 human hinge domainIRDPALVHQRPAPP

[0040] The fusion proteins used in the invention comprise a sushi domain of the IL- 15 receptor alpha-subunit (R) together with the hinge domain and an IL- 15 (I) connected using a peptidic linker (L). Example sequences of fusion proteins useful in the invention are given as SEQ ID No: 26-45.SEQ ID No: 26 RLIITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGG SGGGGSGGGS GGGGSLQNWV NVISDLKKIE DLIQSMHIDA TLYTESDVHP SCKVTAMKCF LLELQVISLE SGDASIHDTV ENLIILANNS LSSNGNVTES GCKECEELEE KNIKEFLQSF VHIVQMFINT SSEQ ID No: 27 RLI [N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSSG GSGGGGSGGG SGGGGSLQNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 28 RLI [N72D]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSSG GSGGGGSGGG SGGGGSLQNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 29 RLI [N72D,N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSSG GSGGGGSGGG SGGGGSLQNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 30 optimized RLI 1ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGS SGSSGSSGSS GSSGSSINNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 31 optimized RLI 1 [N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGS SGSSGSSGSS GSSGSSINNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 32 optimized RLI 1 [N72D]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGS SGSSGSSGSS GSSGSSINNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 33 optimized RLI 1 [N72D,N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGS SGSSGSSGSS GSSGSSINNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSAGNVTE SGCKECEELE EKNIKEFLQSFVHIVQMFIN TSSEQ ID No: 34 optimized RLI 2ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGS SSGSSSGSSS GSSSGSSNNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 35 optimized RLI 2 [N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGS SSGSSSGSSS GSSSGSSNNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 36 optimized RLI 2 [N72D]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGS SSGSSSGSSS GSSSGSSNNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 37 optimized RLI 2 [N72D,N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGS SSGSSSGSSS GSSSGSSNNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 38 optimized RLI 3ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGG GSSGGGSSGG GSSGGGSNNWVNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 39 optimized RLI 3 [N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGG GSSGGGSSGG GSSGGGSNNWVNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILANN SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 40 optimized RLI 3 [N72D]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGG GSSGGGSSGG GSSGGGSNNWVNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 41 optimized RLI 3 [N72D,N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGG GSSGGGSSGG GSSGGGSNNWVNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 42 optimized RLI 4ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKATNVAHWTTPS LKCIRDPALV HQRPAPPSGG GGSGGSGGSG GSGSSGSNNWVNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISLESGDASIHDT VENLIILANN SLSSNGNVTE SGCKECEELE EKNIKEFLQSFVHIVQMFIN TSSEQ ID No: 43 optimized REI 4 [N77A]ITCPPPMSVE HADIWVKSYS EYSRERYICN SGFKRKAGTS SETECVENKA TNVAHWTTPS EKCIRDPAEV HQRPAPPSGG GGSGGSGGSG GSGSSGSNNW VNVISDEKKI EDLIQSMHID ATEYTESDVH PSCKVTAMKC FEEEEQVISE ESGDASIHDT VENLIILANN SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 44 optimized RLI 4 [N72D]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGG GGSGGSGGSG GSGSSGSNNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSNGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TSSEQ ID No: 45 optimized RLI 4 [N72D,N77A]ITCPPPMSVE HADIWVKSYS LYSRERYICN SGFKRKAGTS SLTECVLNKA TNVAHWTTPS LKCIRDPALV HQRPAPPSGG GGSGGSGGSG GSGSSGSNNW VNVISDLKKI EDLIQSMHID ATLYTESDVH PSCKVTAMKC FLLELQVISL ESGDASIHDT VENLIILAND SLSSAGNVTE SGCKECEELE EKNIKEFLQS FVHIVQMFIN TS

[0041] The above sequences may be modified slightly without affecting their usefulness in the invention, for example to improve protein stability, expression, or other properties. In one embodiment, the sequence of S shows at least 85% sequence identity to one of SEQ ID Nos: 1-45. In some embodiments, the sequence of S shows at least 90% sequence identity to one of SEQ ID Nos: 1-45. In certain embodiments, the sequence of S shows at least 95% sequence identity to one of SEQ ID Nos: 1-45.

[0042] The water-soluble polymers P used in the invention are synthetic or natural polymers having a molecular weight from about 1,000 to about 100,000 daltons, preferablyfrom about 10,000 to about 60,000 daltons. In certain embodiments, P is a linear, multi-arm, or branched polyethylene glycol (PEG) having an average molecular weight between about 10,000 and about 60,000 Daltons. In certain embodiments, P is a linear or branched PEG having an average molecular weight between about 20,000 and 40,000 Daltons. Such high- molecular weight synthetic polymers are typically polydisperse, such that they comprise a mixture of polymers having different numbers of monomer units centered about an average number that describes the average molecular weight. For example, a linear PEG of average molecular weight 40,000 Dalton may comprise a distribution of polymers having 900 ± 100 monomer units. The dispersity or poly dispersity index (PDI) is the ratio of the weight average molecular weight (Mw) to the number average molecular weight (Mn); PEGs having a dispersity of 1.1 or lower are commercially available and useful in the present invention. See for example Zacchigna et al. Polymers 2011 3: 1076-90, which is incorporated herein by reference.

[0043] In one embodiment, P is a linear PEG having the general formula Rc- (OCH2CH2)y-Z wherein Rcis a Ci-Ce alkyl capping group, y is between 400-1500, and Z is a reactive functionality useful for conjugation. In other embodiments, P is a branched PEG having from 2-8 branches. The branches may be connected using any of several connecting fragments known in the art, for example a 2-branched PEG may consist of two PEG chains connected through a tri-functional fragment wherein two functionalities connect to PEG chains and the third provides reactive group Z or a means of attaching to Z. Examples of such tri-functional fragments include glycerols and amino acids such as lysine and ornithine. More complex branching arrangements may be constructed by combining several simpler branched molecules; for example, a 4-branched PEG may be constructed by the appropriate combination of two 2-branched PEGs onto another trifunctional fragment. Many such branched PEGs are commercially available.

[0044] Reactive functionality Z may comprise a group suitable for direct conjugation to a natural protein, for example an aldehyde, succinimidyl ester or carbonate, or maleimide, or may comprise a group that is cross -reactive with a cognate group Z’ that has been added to the protein by chemical or genetic modification. In one embodiment, Z is an aldehyde and the conjugation is performed using reductive alkylation of x amine groups on the sushi domain S to provide a conjugate of formula (I) wherein L is absent (i.e., Px-S):When x = 1, it is known in the art that the N-terminal amine group of a polypeptide is preferentially alkylated (Kinstler et al., Adv Drug Deliv Rev 2002 54(4): 477-85; Ambrogelly et al. Protein J 2013 32(5): 337-42). Thus, in this embodiment of the invention a mixture of the polypeptide comprising an N-terminal sushi domain and the water-soluble polymer P comprising a reactive aldehyde group Z = CHO is reacted with a reductant such as sodium cyanoborohydride or a borane complex such as pyridine borane or 2-picoline borane under conditions so as to obtain preferential reductive alkylation on the N-terminal amine group and provide a conjugate of formula (I) Pi-S wherein L is absent and x = 1.

[0045] In another embodiment, the reductive alkylation and polymer conjugation are performed in two steps. In a first step, a mixture of the polypeptide comprising an N-terminal sushi domain S and a linker L having the formula (II)Z’ -(CH2CH2O)n(CH2)pC(=O)-NH-(CH2)r(C=O)H (II) wherein Z’ is a reactive functionality that is cross -reactive with reactive functionality Z on P to form a connecting group Z*; n = 0-100; p = 1-6; and r = 1-6; is treated with a reductant such as sodium cyanoborohydride or a borane complex as described above to provide a linker-protein Lx-S having formula (III)[Z’-(CH2CH2O)n(CH2)PC(=O)-NH-(CH2)rCH2-NH]x-S (III) wherein x = 1-5 amine groups on polypeptide S have been reductively alkylated with linker L to provide formula (III), Lx-S. When x = 1, it is known in the art that the N-terminal amine group of a polypeptide is preferentially alkylated as described above. The reductive alkylation using sodium cyanoborohydride is optimally performed using 1-3 equivalents of a linker of formula (II) at temperatures between 0 and 22 °C, preferably at 2-8 °C, buffered at a pH between 4 and 8, preferably at a pH between 5 and 7. The resulting linker-polypeptide LXS is isolated, optionally as a mixture wherein x = 0-5, and then in a second step is reacted with a water-soluble polymer P-Z comprising an cognate-reactive group Z that reacts solely with Z’ to form a connecting functionality Z* so as to form the conjugate [P-L]x-S of formula (I).

[0046] The cognate pairs of reactive connecting groups Z and Z’ are chosen so as to provide orthogonal reactivity towards functional groups otherwise present on polypeptide S, and when allowed to react for a connecting group Z* that joins P and L. As examples, when one of Z and Z’ is an alkyne or cycloalkyne such as a cyclooctyne or bicyclo[6.1.0]nonyne, the other of Z and Z’ is azide and Z* is a 1,2,3-triazole. When one of Z and Z’ is thiol, the other of Z and Z’ is maleimide, vinylsulfone, or haloacetamide and Z* is a thioether. In one embodiment, one of Z and Z’ is a cyclooctyne such as 5-hydroxycyclooctyne or diazabicyclooctyne (DBCO), or a bicyclo[6.1.0]nonyne such as BCN ((lR,8S,9s)- bicyclo[6.1.0]non-4-yn-9-ylmethanol), and the other of Z and Z’ is an azide.

[0047] The resulting conjugates [P-L]x-S having formula (I) (including Px-S wherein L is absent) may be purified using standard chromatographic techniques, including ion-exchange and size exclusion chromatographies. As addition of PEG to polypeptides generally renders them less adherent to ion-exchange media, with increasing PEGylation leading to decreasing adherence, it is typically possible to separate the [P-L]x-S species by ion-exchange chromatography, with [P-L]x-S species requiring higher salt concentrations to elute as x decreases. Size exclusion chromatography can also be used to further separate various [P- L]x-S species.

[0048] Biological activity of the conjugates of formula (I) is measured in vitro using a CTLL-2 cell-based proliferation assay or a commercial U2OS dimerization bioassay using an IL2 / IL15bg assay kit. The in vitro activity of [P-L]i-S wherein P is a 40kDa linear PEG, x = 1, and L-S is formula (III) wherein Z’ = N3, n=4, p=2, r=2, and S = SEQ ID No: 25 is found to be reduced relative to native polypeptide S: in the CTLL-2 assay, this [P-L]i-S showed a 3.5x reduction in potency relative to S while in the U2OS assay the reduction was 7x.

[0049] Despite this reduction in potency, pharmacodynamic measurements revealed a dramatic and unexpected increase in the ability of [P-L]i-S to stimulate the in vivo proliferation and maintenance of T cells and NK cells in rodent models. Time courses for the number of several immune cell types were measured by flow cytometry after s.c. administration to mice of either [P-L]i-S, free S, or a slow-release form of S described in W02020 / 219943. As the time courses for the various cell types differ, it is advantageous to compare the integrated (cell number)»(time) or “AUC” (area under the curve) for each cell type analogously to a pharmacokinetic experiment measuring total drug exposure. Table 3 shows that [P-L]i-S produces 22x the AUC for NK cells, 13x the AUC for CD8+T cells, and21x the AUC for CD44hlCD8+T cells relative to an equal dose of free S. The fraction of these cells expressing the Ki67+ proliferation marker was similarly increased.

[0050] As it was possible that the dramatic increase in pharmacodynamic effect of the conjugate over free protein might be due to a simple increase in in vivo half-life as anticipated for PEGylation of a protein, the pharmacokinetics of the conjugate was compared with a previously-described slow-release preparation of the same protein (PCT Publication WO2020 / 219943) (Table 2). This demonstrated the half-life extension for [P-L]i-S expected from PEGylation (ti / 2 = 15 h) as compared to free protein (ti / 2 = 3 h); however, the slow- release preparation gave an even longer half-life for released S (ti / 2 = 30 h). Thus, a comparison of the pharmacodynamic effects of [P-L]i-S and the slow-release preparation of S at equal doses showed superiority of [P-L]i-S that would be unexpected if the improved properties of [P-L]i-S were due simply to half-life extension arising from PEGylation. Strikingly, while the AUCs for NK cells were similar ([P-L]i-S being 2x that of the slow- release preparation), the AUC for CD8+T cells for [P-L]i-S was 8x and the AUC for CD44hlCD8+T cells was lOx that from the slow-release preparation. The fraction of these cells expressing the Ki67+ proliferation marker was similarly increased. [P-L]i-S thus unexpectedly shows selectivity for increasing the number of CD8+T cells over NK cells that is not observed with the slow-release preparation of S. This selective increase in CD8+T cells is observed in the target organ spleen as well as in the peripheral blood, as shown in Table 4. Moreover, reported half-life extended IL- 15 agonists do not show the selective increase of CD8+T cells over NK cells, rather they typically show a selective increase in NK cells over CD8+T cells, indicating the PD response [P-L]i-S would not have been predicted. This selective expansion of antitumor CD8+T cells over immunotoxic NK cells may translate into improved efficacy for [P-L]i-S over free S or slow-release preparations of S or other half-life extended IL- 15 agonists.

[0051] Also unexpected was the observation of increased tolerability for [P-L]i-S relative to the slow-release preparation of S. Whereas s.c. administration of slow-release preparation of S at doses greater than 10 pg resulted in severe injection site reactions with formation of necrotic skin lesions, [P-L]i-S was well-tolerated up to 20 pg; while 40 pg still did not result in injection- site lesions, some body weight loss was observed.

[0052] The conjugates of the invention are expected to be useful in the treatment of cancers and other diseases and conditions characterized by undesired cellular proliferation.The selective proliferation of antitumor CD8+T cells over immunotoxic NK cells may provide efficacy while reducing the toxicity often associated with immunotherapy. The conjugates of the invention may also find use in the treatment of various infectious diseases including bacterial, viral, and parasitic infections, and as vaccine adjuvants through their effects on naive T cells and CD8+memory T cells.

[0053] The conjugates of the invention may be formulated using standard methods known in the art, and may be provided as sterile solutions for injection. Administration may be through any of the standard routes, including intravenous, subcutaneous, intramuscular, intraarticular, and intravitreal injections.

[0054] Treatments with the conjugates of the invention may be enhanced through combination with appropriate antibodies, for example anti-PD-1, anti-PD-Ll, anti-CD40, anti-CTLA-4, anti-OX40, anti-CD20, anti-CD52, anti-CCR4, and anti-CD38. IL- 15 cytokine therapeutics showing selective expansion of CD8+ T cell over NK cells may combine with appropriate therapeutic antibodies to have potent and durable anti-tumor responses.Numerous immune checkpoints, naturally occurring negative regulatory pathways, limit the immunological response of T cells and their effector function. Immune checkpoints including PD-1, CTLA-4, LAG3, and CD40, help balance T-cell activation and effector function through inhibitory signaling. Antibodies targeting these checkpoints have been shown to increase effector T cells and reactive exhausted T cells. Therefore, the combination of CD8+ T cell agonists with antibodies targeting immune checkpoints will not only expand the absolute number of CD8+ T cells but also increase the activation state and effector function of the cells. Taken together, these combinations potential to achieve effective and durable responses in the clinical oncology setting while avoiding NK cell dependent toxicities. Examples of anti-PD-1 antibodies include pembrolizumab (Keytruda), nivolumab (Opdivo), cemiplimab (Libtayo), and dostarlimab (Jemperli). Examples of anti-PD-Ll antibodies include atezolizumab, durvalumab (Imfinzi), and avelumab (Bavencio). Examples of anti- CTLA-4 antibodies include Ipilimumab (Yervoy) and tremelimumab (Imjudo).

[0055] As shown in Example 8, am exemplary pegylated conjugate of the disclosure combined with anti-PD- 1 therapy significantly extended median survival and achieved complete tumor regression in a subset of mice, demonstrating its potential to synergize with immune checkpoint inhibitors.

[0056] Beyond checkpoint inhibition regimens, pegylated conjugates of the disclosure hold promise in adoptive cell therapies by sustaining CAR-T cell persistence and functionality, as well as in minimal residual disease (MRD) settings by maintaining memory CD8+ T-cell populations for long-term immune surveillance and reducing relapse risks. By promoting sustained memory CD8+ T-cell responses and modulating innate cytotoxicity, pegylated conjugates of the disclosure offers advantages over existing IL- 15 agonists in enhancing tumor- specific responses while reducing toxicity risks across diverse cancer contexts.

[0057] In some embodiments, the pegylated conjugates of the disclosure enhance CAR-T cell antitumor activity. In some embodiments the CAR-T cell therapy and the pegylated conjugate are administered to a patient with a hematological malignancy. In some embodiments, the conjugates are administered to patients with a B cell malignancy. In some embodiments, the conjugates are administered to patients with leukemia, lymphoma or multiple myeloma. In some embodiments, the conjugates are delivered in patients having a solid tumor. In some such embodiments, the solid tumor is lung cancer, kidney cancer, bone cancer, pancreatic cancer, colorectal cancer, prostate cancer, breast cancer, ovarian cancer, or gastrointestinal cancer. In other embodiments, the solid tumor is a carcinoma, sarcoma, lymphoma, glioma, glioblastoma, or neuroblastoma. In some embodiments, T cells in combination with pegylated conjugates of the disclosure are expanded ex vivo and administered to a patient in need thereof (e.g., a human cancer patient).

[0058] In some embodiments, the pegylated conjugates of the disclosure are used to treat cancer patients that retain MRD. In some such embodiments, treatment prevents or minimizes relapse of the cancer. In some embodiments the cancer associated with MRD is multiple myeloma. Without being bound by theory, it is hypothesized that the pegylated conjugates of the disclosure prevent relapse by sustaining memory T-cell homeostasis and promoting adaptive immunity.ENUMERATED EMBODIMENTS

[0059] The following enumerated embodiments are representative of some aspects of the invention.1. A water-soluble construct comprising a polymer-protein conjugate, wherein the polymer-protein conjugate comprises a water-soluble polymer covalently bonded to an IL- 15 receptor alpha sushi domain (IL15RocSu).2. The water-soluble construct of embodiment 1 , wherein the water-soluble polymer is selected from a polyethylene glycol (PEG), polysaccharide, polypeptide, polypropylene glycol), poly (enol), poly(vinylpyrrolidone), poly(hydroxyalkyl methacrylamide), poly(hydroxyalkyl methacrylate), vinyl alcohol, polyphosphazene, poly(alpha-hydroxy acid), polyoxazoline, poly(N-acrylomorpholine), ethylene glycol / propylene glycol copolymer, poly(hydroxyalkyl methacrylate), poly(oxyethylated polyol), polyacrylic acid, polyacrylic acid copolymers, poly(maleic acid), poly(vinyl phosphoric acid), poly(2-vinyl-l- methylpyridinium bromide), poly(l-lysine hydrobromide), poly(ethylene oxide-b-propylene oxide), poly(styrenesulfonic acid), poly(N-vinyl acetamide), poly(vinyl acetate), poly(vinylphosphonic acid), poly(vinyl methyl ether), and poly(l-lysine hydrobromide).3. The water-soluble construct of embodiment 1, wherein the water-soluble polymer is a polyethylene glycol (PEG) having an average molecular weight between about 10,000 and about 60,000 Daltons.4. The water-soluble construct of embodiment 3, wherein the PEG is linear.5. The water-soluble construct of embodiment 3, wherein the PEG is branched.6. The water-soluble construct of embodiment 3, wherein the PEG is a multi-armedPEG.7. The water-soluble construct of any one of embodiments 1-6, wherein IL15RocSu has a sequence with at least 85% sequence identity to SEQ ID No. 2.8. The water-soluble construct of any one of embodiments 1-6, wherein IL15RocSu has a sequence with at least 90% sequence identity to SEQ ID No. 2.9. The water-soluble construct of any one of embodiments 1-6, wherein IL15RocSu has a sequence with at least 95% sequence identity to SEQ ID No. 2.10. The water-soluble construct of any one of embodiments 1-6, wherein IL15RocSu has a sequence of SEQ ID No. 2.11. The water-soluble construct of any one of embodiments 1-10, wherein the IL15RocSu is directly bonded to the water-soluble polymer.12. The water-soluble construct of embodiment 11, wherein the bond between the IL15RocSu and the water-soluble polymer is formed via a reaction between an aldehyde, succinimidyl ester, carbonate or maleimide moiety on the water-soluble polymer and an amino group on the IL15RocSu.13. The water-soluble construct of embodiment 12, wherein the amino group on the IL15RocSu is an N-terminal amino group.14. The water-soluble construct of embodiment 12 or embodiment 13, wherein the reaction is a reductive alkylation between an aldehyde on the water-soluble polymer and an amino group on the IL15RocSu.15. The water-soluble construct of any one of embodiments 11-14, wherein the bond between the IL15RocSu and the water-soluble polymer is not cleaved in vivo following administration to a human subject.16. The water-soluble construct of any one of embodiments 1-10, wherein the IL15RocSu is bonded to the water-soluble polymer through a linker.17. The water-soluble construct of embodiment 16, wherein the linker is a non-cleavable linker.18. The water-soluble construct of embodiments 16 or embodiment 17, wherein the linker is bonded to the N-terminus of the IL15RocSu.19. The water-soluble construct of any one of embodiments 16-18, wherein the linker comprises a click product formed from a reaction between a first click handle and a second click handle.20. The water-soluble construct of embodiment 19, wherein the first click handle comprises an alkyne or cycloalkyne group and the second click handle comprises an azide group.21. The water-soluble construct of embodiment 20, wherein the first click handle comprises a 5 -hydroxy cyclooctyne or diazabicyclooctyne (DBCO).22. The water-soluble construct of embodiment 19, wherein the click product comprises a 1,2,3-triazole moiety.23. The water-soluble construct of any one of embodiments 1-22, further comprising IL- 15 or a mutein thereof.24. The water-soluble construct of embodiment 23, wherein the IL- 15 or a mutein thereof has a sequence with at least 85% sequence identity to SEQ ID No. 5.25. The water-soluble construct of embodiment 23, wherein the IL- 15 or a mutein thereof has a sequence with at least 90% sequence identity to SEQ ID No. 5.26. The water-soluble construct of embodiment 23, wherein the IL- 15 or a mutein thereof has a sequence with at least 95% sequence identity to SEQ ID No. 5.27. The water-soluble construct of embodiment 23, wherein the IL- 15 has a sequence of SEQ ID No. 5.28. The water-soluble construct of embodiment 23, wherein the mutein of IL-15 has a sequence of SEQ ID No. 7. 129. The water-soluble construct of embodiment 23, wherein the mutein of IL- 15 has a sequence of SEQ ID No. 8.30. The water-soluble construct of embodiment 23, wherein the mutein of IL- 15 has a sequence of SEQ ID No. 9.31. The water-soluble construct of any one of embodiments 23-30, wherein the IL- 15 is noncovalently associated with the polymer-protein conjugate.32. The water-soluble construct of any one of embodiments 23-30, wherein the IL- 15 is covalently associated with the polymer-protein conjugate.33. The water-soluble construct of embodiment 32, wherein the IL- 15 is covalently linked to the IL15RocSu.34. The water-soluble construct of embodiment 33, wherein the IL- 15 is covalently linked to IL15RocSu through a peptidic linker.35. The water-soluble construct of embodiment 34, wherein the peptidic linker has a sequence selected from SEQ ID No. 20, SEQ ID No. 21, SEQ ID No. 22, SEQ ID No. 23, or SEQ ID No. 24.36. A method of treating cancer comprising treatment of a patient in need thereof with a water-soluble construct of any one of embodiments 1-35.37. The method of embodiment 36, further comprising administering to the patient an immune checkpoint inhibitor.38. The method of embodiment 37, wherein the immune checkpoint inhibitor is a PD-1 inhibitor, a PD-L1 inhibitor or a CTLA-4 inhibitor.39. The method of embodiment 38, wherein the immune checkpoint inhibitor is an anti- PD-1 antibody.40. The method of embodiment 39, wherein the anti-PD-1 antibody is selected from pembrolizumab, nivolumab, cemiplimab, atezolizumab, dostarlimab, durvalumab, and avelumab.41. The method of embodiment 38, wherein the immune checkpoint inhibitor is an anti- PD-L1 antibody.42. The method of embodiments 38, wherein the immune checkpoint inhibitor is an anti- CTLA4 antibody.43. The method of embodiment 42, wherein the anti-CTLA4 antibody is ipilimumab or tremelimumab.44. The method of any one of embodiments 36-43, further comprising administering an anti-CD40 antibody, an anti-OX40 antibody, an anti-CD20 antibody, an anti-CD52 antibody, an anti-CCR4 antibody or an anti-CD38 antibody to the patient.45. The method of any one of embodiments 36-44, wherein the water-soluble construct is administered subcutaneously to the patient.46. A conjugate comprising a polyethylene glycol covalently and stably bonded to a sushi domain from an IL- 15 receptor alpha subunit.47. The conjugate of embodiment 46, wherein the polyethylene glycol is linear or branched, and has an average molecular weight between about 10,000 and about 60,000 Daltons.48. The conjugate of embodiment 46, wherein the polyethylene glycol is bonded to the sushi domain through a linker.49. The conjugate of embodiment 46, further comprising an IL- 15.50. The conjugate of embodiment 49, wherein the IL- 15 is noncovalently associated with the PEGylated sushi domain.51. The conjugate of embodiment 49, wherein the IL- 15 is part of a fusion protein comprising the sushi domain, a peptidic linker, and an IL- 15.52. The conjugate of any one of embodiments 46-51, wherein the sushi domain has a sequence with at least 85% sequence identity to SEQ ID No: 2.53. The conjugate of any one of embodiments 46-51, wherein the IL-15 has a sequence with at least 85% sequence identity to one of SEQ ID No: 5-19.54. The conjugate of embodiment 50, wherein the IL- 15 has a sequence with at least 85% sequence identity to one of SEQ ID No: 7-955. The conjugate of embodiment 51, wherein the fusion protein has a sequence with at least 85% sequence identity to one of SEQ ID No: 26-45.56. The conjugate of embodiment 51, consisting of a linear polyethylene glycol having an average molecular weight of 40,000 Daltons covalently and stably bonded to the N-terminal amine of SEQ ID No: 27.57. A method for the preparation of a conjugate of any one of embodiments 46-56, comprising: contacting a protein comprising a sushi domain S with a linker of formula (II) comprising a reactive group Z in the presence of a reducing agent so as to form a linkerprotein via reductive alkylation of the N-terminal amine of the sushi domain; optionally isolating the linker-protein; and contacting the linker protein with an activated polyethylglycol comprising cognate reactive group Z’ under conditions such that Z and Z’ react to form connecting group Z*.58. The method of embodiment 57, wherein Z is an azide and Z’ is a cyclooctyne or bicyclononyne.59. The method of embodiment 57, wherein S is a sushi domain having 85% sequence identity to one of SEQ ID No: 2.60. The method of embodiment 57, wherein S is a fusion protein having 85% sequence identity to one of SEQ ID No: 26-45.61. A method for the preparation of a conjugate of any one of embodiments 46-56, wherein S is a sushi domain noncovalently associated with an IL- 15, comprising the step of contacting a PEGylated sushi domain with an IL- 15 under conditions wherein the noncovalent complex is formed.62. A method of increasing the number of CD8+T cells in a subject comprising the step of administering a conjugate of any one of embodiments 45-56 to the subject.63. The method of embodiment 62, where in the conjugate comprises a polyethylene glycol of average molecular weight of between about 10000 and about 60000 Daltons attached to the N-terminal amine of a fusion protein having SEQ ID No: 26-45.64. The method of embodiment 62, wherein the conjugate comprises a polyethylene glycol of average molecular weight of 40000 Daltons attached to the N-terminal amine of a fusion protein having SEQ ID No: 27.65. A method of treating cancer or other condition benefitting from an expansion of CD8+T cells in a patient in need thereof comprising administration of a conjugate of any one of embodiments 1-11 to the patient.66. The method of embodiment 65, further comprising administering an immune checkpoint inhibitor.67. The method of embodiment 66, wherein the immune checkpoint inhibitor is a PD-1 inhibitor, PD-L1 inhibitor, or CTLA-4 inhibitor.68. The method of embodiment 67, wherein the immune checkpoint inhibitor is selected from pembrolizumab, nivolumab, cemiplimab, atezolizumab, dostarlimab, and avelumab.69. The method of embodiment 65, further comprising administering an anti-CD40 antibody, anti-OX40 antibody, anti-CD20 antibody, anti-CD52 antibody, anti-CCR4 antibody, or anti-CD38 antibody.EXAMPLES

[0060] The presently disclosed subject matter will be better understood by reference to the following Examples, which are provided as exemplary, and not by way of limitation.Preparation 1N3-PEG4-CHOLinker of formula (II) wherein Z’ = N3; n=4, p=2, and r=2

[0061] A solution of azido-PEG4-NHS ester (BroadPharm BP-20518) (41 mg, 110 umol) and 3,3-diethoxy-l-propylamine (37 uL, 230 umol) in 1.1 mL of acetonitrile was kept for 30 min, then partitioned between ethyl acetate and 5% aq. KHSO4. The organic phase was washed successively with water, sat. aq. NaHCO ,, and brine, then dried over MgSO4, filtered, and evaporated to provide 32 mg (69% of expected) of crude acetal. This was dissolved in CHCI3 (0.51 mL), and water (0.13 mL), and trifluoroacetic acid (0.13 mL) were added. After stirring vigorously for 1.5 h, the mixture was concentrated to dryness. The product was purified on SiO2 (4 g) using a gradient of 0-80% acetone in CH2Q2 providing the title compound (20 mg, 58 umol, 53% overall) as a pale-yellow oil. LC-MS: [M+H]+347.6.Preparation 2MeO-PEG4okDa-BCNWater-soluble polymer wherein Rc= Me; x -900; and Z = bicyclo [6.1.0] nonyne

[0062] A mixture of MeO-PEG40kDa-NH2 (200 mg, 5 umol), lR,8S,9s)- bicyclo[6.1.0]non-4-yn-9-ylmethyl succinimidyl carbonate (BCN-OSu) (2 mg, 6.9 umol, 1.4 Eq), and N,N-diisopropylethylamine (2 uL, 11.5 umol, 2.3 Eq) in 2 mL of acetonitrile was kept for 30 min, then evaporated. The residue was dissolved in 2 mL of THF and precipitated by drop wise addition to stirred MTBE (10 mL). The precipitate was collected and dried under vacuum to provide the product.

[0063] Other cyclooctyne-activated PEGs are prepared similarly, using the appropriatePEG-amine and an active ester or carbonate of the cyclooctyne, for example 5- hydroxycyclooctyne succinimidyl carbonate or dibenzocyclooctyne succinimidyl ester.Example 1Preparation of Lx-S of formula III wherein S = SEQ ID No: 27, Z’ = N3; n=4, p=2, and r=2

[0064] The following procedures are illustrated for the polypeptide of SEQ ID No: 27, but are generally applicable to any polypeptide S of the invention by substitution of the polypeptide having the appropriate SEQ ID No.

[0065] Optimization ofLi-S yield. In a total volume of 50 pL, reactions contained 39 nmol (850 pg, 780 pM, 1 equivalent) SEQ ID No: 27 in phosphate buffered saline (PBS, 9.8 mM phosphate, 137 mM NaCl, 2.7 mM KC1, pH 7.3) or 25 mM Citrate, pH 6.0, 500 mM NaCl and 0.05% Tween-20 (Buffer Cit), 1 to 5 equivalents of N3-PEG4-CHO (Preparation A; 13.5 - 67.5 pg, 0.78 - 3.9 mM), and 520 nmol (31 pg, 10 mM) of NaCNBH ,. After 14 hours 4°C, mixtures were analyzed by gel-shift assay as follows. Samples were treated with 10 equivalents of DBCO-PEGskDa (prepared as in Preparation B substituting DBCO-NHS ester for BCN-OSu) or BCN-PEG40kDa for 4 hours at 37 °C, and analyzed by SDS-PAGE. Here, DBCO-PEGsk or BCN-PEG40kDa reacts with the azide of linker-protein L-S by SPAAC and slows migration of the pegylated protein on SDS-PAGE; multiple PEG addition slows migration proportionally such that the band for L2-S moves more slowly than the band for Li- S, etc. The bands were quantitated by densitometry after staining with Coomassie Blue. SDS-PAGE analysis of the reaction mixtures showed the buffer had no significant effect on the reaction yield. Reaction mixtures that used 1 molar equivalents of linker showed a mixture comprising 54% unreacted S + 43% mono-alkylated (Li-S) + 3% di-alkylated (L2-S), whereas reactions that used 3 molar equivalents of linker showed a mixture comprising 51% Li-S + 38% L2-S + 11% tri-alkylated (L3-S) protein.

[0066] Preparation ofLi-S. In 0.41 111L of Buffer Cit, reaction mixtures contained 217 nmol (5 mg, 530 pM) SEQ ID No: 27, 217 nmol (75 pg, 530 pM, 1 Eq) of N3-PEG4-CHO and 5.2 pmol NaCNBH? (312 pg, 10 mM). The reaction was allowed to proceed 14 hours at 4 °C in the dark. Excess reagents were removed using a 3.5 mL PD-10 column (Cytiva) previously equilibrated in 25 mM Tris, pH 8.0 (Buffer A). Protein concentration was determined by A280 (S2so= 22960 M1cm1) using a NanoDrop spectrophotometer. Gel-shift analysis determined this Lx-S to comprise 48% unreacted S + 40% Li-S + 12% L2-S.Example 2 Conjugation of Lx-SPreparation of [P-L]i-S wherein P = 40 kDa PEG (n~910); Z* = triazole; and S = SEQ ID

[0067] The following procedures are illustrated for the polypeptide of SEQ ID No: 27, but are generally applicable to any polypeptide S of the invention by substitution of the polypeptide having the appropriate SEQ ID No.

[0068] In 1.5 mL of 25 mM Tris pH 8.0, 170 nmol of MeO-PEG40kDa-BCN (Preparation 2) was added to the Lx-S reaction mixture of Example 1 containing 217 nmol (~5 mg) of 40% Li-S and 12% L2-S. The reaction proceeded for 48 h at room temperature. The unreacted MeO-PEG-BCN was removed by incubation with 200 nmol of azido-Agarose beads (Click Chemistry Tools) for 18 hours at room temperature, followed by filtration with 0.2 pm pore size filters of both PEG40kDa- Agarose beads and excess azido-Agarose beads.

[0069] The conjugate was purified by ion exchange chromatography. A HiTrap Q FF (7.5 x 0.7 cm, 3 mL column volume (CV), GE Healthcare) resin was equilibrated with 5 CV of 25 mM Tris, pH 8.0 (Buffer A). The [P-L]x-S sample (x = 0-3) was loaded at 150 cm / h and unbound material was eluted with 3 CV of Buffer A. Using a flow rate of 150 cm / h, the sample was eluted using a linear gradient 0-5% Buffer B (25 mM Tris, pH 8, 1 M NaCl) over 10 CV followed a second linear gradient from 5% to 35 % Buffer B over 5 CV. Protein content and purity in collected fractions were analyzed by SDS-PAGE. Fractions containing pure [P-L]i-S were pooled, concentrated and buffer exchanged to PBS by diafiltration with a 10 kDa pore size Amicon filter. Purified [P-L]i-S was stored at -80 °C.Example 3 [P-L]x-S activity assays

[0070] The potencies of [P-L]x-S conjugates were tested using two cell-based assays, an engineered U2OS cell line expressing the dimeric IL-2 / IL- 15 R / 5% receptor and a cell proliferation assay of CTLL-2 cells expressing the IL-2oc and IL-2 / IL- 15R^ receptors. The following procedures are illustrated for the polypeptide of SEQ ID No: 27, but are generally applicable to any polypeptide S of the invention by substitution of the appropriate SEQ ID No.

[0071] CTLL-2 cell-based proliferation assay. CTLL-2 cells were cultured in RPMI-1640 medium supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, 10% fetal bovine serum, and 10% T-STIM with Con A. Passages were performed twice a week and cell density was maintained between 20,000 and 200,000 cells / ml. Prior to assaying, cells were washed three times with RPMI-1640 media containing 10% fetal bovine serum, 2 mM L- glutamine and 1 mM sodium pyruvate. Cells were then resuspended in the RPMI-1640 supplemented media at 500,000 cells / ml and 50,000 cells / well were plated (100 pL). Plates were incubated for four hours at 37 °C, 5 % CO2 before addition of 100 l / well of an IL- 15 agonist serially diluted in RPMI-1640 supplemented media. After addition of agonists, plates were incubated 48 hours at 37 °C 5 % CO2. Then, CellTiter96 Aqueous One Solution was added (40 pL) and the plate was incubated for four hours at 37°C, 5% CO2. The plate was then read at 490 nm using a Spectramax i3 plate reader. EC50 values were estimated by GraphPad Prism.

[0072] U2OS Dimerization Bioassay. A U2OS cell-based assay kit for IL-2RPy binding was performed according to the manufacturer’s instructions (DiscoverX, Part #93- 0998E3CP5). Briefly, cells were plated (100 pL, -5,000 cells / well) in 96 well white-walled assay plates and incubated for 48 hours at 37°C, 5% CO2. Cells were treated with varying concentrations of the agonist and incubated for an additional 6 hours at 37°C, 5% CO2. Treated cells were then incubated with the PathHunter reagent substrate for 1 hour at 20°C, protected from light. Plates were read for chemiluminescence signal detection using a Spectramax i3 plate reader using a 250 ms integration time.

[0073] At saturating concentrations, S and [P-L]i-S resulted in similar maximal levels of response of the U2OS cells expressing IL-2 / IL- l 5R ?%. However, [P-L]i-S was 7-fold less potent (EC50 = 1.2 nM) than free S (EC50 = 0.17 nM) (Figure 5A). Similarly, both free S and [P-L]i-S resulted in similar maximal levels of cell proliferation in CTLL-2 cells expressing the IL-2oc and IL-2 / IL- l 5R ?%. However, the potency of [P-L]i-S (EC50 = 1.5 nM) was -3.5 fold less potent than free S (EC50 = 0.42 nM) and -500 fold less potent than IL- 15 (EC50 = 2.9 pM) (Figure 5B). Thus, pegylation of S attenuates bioactivity, although [P-L]i-S remains an agonist of the IL-2 / IL-15RPyc.Example 4Pharmacokinetics of [P-L]i-S in immunocompetent mice

[0074] Pharmacokinetics of the conjugate [P-L]i-S from Example 2, administered either i.v. or s.c., were measured in normal, male C57BL / 6J mice. Dosing solutions (1.7 nmol / mL - 17.2 nmol / mL) were prepared by diluting the [P-L]i-S stock in PBS, followed by filter sterilization. Protein concentration in dosing solutions was assessed by A280 (£2x0= 22,960 M1cm1). The concentration versus time plot from a single intravenous or subcutaneous injection of 10 pg [P-L]i-S (0.43 nmol) show high plasma concentrations are achieved (Figure 6). When dosed i.v., a CMAX of 240 nM and an elimination ti / 2 of 15 h was observed. When dosed s.c., a CMAX of 97 nM and a half-life of 15 h was measured (Table 1). The bioavailability of [P-L]i-S administered s.c., determined using the AUC, was 46%. Assuming dosing linearity PK parameters, when compared to previously reported pharmacokinetic parameters of S having SEQ ID No: 24, a 5-fold increase in half-life was observed.Moreover, the total exposure of [P-L]i-S, dosed s.c. or i.v., was >35 and 80-fold more than that of free S, respectively (Table 1).Table 1. Summary of pharmacokinetic parameters; S = SEQ ID No: 27Cmax (nM) 240 97 4.9 tmax8 24 1(h) tl / 215 15 3(h)Total AUC7811 3611 38 (nM.h)%F 46 a. Data reported by Bessard et al. Mol. Cancer Ther. 2009.Example 5Dose titration Pharmacodynamics of [P-L]i-S

[0075] Single doses (0.17-1.72 nmol) of the conjugate of Example 2 were administered subcutaneously to normal, male C57BL / 6J mice. Blood samples were collected in EDTA collection tubes over a predefined time course (-2, 3, 5, 7, 14, 21, and 28 days post dose). ForFlow Cytometry Analysis, EDTA whole blood (25 pL) was transferred to a 96 deep well plate and incubated with a fixable viability dye to label dead cells. FcR II / II receptors were blocked with CD16 / 32 (20 pL, 1.6 pg) before staining for cell-surface antigens (all reagents from Invitrogen). The whole blood samples were incubated at 4°C for 30 minutes with previously determined optimal antibodies concentrations for surface staining of peripheral blood mononuclear cells (PBMCs) (Table 2). Red blood cells (RBCs) were lysed and PBMCs were fixed by incubation with 2 mL of 1-Step Fix / Lyse solution for 30 minutes at room temperature. Fixed cells were washed once with 2 mL permeabilization buffer and then intracellularly stained (Ki-67-APC) using a 30 minutes incubation period. After the cells were stained, samples were washed twice with 2 mL of FACS buffer. The cells were resuspended in 400 pL FACS buffer and stored at 4°C until analysis. Stained single cell suspensions were read using a Attune NxT flow cytometer (BD Biosciences) and analyzed using Flow Jo cytometry analysis software (TreeStar, Ashland, OR).

[0076] The absolute cell numbers of the samples were determined by direct cell analysis (volumetric counting by the Attune NxT). Positive populations were identified based on fluorescence minus one control. CD8+and memory CD8 T cells were identified as CD3+CD4'CD8+and CD3+CD4'CD8+CD44hlgh, respectively; Treg cells were identified as CD3+CD8'CD4+CD25+FoxP3+; NK cells were identified as CD3'NK1.1+. Cells with proliferative capacity were defined as Ki67+.Table 2. Antibodies used for immunopheno typing.

[0077] A dose titration response of the target immune cells to treatment with the conjugate [P-L]i-S of Example 2 five days post treatment was compared to historical data of mice treated with the slow-release preparation of PCT Publication WO2020 / 219943 (10 pg, 0.43 nmol), or a bolus SC injection of 10 pg free S or two doses of 2 pg (2x 0.09 nmol) free S separated by 48 h. Pharmacodynamic measurements focused on the expansion of CD8+and CD44hlCD8+T cells, and NK cells. On day 5 of the study, mice were bled and the PBMCs were immunophenotyped using flow cytometry. A large expansion of peripheral lymphocytes was measured in mice treated with [P-L]i-S, which was not observed with bolus doses of free S or the slow-release preparation of S conjugate (Figure 8). The high lymphocyte counts returned to baseline levels by day 21 (Figure 8). In addition to increase lymphocyte counts, a robust increase in the frequency (% total PBMCs) as well as the absolute number of CD8+T cells, CD44hlCD8+T cells and NK cells (Figure 9A&B) was observed in mice treated with [P- E]i-S. The fold-change in absolute cell numbers (Figure 9C) showed a ~50- and ~200-fold expansion for CD8+T cells and CD44hlCD8+T cells, respectively, and 60-fold expansion for NK cells. Regarding CD4+T cells and B cells on day 5, a small increase in the absolute cell number of both cell types was evident (Figure 10). Thus, the large increase in the total number of lymphocytes was due to the expansion of target immune cells, particularly CD8+T cells, CD44hlCD8+T cells, and NK cells.Example 6Eongitudinal Pharmacodynamic response of [P-E]i-S

[0078] The longitudinal pharmacodynamic response following a single administration of the [P-E]i-S of Example 2 (10 pg, 0.43 nmol) was measured. Over a period of 28 days, mice were bled and the PBMCs were stained and immunophenotyped by flow cytometry to measure the total exposure of the target immune cells and the duration in which the cells remained expanded above the baseline measure. This data was then compared to that of two doses of free S (2x 2 pg, 2x 0.09 nmol) separated by 48 h or a single injection of equal molar doses of free S (10 pg, 0.43 nmol) or the slow-release preparation of S (10 pg, 0.43 nmol). As in Example 5, five days after treatment, [P-E]i-S induced ~20-fold expansion in the total number of lymphocytes, while free S and the slow-release preparation of S induced less than a 2-fold expansion. In all treatment groups, the total lymphocytes count returned to basal levels by day 14.

[0079] Treatment with [P-L]i-S not only expanded the total number of lymphocytes, but also induced a massive expansion in the absolute number of NK, CD8+T cells and CD44hlCD8+T cells (Figure 11A). On day 5, NK cells, CD8+T cells and CD44hlCD8+T cells were 100-, 65, 240-fold over the baseline measurement, respectively (Figure 11B). NK cells remained above baseline for up to 14 days, whereas CD8+T cells and CD44hlCD8+T cells remained above baseline for at least 28 days. Impressively, CD44hlCD8+T cell population was 5-fold above the baseline count on day 28. Conversely, a modest 2- to 6-fold change in all absolute cell counts was measured from bolus injections of free S or the slow-release preparation of S, all of which returned to baseline levels by day 14. The proliferation response of the target immune cells was also measured using the intracellular Ki-67 marker (Figure 11C). Bolus injections of free S resulted in a high percentage (75-85%) of Ki67+NK, CD8+and CD44hlCD8+T cells. [P-L]i-S also gave a similar high percentage of Ki67+CD8+and CD44hlCD8+T cells, but only approximately 50% of NK cells were Ki67+. However, the duration of proliferation lasted at least 21 days. The ratio of CD8+to CD4+T cells shows that [P-L]i-S has a significant and long-lasting expansion of CD8+T cells compared to CD4+T cell, which is not observed for free S or the slow releasing S (Figure 12A). The ratio of the CD8+and CD44hlCD8+T cells to NK cells best exemplifies that [P-L]i-S favors expansion of cytotoxic T cells over NKs compared to free S and the slow-release preparation of S (Figure 12B &C). In addition, these ratios show [P-L]i-S is differentiated from other reported halflife extended IL- 15 agonists reported in literature (Figure 12D &E).

[0080] Interestingly, equimolar dosing of NKTR-255 (PEGylated-IL-15) has the opposite PD effect as [P-L]i-S, favoring high selectivity for NK cells over CD44hlCD8+T cells. Taken together, these findings indicated [P-L]i-S is able to induce the proliferation of target immunes cells, expanding the absolute number of NK, CD8+and CD44hlCD8+T cells more than free S or the slow-release preparation of S, and furthermore uniquely showing selective expansion of cytotoxic T cells over NK cells.

[0081] Finally, AUCo-28d measurements were used compare the pharmacodynamics of [P- L]i-S to other IL-15 agonists (Figure 11, Table 3). The AUCo-28d is calculated from the under the curve and above the baseline cell count over 28 days, and can be used to quantify the extent of cell expansion. [P-L]i-S had a 15- to 40-fold larger AUCo-28d for all target immune cells when compared two doses of free S (2 pg) separated by 48 h and to an equimolar bolus injection of free S. While this may be expected for a pegylated molecule with increased half-life, [P-L]i-S also had AUCo-28d measurements for NK cells, CD8+T cells and CD44hlCD8+T cells that were 2-, 8- and 10- fold higher, respectively, when compared to an equimolar dose of a slow-release preparation of S showing the same half-life for released S. As these two agonists have the same half-life, it is unlikely half-life extension alone is responsible for the large AUCo-28d. Thus, based on the cell phenotypes measured here, [P-L]i-S has superior pharmacodynamics compared to free S (SEQ ID No: 27) and slow-release S both in terms of total number of cells as well as the ratio of T cells to NK cells.Table 3. AUCo-28d in PBMCsaCell Marker 2 pg x2 10 pg '^g10 pgS S i °W[P-L]i-S release 8CD3 NK1.1+1.1 1.0 11 22CD8+2.0 4.3 6.8 54CD44hiCD8+1.4 2.0 4.5 43Ki67+CD3+NK1.1+1.1 0.61 6.3 9.4Ki67+CD8+1.1 1.2 3.5 27Ki67+CD44hiCD8+0.72 0.65 2.3 21 a. AUC values are cells / pLxlO'3xdayExample 7Pharmacodynamic response of [P-L]i-S in the spleen and PBMCs

[0082] Single doses (1 & 10 pg, 0.043 & 0.43 nmol) of the [P-L]i-S conjugate of Example 2 or free S (10 pg, 0.43 nmol) were administered subcutaneously to male C57BL / 6J mice to measure the longitudinal pharmacodynamic response in both the spleen and blood to determine if [P-L]l-S’s effect observed in blood is reflected in a secondary lymphoid organ. Over a predefined time course (-2, 5, 7, 14, & 21days post dose), animals (n=3-4 / group / day) were sacrificed, and the blood and spleen were harvested.

[0083] To immunophenotype PBMCs, EDTA whole blood (25 pL) was transferred to a 96 deep well plate and incubated with a fixable viability dye to label dead cells. FcRylPIII receptors were blocked with CD16 / 32 before staining for cell-surface antigens. The wholeblood samples were incubated at 4°C for 30 minutes with previously determined optimal antibodies concentrations for surface staining of peripheral blood mononuclear cells (PBMCs) (Table 2). RBCs were lysed using IX RBC Lysis Buffer (Invitrogen) before PBMCs were fixed and permeabilized with the Foxp3 Transcription Factor Staining Buffer Set (Invitrogen) for intracellular marker staining following the manufacturer’s instructions. Similarly, single cell suspensions of splenocytes were obtained from harvested spleens following mechanical disruption and filtering through a 40-pm cell strainer. RBCs were then lysed using IX RBC Lysis Buffer (Invitrogen) and ~2.5xl05splenocytes per sample were transferred to a 96 deep well plate. The FcRylLIII receptors were blocked with CD 16 / 32 prior to staining for cell-surface antigens. Samples were incubated at 4°C for 30 minutes with previously determined optimal antibodies concentrations for surface staining of splenocytes (Table 2) and then, cells were fixed and permeabilized with the Foxp3 Transcription Factor Staining Buffer Set (Invitrogen) for intracellular marker staining following supplier instructions. Stained single cell suspensions were read using a Attune NxT flow cytometer (BD Biosciences) and analyzed using FlowJo cytometry analysis software (TreeStar, Ashland, OR) as described in Example 5.

[0084] Following treatement with 10 pg of [P-L]i-S, a large expansion in the total number of lymphocytes was measured on day 5 in both PBMCs (10-fold) and the spleen (2- fold). Immunophenotyping of the PBMCs showed NK, CD8+T cells and CD44hlCD8+T cells were maximally expanded between days 5 and 7, being ~18-, 22- and 100-fold above baseline measurements, respectively (Figure 13A&B). NK cells remained above baseline for up to 14 days, whereas CD8+T cells and CD44hlCD8+T cells remained above baseline for at least 21 days. In the spleen, NK, CD8+T cells and CD44hlCD8+T cells were maximally expanded between days 5 and 7, being ~5-, 6- and 21-fold above baseline measurements, respectively (Figure 14A&B). NK cells remained above the baseline measurement for up to 14 days whereas CD8+and CD44hlCD8+T cells remained above the baseline measurement for at least 21 days. The proliferation response of the target immune cells was also measured using the intracellular Ki-67 marker (Figure 13C and 14C). Both [P-L]i-S and bolus free resulted in a high percentage (75-90%) of Ki67+NK, CD8+and CD44hlCD8+T cells, with the duration of proliferation lasting up to 21 days. In contrast to 10 pg of [P-L]i-S, bolus SC dosing of free S (10 pg) resulted in a 1.5-, 2.5-, and 4-fold expansion of NK cells, CD8+T cells and CD44hlCD8+T cells, respectively, in PBMCs (Figure 13A&B). There was no measured expansion of the NK and CD8+T cells and only a 1.5-fold expansion ofCD44hlCD8+T cells in the spleen (Figure 14A&B). Interesting, treatment with 1 pg of [P- L]i-S resulted in 2- to 4-fold expansion of NK, CD8+T cells and CD44hlCD8+T cells in the spleen. NK and CD8+T cells remained above the baseline measurement for up to 7 days, whereas CD44hlCD8+T cells remained above the baseline measurement for 14 days. These finding shows that compared to free S, 0.1 eq of [P-L]i-S results in superior NK cells, CD8+T cells and CD44hlCD8+T cell expansion in both PBMCs and the spleen.

[0085] Similar to Example 6, the ratio of CD8+and CD44hlCD8+T cells to NK cells shows that [P-L]i-S has a significant and long-lasting expansion of CD8+T cells compared to NK cells (Figure 15 A,B,F,G). In addition, the ratios NK, CD8+and CD44hlCD8+T cells to Tregs following treatment with [P-L]i-S indicate a massive increase of these immune cells compared to immunosuppressive Tregs that is not observed with free S. (Figure 15C-E, H-J). Finally, AUCo-2id measurements show treatment with [P-L]i-S results in a profound increase in exposure of NK, CD8+and CD44hlCD8+T cells in both PBMCs and the spleen (Table 4). Taken together, the similar duration of immune cell expansion and percentage of Ki67+NK, CD8+and CD44hlCD8+T cells in the spleen and PBMCs show that the pharmacodynamic effect of [P-L]i-S is not isolated to a single compartment and [P-L]i-S.Table 4. AUCo-2id in PBMCs and Spleena10 pg 1 pg 10 pgS [P-L]1-S [P-L]1-SBlood Spleen Blood Spleen Blood SpleenCD3 NK1.1+0.50 0.10 1.7 7.1 15 24CD8+4.6 1.4 5.3 20 28 190CD44hiCD8+1.8 3.8 3.0 16 21 150CD4+CD25+FoxP3+0.40 0 0.30 0.70 0.40b3.6Ki67+CD3+NK1.1+0.20 1.7 1.5 10 12 25Ki67+CD8+3.1 7.4 4.2 27 23 170Ki67+CD44hiCD8+1.4 1.2 2.4 11 17 88Ki67+CD4+CD25+FoxP3+0.03 0.2 0.04 0.60 0.10b3.6 a. AUCo-2id values are cellsxl0'6xday for spleen and cells / pLxlO'3xday for PBMCs b. AUCo-2id was calculated without a day 7 Tregmeasurement in PBMCsExample 8Efficacy of [P-L]i-S in combination with anti-PD-1 in CT26 tumor bearing mice

[0086] CT26 tumors were established in the flank of female Balb / c mice via SC injection of CT26 tumor cells (1 xlO5in 100 pL of serum-free medium). When tumor volumes reached ~50mm3(DO), mice treated with anti-PD-1 were administered one dose of anti-PD-1 (200|jg, 20 pL); animals were excluded from the study if tumors responded to the initial anti-PD-1 treatment. All animals were randomized on day 3. The groups were treated as follows. Group 1 was the control group and was left untreated. Group 2 was administered IP anti-PD-1 (200pg, 20 pL) on DO, D5, D7, D13, and D17 starting when tumors reach ~50 mm3. Group 3 was administered SC 0.5 mg / kg [P-L]i-S on day 4. Group 4 was given a combination of SC 0.5 mg / kg [P-L]i-S once on day 4 and IP anti-PD-1 DO, D5, D7, D13, and D17. Group 5 was given a combination of 0.5 mg / kg [P-L]i-S on day 4 and day 11 and IP anti-PD-1 DO, D5, D7, D13, and D17. Tumor volumes were measured two times a week until a humane endpoint was reached and the overall survival of animals was recorded through 60 days.

[0087] The blockade of PD-1 has proven very valuable in tumor model studies and therapeutic applications. A number reports have demonstrated that IL- 15 agonists coupled with anti-PD-1 antibodies enhance anti-tumor activity through the blockade and reactivation of exhausted CD8+ T cells. Here we show [P-L]i-S combined with anti-PD-1 leads to an additive effect in terms of anti-tumor efficacy (Figure 16). Single agent administration of [P- L]i-S did not result in measurable tumor growth inhibition or improve the median survival of animals compared to the untreated group. Administration of six doses of anti-PD-1 resulted in tumor growth inhibition but only extended the median survival of animals by 5 days compare to the untreated group. However, the combination of anti-PD-1 and [P-L]i-S, dosed once or Q7Dx2, resulted in a durable and long-lasting anti-tumor response that led to an 8- to 17 days increase in the median survival of animals compared to the mice treated with only anti-PD- 1. Moreover, all of the mice surviving to 60 days (6 / 16) had complete responses to thetreatment, with no measurable tumor volume; no complete responses were recorded in groups receiving either single agent therapy.

Claims

CLAIMSWhat is claimed is:

1. A conjugate comprising a polyethylene glycol covalently and stably bonded to a sushi domain from an IL- 15 receptor alpha subunit.

2. The conjugate of claim 1, wherein the polyethylene glycol is linear or branched, and has an average molecular weight between about 10,000 and about 60,000 Daltons.

3. The conjugate of claim 1, wherein the polyethylene glycol is bonded to the sushi domain through a linker.

4. The conjugate of claim 1, further comprising an IL- 15.

5. The conjugate of claim 4, wherein the IL- 15 is noncovalently associated with the PEGylated sushi domain.

6. The conjugate of claim 4, wherein the IL-15 is part of a fusion protein comprising the sushi domain, a peptidic linker, and an IL- 15.

7. The conjugate of any one of claims 1-6, wherein the sushi domain has a sequence with at least 85% sequence identity to SEQ ID No: 2.

8. The conjugate of any one of claims 4-6, wherein the IL- 15 has a sequence with at least 85% sequence identity to one of SEQ ID No: 5-19.

9. The conjugate of claim 5, wherein the IL-15 has a sequence with at least 85% sequence identity to one of SEQ ID No: 7-910. The conjugate of claim 6, wherein the fusion protein has a sequence with at least 85% sequence identity to one of SEQ ID No: 26-45.

11. The conjugate of claim 6, consisting of a linear polyethylene glycol having an average molecular weight of 40,000 Daltons covalently and stably bonded to the N-terminal amine of SEQ ID No: 27.

12. A method for the preparation of a conjugate of any one of claims 1-11, comprising:contacting a protein comprising a sushi domain S with a linker of formula (II) comprising a reactive group Z in the presence of a reducing agent so as to form a linkerprotein via reductive alkylation of the N-terminal amine of the sushi domain; optionally isolating the linker-protein; and contacting the linker protein with an activated polyethylglycol comprising cognate reactive group Z’ under conditions such that Z and Z’ react to form connecting group Z*.

13. The method of claim 12, wherein Z is an azide and Z’ is a cyclooctyne or bicyclononyne.

14. The method of claim 12, wherein S is a sushi domain having 85% sequence identity to one of SEQ ID No: 2.

15. The method of claim 12, wherein S is a fusion protein having 85% sequence identity to one of SEQ ID No: 26-45.

16. A method for the preparation of a conjugate of any one of claims 1-11, wherein S is a sushi domain noncovalently associated with an IL- 15, comprising the step of contacting a PEGylated sushi domain with an IL- 15 under conditions wherein the nonco valent complex is formed.

17. A method of increasing the number of CD8+T cells in a subject comprising the step of administering a conjugate of any one of claims 1-11 to the subject.

18. The method of claim 17, where in the conjugate comprises a polyethylene glycol of average molecular weight of between about 10000 and about 60000 Daltons attached to the N-terminal amine of a fusion protein having SEQ ID No: 26-45.

19. The method of claim 17, wherein the conjugate comprises a polyethylene glycol of average molecular weight of 40000 Daltons attached to the N-terminal amine of a fusion protein having SEQ ID No: 27.

20. A method of treating cancer or other condition benefitting from an expansion of CD8+T cells in a patient in need thereof comprising administration of a conjugate of any one of claims 1-11 to the patient.

21. The method of claim 20, further comprising administering an immune checkpoint inhibitor.

22. The method of claim 21, wherein the immune checkpoint inhibitor is a PD-1 inhibitor, PD-L1 inhibitor, or CTLA-4 inhibitor.

23. The method of claim 22, wherein the immune checkpoint inhibitor is selected from pembrolizumab, nivolumab, cemiplimab, atezolizumab, dostarlimab, and avelumab.

24. The method of claim 20, further comprising administering an anti-CD40 antibody, anti-OX40 antibody, anti-CD20 antibody, anti-CD52 antibody, anti-CCR4 antibody, or anti- CD38 antibody.

Images