LAG-3-regulatable il-2 receptor agonists

Abstract

Provided herein are reversibly activatable polypeptides comprising a dual-binding detection domain that facilitates reversible, cue-dependent activation and inactivation of an effector domain, such as IL-2 or IL-15. The dual-binding detection domain can be capable of binding the effector domain and an inducing moiety (such as LAG-3), but configured to only bind either the effector domain or the inducing moiety at one time, facilitating reversible activation and inactivation of the effector domain. This reversible, competitive binding can provide effector activity (for example, IL-2 or IL-15 receptor signaling) that is dependent on or is enhanced by the presence of the inducing moiety. Accordingly, reversibly activatable polypeptides disclosed herein can exhibit reduced background effector activity in the absence of or at a low concentration of inducing moiety, for example, lower IL-2 receptor signaling in the absence of LAG-3, reducing systemic IL-2 signaling and toxicity, while providing therapeutic benefits of targeted IL-2 signaling in the presence of LAG-3.

Description

LAG-3-REGULATABLE IL-2 RECEPTOR AGONISTSCROSS REFERENCE

[0001] This Application claims the benefit of United States Provisional Patent Application No. 63 / 631,329, filed April 8, 2024, which is incorporated herein by reference in its entirety.BACKGROUND

[0002] Interleukin 2 (IL-2) is a powerful growth factor for T cells and enhances their survival, expansion, and effector function. Although IL-2 was approved as a treatment for of renal cell carcinoma in 1992, its use in cancer immunotherapy has been limited by toxi cities arising from its broad systemic activity as well as efficacy in a narrow patient population. Several strategies have been pursued to address the limitations of IL-2, which include attenuating activity, limiting binding to IL-2Ra (non-alpha), and targeting to specific cells with antibody-IL-2 fusions.SUMMARY

[0003] Disclosed herein, in some aspects, is a reversibly activatable polypeptide comprising: (a) an effector domain; and (b) a dual-binding detection domain that is configured to bind the effector domain and an inducing moiety; wherein: (i) in an absence of the inducing moiety, the dual-binding detection domain binds to the effector domain and the effector domain is sterically hindered from binding to an effector partner; and (ii) upon contacting the reversibly activatable polypeptide to the inducing moiety, the dual-binding detection domain binds to the inducing moiety, the dual-binding detection domain is sterically hindered from binding to the effector domain, and the effector domain is activated; wherein the inducing moiety comprises LAG-3.

[0004] In some embodiments, the dual-binding detection domain comprises a dual-binding antibody or antigen-binding fragment thereof comprising a heavy chain complementarity determining region (HCDR)3 with the amino acid sequence of SEQ ID NO: 1663. In some embodiments, the dual-binding detection domain comprises a dual-binding antibody or antigenbinding fragment thereof comprising: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662; and (c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663. In some embodiments, the dualbinding detection domain comprises a dual -binding antibody or antigen-binding fragment thereof comprising: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662; (c) an HCDR3 with the amino acidsequence of SEQ ID NO: 1663; (d) a light chain complementarity determining region (LCDR)1 with the amino acid sequence of SEQ ID NO: 1664; (e) an LCDR2 with the amino acid sequence of SEQ ID NO: 1665 (AAS); and (f) an LCDR3 with the amino acid sequence of SEQ ID NO: 1666. In some embodiments, the effector domain is an IL-2 receptor agonist. In some embodiments, when activated, the effector domain is not sterically hindered from binding to the effector partner. In some embodiments, the effector domain comprises IL-2. In some embodiments, the effector domain comprises IL-15. In some embodiments, the effector domain comprises at least 80% sequence identity to any one of SEQ ID NOs: 1-4 and 79-91. In some embodiments, the effector domain comprises at least 95% sequence identity to any one of SEQ ID NOs: 1-4 and 79-91. In some embodiments, the effector domain comprises at least 98% sequence identity to any one of SEQ ID NOs: 1-4 and 79-91. In some embodiments, the effector domain comprises the amino acid sequence of any one of SEQ ID NOs: 1-4 and 79-91. In some embodiments, the effector partner comprises an IL-2 receptor. In some embodiments, the effector partner comprises IL-2RP. In some embodiments, the effector partner comprises IL- 2Ry. In some embodiments, the effector partner comprises IL-2RP and fL-2Ry. In some embodiments, the effector partner comprises IL-2Ra, IL-2RP and IL-2R . In some embodiments, the dual-binding detection domain binds to a domain or epitope of IL-2 that is an IL-2Rpy binding site. In some embodiments, the dual-binding detection domain does not bind to a domain or epitope of IL-2 that is an IL-2Ra binding site. In some embodiments, the effector domain comprises an IL-2 mutein with modification(s) that reduce or eliminate binding to IL- 2Ra.

[0005] In some embodiments, the dual-binding detection domain comprises a dual-binding antibody or antigen-binding fragment thereof. In some embodiments, the dual-binding detection domain comprises an HCDR3 sequence disclosed in TABLE 25. In some embodiments, the dual-binding detection domain comprises a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 25. In some embodiments, the dual-binding detection domain comprises a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 25. In some embodiments, the dual-binding detection domain comprises a heavy chain variable domain (VH) and a light chain variable domain (VL). In some embodiments, the dual-binding detection domain comprises a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 282. In some embodiments, the dual-binding detection domain comprises a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 282. In some embodiments, the dual-bindingdetection domain comprises a VH with the amino acid sequence of SEQ ID NO: 262 and a VL with the amino acid sequence of SEQ ID NO: 282. In some embodiments, the dual-binding detection domain comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLEs 22 & 23. In some embodiments, the dual -binding detection domain comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLEs 22 & 23. In some embodiments, the dual-binding detection domain comprises a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLEs 22 & 23. In some embodiments, the dual-binding detection domain comprises, consists essentially of, or consists of a Fab. In some embodiments, the dual-binding detection domain comprises, consists essentially of, or consists of an scFv. In some embodiments, the dual-binding detection domain comprises, consists essentially of, or consists of a single domain antibody. In some embodiments, the effector domain is activated when the reversibly activatable polypeptide is contacted to a LAG-3+ cell. In some embodiments, the reversibly activatable polypeptide induces IL-2R signaling, expansion, or activation of LAG-3+ IL-2RPY+or LAG-3+ IL-2RaP'y+ cells. In some embodiments, the effector domain is inactive when the reversibly activatable polypeptide is contacted to a LAG-3 -negative (LAG-3-) cell. In some embodiments, the effector domain is inactive when the reversibly activatable polypeptide is contacted to a LAG-3- IL- 2RPy+ cell. In some embodiments, the reversibly activatable polypeptide does not induce or substantially does not induce expansion or activation of LAG-3- IL-2RPy+ cells. In some embodiments, the reversibly activatable polypeptide induces at least 10-fold higher IL-2 signaling by LAG-3+ IL-2RPY+ or LAG-3+ IL-2RaPY+ cells as compared to LAG-3- IL-2RPY+ or LAG-3- IL-2RUPY+ cells. In some embodiments, the reversibly activatable polypeptide is a LAG-3 antagonist. In some embodiments, the reversibly activatable polypeptide inhibits LAG-3 activity and induces IL-2R signaling. In some embodiments, the reversibly activatable polypeptide further comprises a monospecific antibody or antigen-binding fragment thereof that binds to LAG-3. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises an HCDR3 with the amino acid sequence of SEQ ID NO: 817. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816; and (c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816; (c) an HCDR3 with the amino acid sequenceof SEQ ID NO: 817; (d) an LCDR1 with the amino acid sequence of SEQ ID NO: 818; (e) an LCDR2 with the amino acid sequence of SEQ ID NO: 819 (DVS); and (f) an LCDR3 with the amino acid sequence of SEQ ID NO: 820. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises an HCDR3 sequence disclosed in TABLE 24. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 24. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 24. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 214. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 214. In some embodiments, the monospecific antibody or antigenbinding fragment thereof comprises a VH with the amino acid sequence of SEQ ID NO: 137 and a VL with the amino acid sequence of SEQ ID NO: 214. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLES 20 & 21. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLES 20 & 21. In some embodiments, the monospecific antibody or antigen-binding fragment thereof comprises a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLEs 20 & 21. In some embodiments, the reversibly activatable polypeptide comprises an Fc domain. In some embodiments, the reversibly activatable polypeptide comprises an IgG isotype antibody. In some embodiments, the dual-binding detection domain is a Fab of the IgG isotype antibody. In some embodiments, the reversibly activatable polypeptide comprises a first heavy chain and a first light chain, wherein the first heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the effector domain, a linker, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the first light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the VL and a CL. In some embodiments, the reversibly activatable polypeptide comprises a first heavy chain and a first light chain, wherein the first heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH,CH1, hinge, CH2, CH3, and optionally CH4, and the first light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the effector domain, a linker, the VL, and a CL. In some embodiments, the reversibly activatable polypeptide further comprises a second heavy chain. In some embodiments, the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, an scFv, optionally CHI, hinge, CH2, CH3, and optionally CH4. In some embodiments, the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, (i) a single domain antibody, VHH, or nanobody, (ii) optionally CHI, (iii) hinge, (iv) CH2, (v) CH3, and (vi) optionally CH4. In some embodiments, the reversibly activatable polypeptide further comprises a second heavy chain and a second light chain. In some embodiments, the second heavy chain comprises a VH and the second light chain comprises a VL. In some embodiments, the reversibly activatable polypeptide is heterodimeric and the first heavy chain, and the second heavy chain each comprise a modification that promotes heterodimerization of the first heavy chain with the second heavy chain. In some embodiments, the VH of the second heavy chain and the VL of the second light chain form a second dual-binding detection domain that is configured to bind the effector domain and the inducing moiety. In some embodiments, the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the effector domain, a linker, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the second light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the VL and a CL. In some embodiments, the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the second light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the effector domain, a linker, the VL, and a CL. In some embodiments, the VH of the second heavy chain and the VL of the second light chain form an antigen-binding fragment, optionally wherein the antigen-binding fragment is monospecific. In some embodiments, the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH of the second heavy chain, CHI, hinge, CH2, CH3, and optionally CH4, and the second light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the VL of the second light chain and a CL. In some embodiments, the reversibly activatable polypeptide comprises a set of polypeptide chains, wherein each of the polypeptide chains comprises an amino acid sequence with at least 80% sequence identity to a polypeptide chain disclosed in TABLE 27. In some embodiments, the reversibly activatable polypeptide comprises a set of polypeptide chains, wherein each of the polypeptide chains comprises an amino acid sequence with at least 90% sequence identity to a polypeptide chain disclosed in TABLE 27. In some embodiments, the reversibly activatable polypeptide comprises a set of polypeptide chains disclosed in TABLE 26, wherein each of thepolypeptide chains comprises the amino acid sequence of a polypeptide chain disclosed in TABLE 27.

[0006] Disclosed herein, in some aspects, is a polynucleotide comprising a nucleic acid sequence that encodes the reversibly activatable polypeptide of any one of the preceding embodiments.

[0007] Disclosed herein, in some aspects, is a vector comprising the polynucleotide. In some embodiments, the vector is a non-viral vector. In some embodiments, the vector is a viral vector.

[0008] Disclosed herein, in some aspects, is a pharmaceutical composition comprising the reversibly activatable, the polynucleotide, or the vector of any one of the preceding embodiments, and a pharmaceutically-acceptable excipient (e g., vehicle, carrier, or diluent).

[0009] Disclosed herein, in some aspects, is a method of selectively inducing signaling by an effector partner, the method comprising contacting a population of cells with the reversibly activatable polypeptide of any one of the preceding embodiments, wherein signaling by the effector partner is preferentially induced in cells with surface expression of the inducing moiety.

[0010] Disclosed herein, in some aspects, is a method of selectively inducing signaling by an effector partner, the method comprising administering the reversibly activatable polypeptide, the polynucleotide, the vector, or the pharmaceutical composition of any one of the preceding embodiments to the subject.

[0011] Disclosed herein, in some aspects, is a method of treating a condition in a subject in need thereof, the method comprising administering the reversibly activatable polypeptide, the polynucleotide, the vector, or the pharmaceutical composition of any one of the preceding embodiments to the subject.

[0012] In some embodiments, the condition is a cancer. In some embodiments, the reversibly activatable polypeptide drives expansion and activation of tumor-specific T cells. In some embodiments, the reversibly activatable polypeptide inhibits tumor growth. In some embodiments, toxicity is reduced as compared to a control compound for which the effector domain is active in the absence of the inducing moiety. In some embodiments, toxicity is reduced as compared to a control compound for which the dual-binding detection domain does not bind to the effector domain.

[0013] Disclosed herein, in some aspects, is a combination therapy comprising the reversibly activatable polypeptide of any one of the preceding embodiments and an immune checkpoint inhibitor. In some embodiments, the immune checkpoint inhibitor is a PD1 inhibitor. In some embodiments, the immune checkpoint inhibitor is a PDL1 inhibitor.

[0014] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising an HCDR3 with the amino acid sequence of SEQ ID NO: 1663.

[0015] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662; and (c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663.

[0016] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662; (c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663; (d) a light chain complementarity determining region (LCDR)l with the amino acid sequence of SEQ ID NO: 1664; (e) an LCDR2 with the amino acid sequence of SEQ ID NO: 1665 (AAS); and (f) an LCDR3 with the amino acid sequence of SEQ ID NO: 1666.

[0017] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 282.

[0018] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 282.

[0019] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a VH with the amino acid sequence of SEQ ID NO: 262 and a VL with the amino acid sequence of SEQ ID NO: 282.

[0020] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising an HCDR3 sequence disclosed in TABLE 25.

[0021] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 25.

[0022] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 25.

[0023] In some embodiments, the dual-binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

[0024] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed inTABLEs 22 & 23. In some embodiments, the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

[0025] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLES 22 & 23. In some embodiments, the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

[0026] Disclosed herein, in some aspects, is a dual-binding antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLES 22 & 23. In some embodiments, the dual-binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

[0027] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof that binds to LAG-3, comprising an HCDR3 with the amino acid sequence of SEQ ID NO: 817.

[0028] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof that binds to LAG-3, comprising: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816; and (c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817.

[0029] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof that binds to LAG-3, comprising: (a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815; (b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816; (c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817; (d) an LCDR1 with the amino acid sequence of SEQ ID NO: 818; (e) an LCDR2 with the amino acid sequence of SEQ ID NO: 819 (DVS); and (f) an LCDR3 with the amino acid sequence of SEQ ID NO: 820.

[0030] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof that binds to LAG-3, comprising a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 214.

[0031] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof that binds to LAG-3, comprising a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 214.

[0032] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof that binds to LAG-3, comprising a VH with the amino acid sequence of SEQ ID NO: 137 and a VL with the amino acid sequence of SEQ ID NO: 214.

[0033] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof comprising an HCDR3 sequence disclosed in TABLE 24. In some embodiments, the antibody or antigen-binding fragment thereof binds to LAG-3.

[0034] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof comprising a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 24. In some embodiments, the antibody or antigen-binding fragment thereof binds to LAG-3.

[0035] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof comprising a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 24. In some embodiments, the antibody or antigen-binding fragment thereof binds to LAG-3.

[0036] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLEs 20 & 21. In some embodiments, the antibody or antigen-binding fragment thereof binds to LAG-3.

[0037] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLEs 20 & 21. In some embodiments, the antibody or antigen-binding fragment thereof binds to LAG-3.

[0038] Disclosed herein, in some aspects, is an antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLEs 20 & 21. In some embodiments, the antibody or antigen-binding fragment thereof binds to LAG-3.INCORPORATION BY REFERENCE

[0039] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference.BRIEF DESCRIPTION OF THE DRAWINGS

[0040] FIG. 1 shows a schematic of an illustrative reversibly activatable polypeptide in which effector domains (E) are joined via non-cleavable linkers to the N-termini of (in this case) VH regions of an antibody. The VH and VL regions of each arm of the antibody form dualbinding detection domains that have binding affinity for both the effector and an inducingmoiety. In the absence of the inducing moiety, the dual-binding detection domains can bind to the effector domains, preventing, reducing, sterically hindering, or blocking the effector domains from binding to an effector partner.

[0041] FIG. 2 shows a schematic of the illustrative reversibly activatable polypeptide of FIG. 1 in the presence of inducing moiety (I). In the presence of the inducing moiety, the dualbinding detection domains can bind to the inducing moiety, thereby activating the effector domains (e.g., preventing, reducing, sterically hindering, or blocking the effector domains from binding to the dual-binding detection domains, leaving the effector domains free to bind to the effector partner).

[0042] FIG. 3 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which effector domains are linked to the N-termini of VH domains of both arms of an immunoglobulin (e.g., IgG) scaffold.

[0043] FIG. 4 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which an effector domain is linked to the N-terminus of a VH domain of one arm of a heterodimeric immunoglobulin (e.g., IgG) scaffold.

[0044] FIG. 5 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which effector domains are linked to the N-termini of VL domains of both arms of an immunoglobulin (e.g., IgG) scaffold.

[0045] FIG. 6 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which an effector domain is linked to the N-terminus of a VL domain of one arm of a heterodimeric immunoglobulin (e.g., IgG) scaffold.

[0046] FIG. 7 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which effector domains are linked to the N-terminus of a VH domain of one arm and the VL domain of a second arm of a heterodimeric immunoglobulin (e.g., IgG) scaffold.

[0047] FIG. 8 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which a first heavy chain and a second heavy chain form a Fab, and a second heavy chain is linked to an scFv that is in turn linked to an effector domain, where the scFv is a dual-binding detection domain.

[0048] FIG. 9 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which one arm of a heterodimeric immunoglobulin (e g., IgG) scaffold comprises a Fab with an N-terminally linked effector domain (in this case to the VH), and the second arm comprises an scFv linked N-terminally to the second heavy chain’s hinge or CHI.

[0049] FIG. 10 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which first heavy and light chains form a first Fab, second heavy and light chains form a second Fab, an scFv that is the dual-binding detection domain is linked to the C-terminusof the first heavy chain, and an effector domain is linked to the C-terminus of the second heavy chain.

[0050] FIG. 11 shows an illustrative structure that can be used for a reversibly activatable polypeptide in which a first heavy chain and a second heavy chain form an Fc domain, an scFv that is the dual-binding detection domain is linked to the C-terminus of the first heavy chain, and an effector domain is linked to the C-terminus of the second heavy chain.

[0051] FIG. 12A shows an illustrative structure that can be used for a reversibly activatable polypeptide in which IL-2 is linked to the N-terminus of a dual-binding antibody (DBA) on one arm, with an scFv (e.g., monospecific anti-LAG3 scFv) on the other arm.

[0052] FIG. 12B shows an illustrative common light chain bispecific format that can be used for a reversibly activatable polypeptide, with (i) a DBA and an effector on one arm, and (ii) a monospecific anti-LAG3 Fab on the other arm.

[0053] FIG. 13 illustrates binding of illustrative polypeptides to human LAG3 on mammalian cells. Mean fluorescence intensity (MFI) is plotted in relation to antibody concentration for four constructs disclosed herein, anti-LAG3 monospecific antibodies (AB004531, AB004524, relatlimab), and an isotype control.

[0054] FIG. 14 shows binding of illustrative polypeptides to LAG3 -expressing primary human T cells. Mean fluorescence intensity (MFI) for anti-LAG3 antibodies (with a fluorescently conjugated secondary detection antibody) is plotted in relation to anti-LAG3 antibody concentration.

[0055] FIG. 15 shows binding of illustrative polypeptides to LAG3 -expressing primary human T cells. A heterodimeric Fc fusion format was used, with the anti-LAG3 as an scFv on one arm. Mean fluorescence intensity (MFI) for anti-LAG3 antibodies (with a fluorescently conjugated secondary detection antibody) is plotted in relation to anti-LAG3 antibody concentration.

[0056] FIG. 16 shows binding of illustrative polypeptides to LAG3 -expressing primary human T cells. A common light chain format was used. Mean fluorescence intensity (MFI) for anti-LAG3 antibodies (with a fluorescently conjugated secondary detection antibody) is plotted in relation to anti-LAG3 antibody concentration.

[0057] FIGs. 17A-17F show LAG-3 dependent IL-2 activity of illustrative reversibly activatable polypeptides disclosed herein, as shown by reporter cell assays with IL-2 reporter cells that lack LAG3 or express LAG3. The reversibly activatable polypeptides were serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0058] FIG. 17G shows IL-2 activity of a control construct that lacks a dual-binding detection domain (e.g., isotype-matched to the constructs tested in FIGs. 17A-F, with a linkedIL-2, but lacking IL-2 binding specificity, thus IL-2 is constitutively active). The control construct was serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0059] FIG. 17H shows an illustrative structure that can be used for reversibly activatable polypeptides (including, e.g., for the constructs used to generate FIGs. 17A-F). The design utilizes heterodimeric antibodies with the DBA linked to an IL-2 variant on one arm and an scFv based on favezelimab on the other arm.

[0060] FIGs. 18A-18G show LAG-3 dependent IL-2 activity of illustrative reversibly activatable polypeptides disclosed herein, as shown by reporter cell assays with IL-2 reporter cells that lack LAG3 or express LAG3. The reversibly activatable polypeptides were serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0061] FIG. 18H shows IL-2 activity of a control construct that lacks a dual-binding detection domain (e.g., isotype-matched to the constructs tested in FIGs. 18A-G, with a linked IL-2, but lacking IL-2 binding specificity, thus IL-2 is constitutively active). The control construct was serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0062] FIG. 181 shows an illustrative structure that can be used for reversibly activatable polypeptides (including, e.g., for the constructs used to generate FIGs. 18A-G). The design utilizes, heterodimeric antibodies with the DBA linked to an IL-2 variant on one arm and an scFv based on relatlimab on the other arm.

[0063] FIGs. 19A-19D show LAG-3 dependent IL-2 activity of illustrative reversibly activatable polypeptides disclosed herein, as shown by reporter cell assays with IL-2 reporter cells that lack LAG3 or express LAG3. The reversibly activatable polypeptides were serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0064] FIG. 19E shows IL-2 activity of a control construct that lacks a dual-binding detection domain (e.g., isotype-matched to the constructs tested in FIGs. 19A-D, with a linked IL-2, and lacking an anti-IL-2 binding specificity, thus constitutively active). The control construct was serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0065] FIG. 19F shows an illustrative common light chain structure that can be used for reversibly activatable polypeptides (including, e.g., for the constructs used to generate FIGs. 19A-D). The design utilizes a heterodimeric antibody with the DBA linked to an IL-2 variant on one arm, and a monospecific (e.g., LAG3 -specific) Fab on the other arm, with a common light chain on both arms.

[0066] FIG. 20 shows induction of STAT5 signaling by a reversibly activatable polypeptide (cLAG3-IL2) and a non-conditional IL2-containing isotype control (e.g., lacking an anti-IL-2 binding specificity, thus constitutively active). Primary human CD8+ T cells were incubated with the constructs, with or without pre-treatment with an anti-LAG3 antibody to block the dualbinding detection domain from binding LAG3.

[0067] FIG. 21 provides schematics of illustrative reversibly activatable polypeptides (top panels) and a control construct with the same anti-LAG3 targeting scFv arm but with an isotype control Fab arm that does not bind IL-2.

[0068] FIG. 22A shows average tumor size over time for MC38 tumor-bearing hPD- l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), and an isotype control.

[0069] FIG. 22B shows tumor size of individual animals from each group of the experiment described for FIG. 22A.

[0070] FIG. 23A shows body weight over time for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control. Body weight loss is indicative of toxicity.

[0071] FIG. 23B shows the concentration of blood CD8+ cells for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0072] FIG. 23C shows the concentration of blood NK cells for MC38 tumor-bearing hPD- l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0073] FIG. 23D shows the concentration of blood Tregs for MC38 tumor-bearing hPD- l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0074] FIG. 24 shows the percent change in lung weight for MC38 tumor-bearing hPD- l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451) or constitutively active LAG3-targeted IL-2 (AF008522), relative to an isotype control.

[0075] FIG. 25A shows the intratumoral density of CD8+ T cells and stem-like CD8+ T cells for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3- IL2 constmcts (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0076] FIG. 25B shows the frequency of Ki-67+ cells (e g., indicative of division) and intratumoral GrzB+ cells (e.g., indicative of cytolytic / effector function) of intratumoral CD8+ T cells for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3- IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0077] FIG. 25C shows the intratumoral ratio of CD8+ T cells to Tregs for MC38 tumorbearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0078] FIG. 25D shows the composition of the intratumoral CD8+ T cell population for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control. From top to bottom, the subpopulations shown by the bars are naive, effector, central memory, and effector memory.

[0079] FIG. 25E shows the frequency of splenic (left) and intratumoral (right) M8 tetramerpositive cells for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0080] FIG. 26A shows frequency of intratumoral IFN-Y+ TNF-a+ cells amongst CD8+ T cells for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3- IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control.

[0081] FIG. 26B shows the proportions of intratumoral CD8+ T cells producing 0, 1, 2, or all 3 of the effectors IFN-y, TNF-a, and GrzB for MC38 tumor-bearing hPD-l / hLAG-3 mice treated with reversibly activatable cLAG3-IL2 constructs (AF008825 and AF008451), a constitutively active LAG3-targeted IL-2 (AF008522), or an isotype control. The fraction with no percentages shown is the fraction with zero effectors detected.

[0082] FIG. 27 shows the frequency of intratumoral PD-1 and LAG-3 double-positive cells amongst CD8+ T cells after treatment with a reversibly activatable polypeptide (AF008825).

[0083] FIG. 28A shows average tumor volumes over time for groups of MC38 tumorbearing hPD-l / hLAG-3 mice treated with (i) a reversibly activatable polypeptide (AF008809) alone, (ii) the reversibly activatable polypeptide in combination with an anti -PD-1 antibody, (iii) a combination of an anti-LAG3 antibody and an anti -PD-1 antibody, (iv) a non-conditional LAG-3 -targeted IL-2 (AF008522), or an isotype control.

[0084] FIG. 28B shows tumor volumes over time plotted for individual MC38 tumorbearing hPD-l / hLAG-3 mice treated with (i) a reversibly activatable polypeptide (AF008809) alone, (ii) the reversibly activatable polypeptide in combination with an anti-PD-1 antibody, (iii) a combination of an anti-LAG3 antibody and an anti-PD-1 antibody, (iv) a non-conditional LAG-3 -targeted IL-2 (AF008522), or an isotype control.

[0085] FIG. 29A shows % change in body weight in non-tumor bearing hPD-l / hLAG3 mice following treatment with cLAG3-IL2 constructs (30 mg / kg; AF009302 & AF008809), an isotype control antibody (30 mg / kg), or the maximum tolerated dose of a non-conditional LAG3-targeted IL-2 (2.5 mg / kg; AF008522).

[0086] FIG. 29B shows % change in lung weight in non-tumor bearing hPD-l / hLAG3 mice following treatment with cLAG3-IL2 constructs (30 mg / kg; AF009302 & AF008809), or the maximum tolerated dose of a non-conditional LAG3-targeted IL-2 (2.5 mg / kg; AF008522), relative to an isotype control antibody (30 mg / kg).

[0087] FIG. 29C shows concentration of blood NK cells and CD8+ T cells in non-tumor bearing hPD-l / hLAG3 mice following treatment with cLAG3-IL2 constructs (30 mg / kg; AF009302 & AF008809), an isotype control antibody (30 mg / kg), or the maximum tolerated dose of a non-conditional LAG3-targeted IL-2 (2.5 mg / kg; AF008522).

[0088] FIG. 30A shows IFN-y release by human PBMCs after treatment with reversibly activatable polypeptides (cLAG3-IL2) or a control non-conditional IL-2 construct.

[0089] FIG. 30B shows IFN-y release by human PBMCs of individual donors after treatment for 24 hours with reversibly activatable polypeptides (cLAG3-IL2) or a control non- conditional IL-2 construct.

[0090] FIG. 31 shows serum concentrations of reversibly activatable polypeptides (cLAG3- IL2), non-conditional LAG3-targeted IL-2, and an PSMA-directed isotype control in circulation in C57BL / 6 mice after intravenous dosing.

[0091] FIGs. 32A-32D show LAG-3 dependent IL-2 activity of illustrative reversibly activatable polypeptides disclosed herein, as shown by reporter cell assays with IL-2 reporter cells that lack LAG3 expression (open circles) or express LAG3 (filled squares), for constructs utilizing the DBA kp2_A03 (FIG. 32A) or with sequence modifications to the DBA designed to improve construct stability (FIGs. 32B-D). The reversibly activatable polypeptides were serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0092] FIG. 33A-33D show LAG-3 dependent IL-2 activity of illustrative reversibly activatable polypeptides disclosed herein, as shown by reporter cell assays with IL-2 reporter cells that lack LAG3 expression (open circles) or express LAG3 (filled squares), for constructs utilizing the DBA 7w2_A02 (FIG. 33A) or with sequence modifications to the DBA designed toimprove construct stability (FIGs. 33B-D). The reversibly activatable polypeptides were serially diluted and tested for the ability to induce IL-2R signaling at various concentrations.

[0093] FIGs. 34A-34H show LAG-3 dependent IL-2 activity, as shown by reporter cell assays with IL-2 reporter cells that lack LAG3 expression (open circles) or express LAG3 (filled squares). Activity is shown for illustrative reversibly activatable polypeptides disclosed herein (FIGs. 34A-34F), compared to “always on” controls (FIG. 34G & FIG. 34H).

[0094] FIGs. 35A-35F show binding of six illustrative reversibly activatable polypeptides (one per figure, filled circles) to LAG-3- expressing primary human T cells. Binding is compared to an anti-LAG3 monoclonal antibody (relatlimab, empty squares), or an isotype control antibody (filled diamonds).

[0095] FIG. 36 shows average tumor size over time for hLAG-3 mice implanted with MC38 syngeneic tumor cells. When tumors reached a volume of -125 mm3on day 8 post-implantation, mice were treated intravenously (IV) with cLAG3-IL2 (0.5 mg / kg), cLAG3-IL2 (0.5 mg / kg) in combination with anti-mouse PD-1 (clone 29F.1A12, Bio X Cell; 10 mg / kg), anti-human LAG-3 (10 mg / kg) in combination anti-mouse PD-1 (10 mg / kg), an isotype control antibody (PSMA; 10 mg / kg), or a non-conditionally active non-targeted IL-2 (1 mg / kg). Mice received a second dose of treatments on day 11 post-implantation.DETAILED DESCRIPTION

[0096] T cells are capable of specifically recognizing tumor-derived antigens and are key mediators of anti-tumor immunity. Modulation of T cell function has proven to be a powerful axis in cancer therapy. Indeed, treatment with blocking antibodies directed against the T cell- expressed inhibitory receptors PD-1 and CTLA-4 leads to a reinvigorated anti -turn or immune response. Interleukin 2 (IL-2) is a powerful growth factor for T cells and enhances their survival, expansion, and effector function. Although IL-2 was approved as a treatment for of renal cell carcinoma in 1992, its use in cancer immunotherapy has been limited by toxicities arising from its broad systemic activity as well as efficacy in a narrow patient population. Several strategies have been pursued to address the limitations of IL-2, which include attenuating activity, limiting binding to IL-2Ra (non-alpha), and targeting to specific cells with antibody-IL-2 fusions.

[0097] Multiple groups have developed IL-2 variants with limited binding to IL-2Ra with the goal of directing IL-2 activity away from regulatory T (Treg) cells (see, e.g., Raeber et al. A systematic review of interleukin-2 -based immunotherapies in clinical trials for cancer and autoimmune diseases. eBioMedicine 90, 104539 (2023)). Of these, bempegaldesleukin has progressed the furthest in clinical development. Bempegaldesleukin consists of human IL-2 with six pegylated surface lysines, which are designed to extend its half-life and reduce binding to IL-2Ra. In clinical trials, bempegaldesleukin showed no activity as a monotherapy, and in Phase III trials in combination with nivolumab, it showed no improvement over nivolumab alone. In addition, bempegaldesleukin showed toxicities that were similar to aldesleukin, precluding its administration at higher doses. Clinical development of bempegaldesleukin and two other nonalpha IL-2 variants (NL-201 and THOR-707) has been discontinued due to a lack of efficacy at tolerable doses. Nemvaleukin, which is still under development, consists of a circularly permuted IL-2 fused to the extracellular portion of IL-2Ra. This design sterically occludes IL-2 from binding to the high-affinity IL-2R. Nemvaleukin has shown activity as a monotherapy and in combination with PD-1 blockade in early trials, and a Phase III trial was initiated in 2022.

[0098] Upon activation, T cells upregulate inhibitory receptors to downregulate overexuberant immune responses. These checkpoints, such as PD-1 and LAG-3, can be important for preventing autoimmune disease, however they can also limit T cell function in inappropriate tolerizing settings such as viral infection and cancer.

[0099] Blockade of the PD-1 / PD-L1 pathway with monoclonal antibodies has been approved for the treatment of a wide variety of malignancies. Unfortunately, only a minority of patients show durable responses, and some tumor types are largely refractory. In an effort to overcome mechanisms of resistance to PD-1 / PD-L1 blockade, many therapeutics are being tested in clinical trials in combination with anti-PDl / PD-Ll. Relatlimab, a blocking anti-LAG-3 antibody, was recently approved as a treatment for melanoma in combination with nivolumab. Relatlimab treatment alone did not show clinical activity, however in combination with nivolumab, it significantly improved responses compared to nivolumab alone.

[0100] Some groups have developed approaches that target IL-2 or IL-15 (which also acts on the IL-2RPy receptor) to cells expressing immune checkpoint proteins, e.g., via fusion to an anti -PD-1 or anti-LAG-3 antibody. The goal of these targeting approaches is to preferentially direct cytokine activity to antigen-experienced T cells relative to other IL-2-responsive cell types. Approaches to PD-1 targeting programs include ASKG015 and ASKG812 (AskGene Pharma), PDl-IL2v (Roche), IB 1363 (Innovent Biologies, NCT05460767), IAP0971 (SunHo BioPharmaceutical), SAR445877 (Sanofi), Mableukin-2PD1 (Anwita Bioscience), BPT331 (Bright Peak Therapeutics), PF-07209960 (Pfizer), and SOT201 (Sotio Biotech). LAG-3- targeted programs include aLAG-3-IL-2c (Anwita Bioscience), TNRX257 (Tentarix Biotherapeutics), and LAG-3 x IL15 (Xencor). Salubris Biotherapeutics is testing a CTLA-4- targeted IL-15 (JK08). AB248 is a CD8-targeted IL-2 in development by Asher Bio.

[0101] To improve targeting to the tumor microenvironment, strategies have been developed that utilize protease-activated versions of IL-2, such as WTX-124 (Werewolf Therapeutics) and XTX202 (Xilio Therapeutics), and a PD-1 antibody -targeted, protease-activated IL-2 (XilioTherapeutics). However such mechanisms of conditional activity can have limitations including, for example, irreversible activation, and off-target activation. Additionally, tumor targeting can reduce the effectiveness of a therapeutic, since tumor-specific T cells outside the tumor microenvironment can be activated by systemic treatment and migrate into the tumor.

[0102] Although therapeutic strategies based on IL-2R agonism and / or immune checkpoint (e.g., LAG-3, PD-1) targeting / inhibition have shown promise, many existing strategies only achieve efficacy in limited patient populations, lack durable therapeutic benefit, and are associated with toxi cities or narrow therapeutic windows to achieve efficacy with low or acceptable toxicity.

[0103] Disclosed herein are reversibly activatable polypeptides that can provide context- dependent therapeutic activity, for example, IL-2 receptor signaling that is dependent on the presence of LAG-3. Such polypeptides can exhibit conditional IL-2 activity based on the presence of LAG-3 (and in some ceases, they are referred to as cLAG3-IL2).I. REVERSIBLY ACTIVAT BLE POLYPEPTIDE S

[0104] Provided herein are reversibly activatable polypeptides comprising a dual-binding detection domain that facilitates reversible, cue-dependent activation and inactivation of an effector domain, such as IL-2 or IL-15. The dual -binding detection domain can be capable of binding the effector domain and an inducing moiety, but configured to only bind either the effector domain or the inducing moiety at one time, facilitating reversible activation and inactivation of the effector domain. For example, a reversibly activatable polypeptide can be in an “off’ state by default in the absence of inducing moiety, with the detection domain is bound to the effector domain, thereby blocking the effector domain from binding to its effector partner and rendering it inactive (FIG. 1). For example, an inactivated cytokine can be blocked from binding its receptor and inducing receptor signaling.

[0105] When the reversibly activatable polypeptide is contacted to an inducing moiety, the detection domain can bind to the inducing moiety instead of the effector moiety, resulting in a conformation of the reversibly activatable polypeptide in which the effector domain is activated and capable of binding the effector partner (FIG. 2). This reversible, competitive binding can provide effector activity (for example, IL-2 or IL-15 receptor signaling) that is dependent on or is enhanced by the presence of the inducing moiety, or is enhanced at a higher relative concentration of the inducing moiety. Accordingly, reversibly activatable polypeptides disclosed herein can exhibit reduced background effector activity in the absence of or at a low concentration of inducing moiety, for example, lower IL-2 receptor signaling in the absence of LAG-3, reducing systemic IL-2 signaling and toxicity.

[0106] Binding of an activated effector domain to an effector partner can elicit a therapeutic effect. For example, binding of IL-2 or IL- 15 to an IL-2 receptor or IL- 15 receptor can result in cytokine receptor signaling in a target cell that expresses the cytokine receptor (and, for example, also expresses, comprises at the cell surface, or is in close association with, the inducing moiety, such as LAG-3). Accordingly, a reversibly activatable polypeptide can induce IL-2R signaling in LAG-3+ IL-2RPy+ or LAG-3+ IL-2RaPy+ cells.

[0107] Binding of an activated effector domain to an effector partner can induce a response, for example, STAT (e.g., STAT5) phosphorylation, cellular (e.g., immune cell) proliferation, cellular (e.g., immune cell) survival or persistence, immune activation, cytokine (e.g., pro- inflammatory cytokine) production, chemokine production, chemotaxis, cytolytic activity, killing of cancer cells, killing of infected cells, antiviral activity, reduced exhaustion, enhanced immune effector functions, enhanced anti-cancer immune response, cellular differentiation (e.g., memory and / or effector differentiation), and the like.

[0108] In some embodiments, a reversibly activatable polypeptide can act as a full agonist on the effector partner that is a receptor, for example, can result in a high level of signaling that is comparable to a control molecule in which the effector domain is constitutively active or irreversibly activated (e.g., via cleavage). The full agonist effect can apply, for example, at a medium or high concentration of the inducing moiety (e.g., with selectivity based on the presence and abundance of the inducing moiety).

[0109] Reduced activation or no activation of the effector domain (and, e.g., reduced or no binding to the effector partner) can occur when the reversibly activatable polypeptide is contacted to a LAG-3- or LAG-3 low cell, for example, despite expression of subunits of an IL- 2 or IL- 15 receptor. For example, in some embodiments a reversibly activatable polypeptide does not induce or substantially does not induce expansion or activation of LAG-3- IL-2RPy+, LAG-3- IL-2RaPy+, LAG-3 low IL-2RPy+, or LAG-3 low IL-2RaPy+ cells.

[0110] Reversibly activatable polypeptides can exert effects via an effector domain binding to an effector partner, for example, on a target cell. In some embodiments, the target cell expresses the effector partner, for example, receptor. As disclosed herein, the effector domain can be IL-2 or IL-15, and accordingly the target cell can express a combination of appropriate receptor subunits, for example, IL-2RPy, IL-2RaPy, or IL-15Ra and IL-2RPy.

[0111] In some embodiments, the target cell expresses the inducing moiety or comprises the inducing moiety, e.g., on the cell surface. For example, a target cell can express LAG-3. In some embodiments, the target cell expresses the effector partner and expresses the inducing moiety or comprises the inducing moiety, e.g., on the cell surface. For example, a target cell can be LAG- 3+ (or high) and IL-2RPy+ (or high), LAG-3 + (or high) and IL-2RaPy+ (or high), or LAG-3 +(or high), IL-15Ra+ (or high), and fL-2Rpy+ (or high). The reversibly activatable polypeptide can bind to the inducing moiety (via the detection domain) and the effector partner (via the effector domain), e.g., inducing activity of the effector partner in cis.

[0112] Non-limiting examples of target cells can include activated T cells, LAG-3+ T cells, LAG-3 high T cells, antigen-experienced T cells, effector T cells, effector / memory T cells, memory T cells, central memory T cells, resident memory T cells, CD8+ T cells, activated CD8+ T cells, antigen-experienced CD8+ T cells, CD8+ effector T cells, CD8+ effector / memory T cells, CD8+ memory T cells, CD8+ central memory T cells, CD8+ resident memory T cells, CD4+ T cells (e.g., TH1, TH2, TH9, or TH17), activated CD4+ T cells, antigen-experienced CD4+ T cells, CD4+ effector T cells, CD4+ effector / memory T cells, CD4+ memory T cells, CD4+ central memory T cells, CD4+ resident memory T cells, gamma delta T cells (e.g., LAG-3+ or LAG-3 high), NKT cells (e.g., LAG-3+ or LAG-3 high), invariant natural killer T (iNKT) cells (e.g., LAG-3+ or LAG-3 high), lymphocytes (e.g., LAG-3+ or LAG-3 high), tumor-infiltrating lymphocytes (e g , LAG-3+ or LAG-3 high), plasmacytoid dendritic cells (pDCs; e.g., LAG-3+ or LAG-3 high), activated B cells (e.g., LAG-3+ or LAG-3 high), mucosal-associated invariant T (MAIT) cells, NK cells (e.g. activated NK cells), monocytes, macrophages, and dendritic cells.

[0113] Control, non-target cells can lack the inducing moiety, for example, lack expression the inducing moiety and / or lack the inducing moiety at the cell surface. Control, non-target cells can have a low level of the inducing moiety, for example, a low expression level of the inducing moiety and / or a low level of the inducing moiety at the cell surface. Control, non-target cells can comprise the effector partner, but lack the inducing moiety or have a low level of the inducing moiety.

[0114] In some embodiments, the target cell is adjacent to a cell that expresses the inducing moiety, e.g., in a target condition or niche, such as a tumor microenvironment. In some embodiments, the reversibly activatable polypeptide can bind to the inducing moiety on an adjacent cell (via the detection domain) and the effector partner on the target cell (via the effector domain), e.g., inducing activity of the effector partner in trans. A control, non-target condition or niche can be condition or niche in which both the target cell and adjacent cells lack or express low levels of the inducing moiety, for example, systemic locations distal from the tumor microenvironment.

[0115] The lack of or low level of the inducing moiety can result in no activity or only a low level of activity induced by the reversibly activatable polypeptide regardless of the presence of the effector partner, for example, STAT (e.g., STAT5) phosphorylation, cellular (e.g., immune cell) proliferation, cellular (e.g., immune cell) survival or persistence, immune activation,cytokine (e.g., pro-inflammatory cytokine) production, chemokine production, chemotaxis, cytolytic activity, killing of cancer cells, killing of infected cells, antiviral activity, reduced exhaustion, enhanced immune effector functions, enhanced anti-cancer immune response, or cellular differentiation (e.g., memory and / or effector differentiation).

[0116] Non-limiting examples of control / non-target cells can include naive T cells (e.g., LAG3- or LAG3 low), T cells in the absence of TCR-specific cognate antigen presentation (e.g., LAG3- or LAG3 low), NK cells (e.g., activated NK cells and / or LAG3- or LAG3 low), B lymphocytes (e.g., LAG3- or LAG3 low), macrophages (e.g., LAG3- or LAG3 low), dendritic cells (DCs; e.g., LAG3- or LAG3 low), and monocytes (e g., LAG3- or LAG3 low).

[0117] A reversibly activatable polypeptide can exhibit higher induction of a response mediated by the effector and effector partner in the presence of the inducing moiety than in the absence of the inducing moiety (e.g., LAG-3), or at a high concentration of the inducing moiety versus a low concentration of the inducing moiety. In some embodiments, a reversibly activatable polypeptide exhibits at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3 -fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold higher induction of a response mediated by the effector or effector partner in the presence of the inducing moiety than in the absence of the inducing moiety, or at a high concentration of the inducing moiety versus a low concentration of the inducing moiety. In some embodiments, the higher induction of a response can be as demonstrated by blocking binding of the reversibly activatable polypeptide to the inducing moiety, for example, with an antibody that bocks binding of the dual-binding detection domain to LAG-3.

[0118] In some embodiments, a reversibly activatable polypeptide exhibits at most about 10- fold, at most about 20-fold, at most about 50-fold, at most about 100-fold, at most about 500- fold, at most about 1000-fold, at most about 5000-fold, or at most about 10,000-fold higher induction of a response mediated by the effector or effector partner in the presence of the inducing moiety than in the absence of the inducing moiety, or at a high concentration of the inducing moiety versus a low concentration of the inducing moiety.

[0119] In some embodiments, a reversibly activatable polypeptide exhibits about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about 20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500-fold, about 750- fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-fold higher induction of a response mediated by the effector or effector partner in the presence of theinducing moiety than in the absence of the inducing moiety, or at a high concentration of the inducing moiety versus a low concentration of the inducing moiety.

[0120] In some embodiments, a reversibly activatable polypeptide exhibits about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2-fold to about 1,000-fold, about 2- fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2- fold to about 50-fold, about 5-fold to about 50,000-fold, about 5-fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000-fold, about 50-fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 50,000-fold, about 100- fold to about 10,000-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000- fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500-fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher induction of a response mediated by the effector or effector partner in the presence of the inducing moiety than in the absence of the inducing moiety, or at a high concentration of the inducing moiety versus a low concentration of the inducing moiety.

[0121] An illustrative reversibly activatable polypeptide comprises an IL-2R agonist as an effector domain and utilizes LAG-3 as an inducing moiety, facilitating selective, preferential, or enhanced signaling on LAG-3+ or LAG-3 high cells, for example, as determined by a reporter cell assay.

[0122] In some embodiments, a reversibly activatable polypeptide induces at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold higher IL-2 signaling by LAG-3+ fL-2RPy+ or LAG-3 high IL-2RPy+ cells as compared to LAG-3- IL-2RPy+ cells or LAG-3 low IL-2RPy+ cells.

[0123] In some embodiments, a reversibly activatable polypeptide induces at most about 10- fold, at most about 20-fold, at most about 50-fold, at most about 100-fold, at most about 500- fold, at most about 1000-fold, at most about 5000-fold, or at most about 10,000-fold higher IL-2 signaling by LAG-3+ IL-2RPy+ or LAG-3 high IL-2R.P + cells as compared to LAG-3- IL- 2RPy+ cells or LAG-3 low IL-2RPy+ cells.

[0124] In some embodiments, a reversibly activatable polypeptide induces about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500-fold, about 750- fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-fold higher IL-2 signaling by LAG-3+ IL-2RPy+ or LAG-3 high fL-2RPy+ cells as compared to LAG-3- IL-2Rpy+ cells or LAG-3 low ZL-2Rpy+ cells.

[0125] In some embodiments, a reversibly activatable polypeptide induces about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2-fold to about 1,000-fold, about 2- fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2- fold to about 50-fold, about 5-fold to about 50,000-fold, about 5-fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000-fold, about 50-fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 50,000-fold, about 100- fold to about 10,000-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000- fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500-fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher IL-2 signaling by LAG-3+ IL-2RPy+ or LAG-3 high IL- 2RPy+ cells as compared to LAG-3- IL-2Rpy+ cells or LAG-3 low IL-2RPy+ cells.

[0126] In some embodiments, a reversibly activatable polypeptide induces at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold higher IL-2 signaling by LAG-3+ IL-2ROIPY+ or LAG-3 high TL-2RaPy+ cells as compared to LAG-3- IL-2RaP,y+ cells or LAG-3 low IL-2RaP'y+ cells.

[0127] In some embodiments, a reversibly activatable polypeptide induces at most about 10- fold, at most about 20-fold, at most about 50-fold, at most about 100-fold, at most about 500- fold, at most about 1000-fold, at most about 5000-fold, or at most about 10,000-fold higher IL-2 signaling by LAG-3+ fL-2RaPy+ or LAG-3 high IL-2RaPy+ cells as compared to LAG-3 - IL- 2Ra.Py+ cells or LAG-3 low fL-2Ra.Py+ cells.

[0128] In some embodiments, a reversibly activatable polypeptide induces about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about 20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500-fold, about 750- fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-foldhigher IL-2 signaling by LAG-3+ IL-2RaPy+ or LAG-3 high IL-2RotPy+ cells as compared to LAG-3- IL-2Rapy I cells or LAG-3 low IL-2R«Py l cells.

[0129] In some embodiments, a reversibly activatable polypeptide induces about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2-fold to about 1,000-fold, about 2- fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2- fold to about 50-fold, about 5-fold to about 50,000-fold, about 5-fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000-fold, about 50-fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 50,000-fold, about 100- fold to about 10,000-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000- fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500-fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher IL-2 signaling by LAG-3+ IL-2RaPy+ or LAG-3 high IL- 2RaPy+ cells as compared to LAG-3- IL-2RaPy+ cells or LAG-3 low IL-2RaPy+ cells.

[0130] In some embodiments, a reversibly activatable polypeptide induces at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold higher IL-15 signaling by LAG-3+ IL-2RPy+ or LAG-3 high IL-2RPy+ cells as compared to LAG-3- IL-2RPy+ cells or LAG-3 low IL-2RPy+ cells.

[0131] In some embodiments, a reversibly activatable polypeptide induces at most about 10- fold, at most about 20-fold, at most about 50-fold, at most about 100-fold, at most about 500- fold, at most about 1000-fold, at most about 5000-fold, or at most about 10,000-fold higher IL- 15 signaling by LAG-3+ IL-2RPy+ or LAG-3 high IL-2RPy+ cells as compared to LAG-3- IL- 2RPy+ cells or LAG-3 low IL-2RPy+ cells.

[0132] In some embodiments, a reversibly activatable polypeptide induces about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about 20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500-fold, about 750- fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-fold higher IL- 15 signaling by LAG-3 + ZL-2RPy+ or LAG-3 high IL-2RPy+ cells as compared to LAG-3- IL-2RPy+ cells or LAG-3 low IL-2RPy+ cells.

[0133] In some embodiments, a reversibly activatable polypeptide induces about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2-fold to about 1,000-fold, about 2- fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2- fold to about 50-fold, about 5-fold to about 50,000-fold, about 5-fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000-fold, about 50-fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 50,000-fold, about 100- fold to about 10,000-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000- fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500-fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher IL-15 signaling by LAG-3+ ZL-2RPy+ or LAG-3 high IL- 2RPy+ cells as compared to LAG-3- IL-2Rpy+ cells or LAG-3 low IL-2RP'y+ cells.

[0134] In some embodiments, a reversibly activatable polypeptide inhibits LAG-3 activity or signaling, and induces IL-2R or IL-15R signaling.

[0135] A reversibly activatable polypeptide disclosed herein can comprise a serum half-life of at least 12 hours, at least 24 hours, at least 36 hours, at least 48 hours, at least 72 hours, at least 96 hours, at least 120 hours, at least 168 hours, at least 250 hours, at least 320 hours, or at least 400 hours. The serum half-life can be a human serum half-life, a murine serum half-life, a porcine serum-half life, a bovine serum half-life, a canine serum half-life, a feline serum halflife, or a leporine serum half-life.A. Effector domain

[0136] Reversibly activatable polypeptides disclosed herein comprise an effector domain that can be reversibly activated and inactivated to selectively modulate an effector partner (e.g., receptor) based on the presence or absence of an inducing moiety. An activated effector domain can be capable of binding to an effector partner, for example, and an inactivated effector domain can be sterically hindered (e.g., blocked or obstructed) from binding to the effector partner.

[0137] An effector domain can be or comprise a receptor ligand, and the effector partner can comprise a receptor. An effector domain can be or comprise a cytokine, and the effector partner can comprise a cytokine receptor, for example a class I, heterodimeric, or heterotrimeric cytokine receptor. An effector domain can be or comprise a common gamma chain family cytokine, and the effector partner can be a cytokine receptor that comprises common gammachain. An effector domain can be or comprise a mammalian cytokine, and the effector partner can be a mammalian cytokine receptor. An effector domain can be or comprise a human cytokine, and the effector partner can be a human cytokine receptor. When activated, the effector domain can be capable of binding to and / or inducing signaling by the receptor. When inactivated, the effector domain can exhibit reduced binding to the receptor, for example, can be sterically hindered or blocked from binding to the receptor.

[0138] An effector domain can be or comprise an IL-2 receptor (IL-2R) agonist, for example, IL-2, IL-15, or a peptide agonist of an IL-2R. An effector domain can be or comprise IL-2. In some embodiments, an effector domain is or comprises a wild type IL-2 sequence. Illustrative human wild type IL-2 sequences are provided in SEQ ID NO: 1 (without signal peptide) and SEQ ID NO: 2 (with signal peptide).

[0139] In some embodiments, an effector domain is or comprises a mutein IL-2 sequence. Substitutions present in a mutein can be described relative to a given reference sequence, for example, SEQ ID NO: 1 or 2. In some embodiments, an effector domain comprises an IL-2 mutein or variant comprising the amino acid sequence of SEQ ID NO: 1 except for one or more modifications (e.g., substitution(s), insertion(s), or deletion(s)) relative to SEQ ID NO: 1. In some embodiments, an effector domain comprises an IL-2 mutein or variant comprising an amino acid sequence with at least 80%, at least 85%, at least 90%, at least 95%, at least 97%, or at least 98% sequence identity or similarity to SEQ ID NO: 1 with one or more modifications (e.g., substitution(s), insertion(s), or deletion(s)) relative to SEQ ID NO: 1.

[0140] An IL-2 mutein or variant used in a reversibly activatable polypeptide disclosed herein can comprise modification(s) that alter binding affinity for one or more IL-2 receptor subunits. Native IL-2 can bind to trimeric and dimeric IL-2 receptor (IL-2R). Trimeric IL-2R comprises IL-2Ra (CD25), IL-2RP (CD122), and the common gamma chain (fL-2Ry, CD132), and can also be referred to as high affinity IL-2R or IL-2RaP'y. Dimeric IL-2R comprises IL-2RP and ZL-2Ry, and can also be referred to as ZL-2RPy. ZL-2RaPy can be present at high levels or predominantly expressed on immunosuppressive or immune tolerance-promoting regulatory T (Treg) cells, and accordingly in some embodiments an IL-2 comprises one or more modifications to reduce affinity for IL-2Rot and / or fL-2RaPy, e.g., to reduce IL-2 signaling in Tregs, or enhance affinity for IL-2Ra and / or ZL-2RaPy, e.g., to enhance IL-2 signaling in Tregs. IL-2R.py can be present at high levels or predominantly expressed on effector T cells, memory T cells, and NK cells, and accordingly in some embodiments an IL-2 comprises one or more modifications to increase affinity for ZL-2RPy to increase IL-2 signaling in effector T cells, memory T cells, or NK cells, or to reduce affinity for IL-2RPy to reduce IL-2 signaling in effector T cells, memory T cells, or NK cells.

[0141] In some embodiments, modification(s) that reduce or eliminate binding to IL-2Ra can be advantageous for specificity of the reversibly activatable polypeptide for desired target cells. For example, a dual -binding detection domain can block binding of IL-2 to IL-2RPy by binding to a site or domain of IL-2 that binds to and induces signaling by IL-2RPy, but in some embodiments the dual -binding detection domain does not block binding of IL-2 to IL-2Ra. In such a case, modification(s) that reduce or eliminate binding to IL-2Ra can reduce, eliminate, or prevent binding of the reversibly activatable polypeptide to cells that express the IL-2Ra subunit, thereby increasing specificity of binding of the reversibly activatable polypeptide for cells with the inducing moiety.

[0142] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that reduce binding affinity for IL-2Ra (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0143] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that reduce binding affinity for IL-2RP (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0144] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that reduce binding affinity for fL-2Ry (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0145] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that reduce binding affinity for fL-2RaPy (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0146] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that reduce binding affinity for IL-2RPy (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0147] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that increase binding affinity for IL-2Ra (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0148] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that increase binding affinity for IL-2RP (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0149] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that increase binding affinity for IL-2RY (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0150] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that increase binding affinity for IL-2RaPy (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about500-fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type humanIL-2.

[0151] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that increase binding affinity for IL-2RPy (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, or at least about 10,000-fold, e.g., relative to wild type human IL- 2.

[0152] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that facilitate selectivity for IL-2RPy over IL-2Ra.Py (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20- fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, or at least about 10,000-fold higher affinity for IL-2R.pY as compared to fL-2RaPy.

[0153] In some embodiments, an effector domain comprises an IL-2 comprising one or more modifications that facilitate selectivity for IL-2RaPy over IL-2RPY (e.g., when the effector domain is in the active state), for example, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20- fold, at least about 25-fold, at least about 30-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, or at least about 10,000-fold higher affinity for IL-2RaPy as compared to IL-2RPy.

[0154] The one or more modifications can comprise, for example, modifications of residues T3, R38, K43, E61, E62, and / or C125 relative to SEQ ID NO: 1, or any combination thereof. In some embodiments, the modifications comprise a T3A, R38A, R38D, R38E, K43A, K43E, K43G, E61R, E62G, E62S, or C125A substitution, or a combination thereof. In some embodiments, the modifications comprise modifications (e.g., substitutions) of residues T3, R38, K43, E61, and C125. In some embodiments, the modifications comprise T3A, R38D, K43E, E61R, and C125A substitutions (e.g., as shown in SEQ ID NO: 80). In some embodiments, the modifications comprise a T3A, R38D, K43E, E62S, or C125A substitution, or a combination thereof. In some embodiments, the modifications comprise modifications (e.g., substitutions) of residues T3, R38, K43, E62, and C125. In some embodiments, the modifications comprise T3A, R38D, K43E, E62S, and C125A substitutions (e.g., as shown inSEQ ID NO: 79). Additional illustrative IL-2 muteins or variants are provided in SEQ ID NOs: 81-91.

[0155] The one or more modifications can comprise, for example, modifications of residues T3, L18, D20, Q22, 128, N29, Y31, K35, T37, R38, F42, K43, Y45, K48, T51, E61, E62, V69, N71, L72, Q74, L80, R81, L85, 186, N88, 189, V91, 192, M104, C125, Q126, S130 relative to SEQ ID NO: 1, or any combination thereof. In some embodiments, the modifications comprise a T3A, L18R, D20A, D20T, D20H, Q22E, I28T, N29S, Y31H, K35R, T37A, R38A, R38D, R38E, F42A, F42K, K43A, K43E, K43G, Y45A, Y45R, K48E, E61R, E62G, E62S, V69A, N71R, L72G, Q74P, L80F, R81D, L85V, I86V, N88D, N88R, N88G, I89V, V91D, V91K, I92F, M104V, C125A, C125S, Q126L, Q126F, Q126H, Q126T, S130R, substitution relative to SEQ ID NO: 1, or a combination thereof.

[0156] In some embodiments, an effector domain comprises an IL-2 comprising a C125S substitution. In some embodiments, an effector domain comprises an IL-2 comprising an N88R and / or V91D substitution, e.g., to reduce affinity for IL-2R0. In some embodiments, an effector domain comprises an IL-2 comprising an N88D substitution, e g., to reduce affinity for IL-2Rp. In some embodiments, an effector domain comprises an IL-2 comprising L80F, R81D, L85V, I86V, and I92F substitutions, e.g., to enhance affinity to IL-2Rp. In some embodiments, an effector domain comprises an IL-2 comprising L80F, R81D, L85V, I86V, and I92F substitutions, e.g., to enhance ZL-2R0 affinity, and / or L18R, Q22E, Q126T, and S130R substitutions, e.g., to reduce affinity for gamma chain. In some embodiments, an effector domain comprises an IL-2 comprising L18R, Q22E, and Q126H substitutions, e.g., to reduce affinity for gamma chain. In some embodiments, an effector domain comprises an IL-2 comprising V91K and D20A substitution, e.g. to reduce IL-2RP affinity, and an M104V mutation, e.g., to reduce IL-2Ra affinity. In some embodiments, an effector domain comprises an IL-2 comprising F42A, L80F, R81D, L85V, I86V, and I92F substitutions. In some embodiments, an effector domain comprises an IL-2 comprising an N88D substitution to reduce affinity to IL-2RP, and / or V69A and Q74P mutations to enhance affinity to IL-Ra. In some embodiments, an effector domain comprises an IL-2 comprising a T3 (e.g., T3A) substitution. In some embodiments, an effector domain comprises an IL-2 comprising an R38 (e.g., R38D) substitution. In some embodiments, an effector domain comprises an IL-2 comprising a K43 (e.g., K43E) substitution. In some embodiments, an effector domain comprises an IL-2 comprising an E61 (e.g., E61R) substitution. In some embodiments, an effector domain comprises an IL-2 comprising an E62 (e.g., E62S) substitution. In some embodiments, an effector domain comprises an IL-2 comprising a C125 (e.g., C125A) substitution.

[0157] An effector domain can be or comprise an IL-15 receptor agonist, for example, IL- 15, or a peptide agonist of IL-15R. In some embodiments, an effector domain is or comprises a wild type IL-15 sequence. In some embodiments, an effector domain is or comprises a mutein IL-15 sequence. Illustrative human wild type IL-15 sequences are provided in SEQ ID NO: 3 (without signal peptide) and SEQ ID NO: 4 (with signal peptide).

[0158] In some embodiments, an effector domain is or comprises a mutein IL- 15 sequence. Substitutions present in a mutein can be described relative to a given reference sequence, for example, SEQ ID NO: 3 or 4. In some embodiments, an effector domain comprises an IL-15 mutein comprising the amino acid sequence of SEQ ID NO: 3 except for one or more modifications (e.g., substitution(s), insertion(s), or deletion(s)) relative to SEQ ID NO: 3. In some embodiments, an effector domain comprises an IL-15 mutein comprising an amino acid sequence with at least 80%, at least 85%, at least 90%, at least 95%, at least 97%, or at least 98% sequence identity or similarity to SEQ ID NO: 3 with one or more modifications (e.g., substitution(s), insertion(s), or deletion(s)) relative to SEQ ID NO: 3.

[0159] An IL- 15 mutein used in a reversibly activatable polypeptide disclosed herein can comprise modification(s) that alter binding affinity for one or more receptor subunits. Native IL- 15 can bind to heterotrimeric IL-15R comprising IL-15Ra, IL-2RP (CD122), and the common gamma chain (fL-2Ry, CD132). IL-15 bound to IL-15Ra can also be presented in trans to cells that express fL-2RPy but minimal IL-15Ra.

[0160] In some embodiments, an effector domain comprises an IL-15 comprising one or more modifications that reduce binding affinity for IL-15Ra, IL-2R , and / or fL-2Ry (e.g., when the effector domain is in the active state), as compared to wild type human IL- 15. In some embodiments, an effector domain comprises an IL-15 comprising one or more modifications that increase binding affinity for IL-15Ra, IL-2R|3, and / or IL-2Ry (e.g., when the effector domain is in the active state), as compared to wild type human IL-15. In some embodiments, an effector domain comprises an IL- 15 comprising one or more modifications that reduce binding affinity for IL-2Rpy (e.g., when the effector domain is in the active state), as compared to wild type human IL-15. In some embodiments, an effector domain comprises an IL-15 comprising one or more modifications that increase binding affinity for IL-2RPy (e.g., when the effector domain is in the active state), as compared to wild type human IL-15.

[0161] The one or more modifications can comprise, for example, modifications of residue N72 relative to SEQ ID NO: 3. In some embodiments, an effector domain comprises an IL-15 comprising an N72D substitution.

[0162] TABLE 1 provides sequences of illustrative effector domains and reference IL-2 and IL- 15 sequences.TABLE 1

[0163] An effector partner can comprise IL-2Ra. An effector partner can comprise IL-2Rp. An effector partner can comprise IL-2Ry. An effector partner can comprise IL-2Ra, IL-2RP, and IL-2 Ry (e.g., ZL-2RaPy). An effector partner can comprise IL-2RP and fL-2Ry (e.g., IL-2RPy). An effector partner can comprise IL-15Ra. An effector partner can comprise IL-15Ra, IL-2RP, and ZL-2Ry.B. Dual-binding detection domain

[0164] Reversibly activatable polypeptides disclosed herein comprise a dual-binding detection domain that can be configured to bind an effector domain and an inducing moiety. The effector domain and the inducing moiety can compete for binding to the dual-binding detection domain. The dual-binding nature of the detection domain and the fact that it can bind only one of the effector domain and inducing moiety at a time can facilitate reversible activation of the effector domain. For example, a reversibly activatable polypeptide can be in an “off” state by default, in which the detection domain is bound to the effector domain, thereby rendering the effector domain inactive and sterically hindering the effector domain from binding to its effector partner (FIG. 1). When the reversibly activatable polypeptide encounters an inducing moiety (e.g., LAG-3), the detection domain can bind to the inducing moiety instead of the effector moiety, resulting in a conformation of the reversibly activatable polypeptide in which the effector domain is activated (FIG. 2). For example, binding of the detection domain to the inducing moiety can sterically hinder the detection domain from binding the effector domain, thereby rendering the effector domain free to bind to its effector partner.

[0165] An inducing moiety can be a molecule that is present at a niche of interest within a subject, for example, present at higher levels in the niche of interest than a control niche, or present at the niche of interest and absent in the control niche, facilitating selective activity of the effector domain in the niche. In some embodiments, an inducing moiety is present at high levels in a tumor microenvironment (TME) and at low levels systemically or outside the tumor microenvironment, facilitating selective activation of an effector domain in the TME. In some embodiments, an inducing moiety is expressed by or present at high levels on activated immune cells, such as antigen-activated T cells, but is absent or expressed at low levels on non-activated immune cells (e.g., present at high levels on the surface of activated or antigen-experienced T cells, but absent or present at lower levels on non-activated or naive T cells). In some embodiments, an inducing moiety is expressed by or present on cancer cells, or cells present in the TME. In some embodiments, an inducing moiety is expressed by or present at high levels on target cells, but not expressed or present at low levels on control, non-target cells.

[0166] An inducing moiety targeted by a reversibly activatable polypeptide can comprise or can be LAG-3. LAG-3 is an immune inhibitory receptor expressed on the surface of certain immune cells upon activation, for example, CD4+ T cells, CD8+ T cells, gamma delta T cells, natural killer T (NKT) cells, plasmacytoid dendritic cells (pDCs), and activated B cells. LAG-3 delivers inhibitory signals upon binding to ligands, such as FGL1. LAG-3 signaling negatively regulates the proliferation, activation, and effector function of immune effector cells, including CD8+ T cells, CD4+ T-cells, and pDCs.

[0167] LAG-3 and other immune inhibitory receptors and immune checkpoints have been found at high levels in cancers (e.g., in the TME), and can contribute to downregulation of immune activation that allows cancers or pathogens to persist, grow, and evade targeting or elimination by the immune system.

[0168] Individual receptors / checkpoints can have overlapping effects on the immune system, however these effects are not identical. For example, knockout mutations in the gene for either LAG-3 or PD1 alone can have relatively minimal impact on steady-state immune cell function, while a combined knockout can result in diminished self-tolerance characterized by autoimmune infiltrates in multiple organs and eventual lethality, demonstrating an additive effect of targeting both receptors. PD-1 and LAG-3 are expressed on similar but not identical cell populations. For example, PD-1 can be found on activated NK cells, B lymphocytes, macrophages, dendritic cells (DCs), and monocytes, while LAG-3 can be expressed by gamma delta T cells, NKT cells, pDCs, and activated B cells. Moreover, while both PD-1 and LAG-3 can be induced on CD4+ and CD8+ T cells upon activation, they are expressed in overlapping but non-identical subpopulations of activated T cells with distinct functional characteristics (see,e.g., Grosso, et al. Functionally Distinct LAG-3 and PD-1 Subsets on Activated and Chronically Stimulated CD8 T Cellsl. J. Immunol. 182, 6659-6669 (2009); Maruhashi et al., LAG-3: from molecular functions to clinical applications. J. Immunother. Cancer 8, e001014 (2020);Matsuzaki et al. Tumor-infiltrating NY-ESO-l-specific CD8+ T cells are negatively regulated by LAG-3 and PD-1 in human ovarian cancer. Proc. Natl. Acad. Sci. 107, 7875-7880 (2010); and Woo et al. Immune inhibitory molecules LAG-3 and PD-1 synergistically regulate T cell function to promote tumoral immune escape. Cancer Res. 72, 917-927 (2012); each of which is incorporated herein by reference).

[0169] In some embodiments, a reversibly activatable polypeptide disclosed herein that utilizes LAG-3 as an inducing moiety has advantageous effects as compared to a control polypeptide that uses another immune checkpoint / immune inhibitory receptor in a comparable fashion, for example, PD1 or PDL1. Illustrative, non-limiting examples of bases of the advantageous effects can include expression of the immune inhibitory receptor by different cell populations or subpopulations, differential expression levels or persistence of expression of the receptors (e.g., in particular patients or cancer types), and the ability to combine the reversibly activatable polypeptide with a compatible immune checkpoint inhibitor, e.g., as part of the same molecule or as a separate agent administered to a subject. For example, in some embodiments a reversibly activatable polypeptide disclosed herein can be combined with a PD1 or PDL1 inhibitor, resulting in improved efficacy.

[0170] Agents that inhibit LAG-3 signaling have also been explored clinically as candidate immune checkpoint inhibitors. In some embodiments, a dual-binding detection domain is antagonistic to LAG-3 signaling when bound to LAG-3, such that a reversibly activatable polypeptide can elicit a combination of LAG-3 inhibition and effector partner (e.g., IL-2R) activation.

[0171] An inducing moiety disclosed herein can be an immunoglobulin super family member (e g., LAG-3 is an Ig superfamily member comprising 4 extracellular Ig-like domains). An inducing moiety disclosed herein can be a LAG-3 family member. For example, an inducing moiety can be a mammalian LAG-3, such as human LAG-3. The LAG-3 can be a wild type or native LAG-3. In some embodiments the inducing moiety (e.g., LAG-3) comprises an allelic variant, isoform, or, e.g., splice variant. An inducing moiety can be a cell surface protein, receptor, receptor ligand, or can be expressed, present, or presented on a cell or tissue surface. In some embodiments, an inducing moiety is secreted.

[0172] A dual-binding detection domain can exist in different states. For example, when an inducing moiety and an effector domain are each present, the detection domain: (i) can be in a first state with the inducing moiety is bound to the detection domain and the effector domain isantagonized, blocked, or sterically hindered from binding to the detection domain, (ii) can be in a second state in which the effector domain is bound to the detection domain and the inducing moiety is antagonized, blocked, or sterically hindered from binding to the detection domain, or (iii) can be in a third state in which neither the effector domain or inducing moiety is bound to the detection domain. In some embodiments, the detection domain can fluctuate between the first and second state, or between the first, second, and third states, and the (e.g., average) time spent in each state can depend on certain factors. In some embodiments, where a plurality of dual-binding detection domains (or reversibly activatable polypeptides comprising the dualbinding detection domains) are present, a percentage of the detection domains are present in each of the first and second states, or in each of the first, second, and third states, and the (e.g., average) percentage of detection domains present in each state can depend on certain factors.

[0173] Non-limiting examples of factors that can influence the time spent in each state and / or the percentage of detection domains in each state can include the absolute and relative affinity of the effector domain and inducing moiety for the detection domain, the on rates or off rates for binding of the effector domain or inducing moiety for the detection domain, abundance or concentration of the inducing moiety, and physical or chemical factors influencing accessibility of the effector domain and / or the inducing moiety to the detection domain’s binding site, such as, e g., length and flexibility of a linker that joins the effector domain to the rest of the reversibly activatable polypeptide, and physiochemical properties of amino acids at and near the binding site (e.g., affecting charge, physical bulk, and solvent accessibility). Each of these factors that influence propensity for and prevalence of the different states can be modified to obtain a reversibly activatable polypeptide with desirable properties, for example, that exhibits low background induction of signaling by an effector partner when the inducing moiety is absent or present at a lower level, and a desirable level of signaling when the inducing moiety is present or present at a higher level.

[0174] The absolute and / or relative binding affinities of a dual-binding detection domain for the effector domain and / or inducing moiety can be configured to achieve a desired sensitivity, for example, threshold level of inducing moiety at which the effector domain is activated.

[0175] For example, a dual -binding detection domain can have a higher affinity for an inducing moiety (e.g., LAG-3) than for the effector domain (e.g., cytokine such as IL-2). In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3- fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 50-fold, at least about 100-fold, at least about 200-fold at least about 500-fold, at least about 750-fold, at least about 1000-fold, at least about 2500-fold, at leastabout 5000-fold, at least about 10,000-fold, or at least about 50,000-fold higher than the affinity of the dual-binding detection domain for the effector domain.

[0176] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is at most about 10%, at most about 25%, at most about 50%, at most about 2- fold, at most about 3-fold, at most about 5-fold, at most about 10-fold, at most about 15-fold, at most about 20-fold, at most about 25-fold, at most about 50-fold, at most about 100-fold, at most about 200-fold at most about 500-fold, at most about 750-fold, at most about 1000-fold, at most about 2500-fold, at most about 5000-fold, at most about 10,000-fold, or at most about 50,000- fold higher than the affinity of the dual-binding detection domain for the effector domain.

[0177] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about 20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500-fold, about 750-fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-fold higher than the affinity of the dual -binding detection domain for the effector domain.

[0178] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2- fold to about 1,000-fold, about 2-fold to about 500-fold, about 2-fold to about 250-fold, about 2- fold to about 100-fold, about 2-fold to about 50-fold, about 5-fold to about 50,000-fold, about 5- fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000- fold, about 50-fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 50,000-fold, about 100-fold to about 10,000-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000-fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500- fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher than the affinity of the dual-binding detection domain for the effector domain.

[0179] In some embodiments, a dual-binding detection domain can have a lower affinity for an inducing moiety (e.g., LAG-3) than for the effector domain (e.g., cytokine such as IL-2). In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3- fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, atleast about 100-fold, at least about 500-fold, at least about 1000-fold, or at least about 5000-fold lower than the affinity of the dual-binding detection domain for the effector domain.

[0180] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is at most about 10%, at most about 25%, at most about 50%, at most about 2- fold, at most about 3 -fold, at most about 5-fold, at most about 10-fold, at most about 20-fold, at most about 50-fold, at most about 100-fold, at most about 500-fold, at most about 1000-fold, or at most about 5000-fold lower than the affinity of the dual-binding detection domain for the effector domain.

[0181] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 20-fold, about 50-fold, about 100-fold, about 500-fold, about 1000-fold, or about 5000-fold lower than the affinity of the dual-binding detection domain for the effector domain.

[0182] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2-fold to about 50-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000-fold, or about 100-fold to about 500-fold lower than the affinity of the dual-binding detection domain for the effector domain.

[0183] In some embodiments, the affinity of a dual-binding detection domain for an inducing moiety is similar to the affinity of the dual-binding detection domain for the effector domain, for example, within about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 20-fold, about 50-fold, about 100-fold.

[0184] The apparent affinity, absolute affinity, and / or relative affinity of an effector domain for the effector partner can be modulated by the presence versus absence of the inducing moiety. For example, the effector domain of a reversibly activatable polypeptide can bind the effector partner with at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 50-fold, at least about 100-fold, at least about 200-fold at least about 500-fold, at least about 750-fold, at least about 1000-fold, at least about 2500-fold, at least about 5000-fold, at least about 10,000-fold, or at least about 50,000-foldhigher affinity (e.g., apparent, absolute, or relative) in the presence of the inducing moiety than in the absence of the inducing moiety.

[0185] In some embodiments, the effector domain of a reversibly activatable polypeptide binds the effector partner with at most about 10%, at most about 25%, at most about 50%, at most about 2-fold, at most about 3-fold, at most about 5-fold, at most about 10-fold, at most about 15-fold, at most about 20-fold, at most about 25-fold, at most about 50-fold, at most about 100-fold, at most about 200-fold at most about 500-fold, at most about 750-fold, at most about 1000-fold, at most about 2500-fold, at most about 5000-fold, at most about 10,000-fold, or at most about 50,000-fold higher affinity (e.g., apparent, absolute, or relative) in the presence of the inducing moiety than in the absence of the inducing moiety.

[0186] In some embodiments, the effector domain of a reversibly activatable polypeptide binds the effector partner with about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about 20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500-fold, about 750-fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-fold higher affinity (e.g., apparent, absolute, or relative) in the presence of the inducing moiety than in the absence of the inducing moiety.

[0187] In some embodiments, the effector domain of a reversibly activatable polypeptide binds the effector partner with about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2-fold to about 1,000-fold, about 2-fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2-fold to about 50-fold, about 5-fold to about 50,000-fold, about 5-fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000- fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100- fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000-fold, about 50-fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100- fold to about 50,000-fold, about 100-fold to about 10,000-fold, about 100-fold to about 5,000- fold, about 100-fold to about 1,000-fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500-fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher affinity (e g., apparent, absolute, or relative) in the presence of the inducing moiety than in the absence of the inducing moiety.

[0188] The absolute and / or relative binding affinities of an effector domain for the dualbinding detection domain and / or effector partner can be configured to achieve a desired potency.For example, where the effector domain is a cytokine (e.g., IL-2), affinity of the dual-binding detection domain for the cytokine can be selected, and / or cytokine muteins with altered affinity for cytokine receptor(s) or receptor subunit(s) can be selected, to alter the amount of signaling induced by cytokine receptors (e.g., IL-2RaPy or IL-2RPy) in the presence of a given amount of inducing moiety.

[0189] For example, a dual-binding detection domain can bind the effector domain at a lower affinity than the effector domain binds to its effector partner (e.g., IL-2Ra, IL-2RP, IL- 2Ry, IL-2RaPy, IL-2RPy, IL-15Ra, or IL-15Ra / IL-2RPy). In some embodiments, the affinity of the effector domain for the effector partner is at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10- fold, at least about 15-fold, at least about 20-fold, at least about 25-fold, at least about 50-fold, at least about 100-fold, at least about 200-fold at least about 500-fold, at least about 750-fold, at least about 1000-fold, at least about 2500-fold, at least about 5000-fold, at least about 10,000- fold, or at least about 50,000-fold higher than the affinity of the effector domain for the dualbinding detection domain, e.g., in the presence of a given amount of inducing moiety.

[0190] In some embodiments, the affinity of the effector domain for the effector partner is at most about 10%, at most about 25%, at most about 50%, at most about 2-fold, at most about 3- fold, at most about 5-fold, at most about 10-fold, at most about 15-fold, at most about 20-fold, at most about 25-fold, at most about 50-fold, at most about 100-fold, at most about 200-fold at most about 500-fold, at most about 750-fold, at most about 1000-fold, at most about 2500-fold, at most about 5000-fold, at most about 10,000-fold, or at most about 50,000-fold higher than the affinity of the effector domain for the dual -binding detection domain, e.g., in the presence of a given amount of inducing moiety.

[0191] In some embodiments, the affinity of the effector domain for the effector partner is about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 15-fold, about 20-fold, about 25-fold, about 50-fold, about 100-fold, about 200-fold about 500- fold, about 750-fold, about 1000-fold, about 2500-fold, about 5000-fold, about 10,000-fold, or about 50,000-fold higher than the affinity of the effector domain for the dual-binding detection domain, e.g., in the presence of a given amount of inducing moiety.

[0192] In some embodiments, the affinity of the effector domain for the effector partner is about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 50,000-fold, about 2-fold to about 10,000-fold, about 2-fold to about 5,000-fold, about 2-fold to about 1,000- fold, about 2-fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100- fold, about 2-fold to about 50-fold, about 5-fold to about 50,000-fold, about 5-fold to about 10,000-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold toabout 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50-fold to about 50,000-fold, about 50-fold to about 10,000-fold, about 50- fold to about 5,000-fold, about 50-fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 50,000-fold, about 100-fold to about 10,000-fold, about 100-fold to about 5,000-fold, about 100- fold to about 1,000-fold, about 100-fold to about 500-fold, about 100-fold to about 250-fold, about 500-fold to about 50,000-fold, about 500-fold to about 10,000-fold, about 500-fold to about 5,000-fold, or about 500-fold to about 1,000-fold higher than the affinity of the effector domain for the dual-binding detection domain, e.g., in the presence of a given amount of inducing moiety.

[0193] In some embodiments, an effector domain can have a lower affinity for the effector partner than for the detection domain, e.g., in the absence of the inducing moiety or at a given (e.g., low) concentration of inducing moiety. In some embodiments, the affinity of the effector domain for the effector partner is at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3 -fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold lower than the affinity of the effector domain for the dual-binding detection domain, e.g., in the absence of the inducing moiety or at a given (e.g., low) concentration of inducing moiety.

[0194] In some embodiments, the affinity of the effector domain for the effector partner is at most about 10%, at most about 25%, at most about 50%, at most about 2-fold, at most about 3- fold, at most about 5-fold, at most about 10-fold, at most about 20-fold, at most about 50-fold, at most about 100-fold, at most about 500-fold, at most 1000-fold, at most 5000-fold, or at most 10,000-fold lower than the affinity of the effector domain for the dual -binding detection domain, e.g., in the absence of the inducing moiety or at a given (e.g., low) concentration of inducing moiety.

[0195] In some embodiments, the affinity of the effector domain for the effector partner is about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10-fold, about 20-fold, about 50-fold, about 100-fold, about 500-fold, about 1000-fold, about 5000-fold, or about 10,000-fold lower than the affinity of the effector domain for the dual-binding detection domain, e.g., in the absence of the inducing moiety or at a given (e.g., low) concentration of inducing moiety.

[0196] In some embodiments, the affinity of the effector domain for the effector partner is about 10% to about 100-fold, about 10% to about 10-fold, about 2-fold to about 500-fold, about 2-fold to about 250-fold, about 2-fold to about 100-fold, about 2-fold to about 50-fold, about 5-fold to about 5,000-fold, about 5-fold to about 1,000-fold, about 5-fold to about 500-fold, about 5-fold to about 250-fold, about 5-fold to about 100-fold, about 5-fold to about 50-fold, about 50- fold to about 1,000-fold, about 50-fold to about 500-fold, about 50-fold to about 250-fold, about 50-fold to about 100-fold, about 100-fold to about 5,000-fold, about 100-fold to about 1,000- fold, or about 100-fold to about 500-fold lower than the affinity of the effector domain for the dual-binding detection domain, e.g., in the absence of the inducing moiety or at a given (e.g., low) concentration of inducing moiety.

[0197] In some embodiments, the affinity of the effector domain for the effector partner is similar to the affinity of the effector domain for the dual-binding detection domain, for example, within about 10%, about 25%, about 50%, about 2-fold, about 3-fold, about 5-fold, about 10- fold, about 20-fold, about 50-fold, about 100-fold.

[0198] In some embodiments, binding of the dual-binding detection domain to the effector domain physically blocks or obstructs binding of the effector domain or a functional fragment thereof to the effector partner (e.g., physically blocks or obstructs a receptor-binding part of a cytokine from binding to the corresponding cytokine receptor or a subunit thereof).

[0199] In some embodiments, when the dual-binding detection domain binds the effector domain, the effector domain occupies the binding site of the dual-binding detection domain (for example, occupies the antigen-binding site or paratope of a dual-binding antibody or antigenbinding fragment thereof.

[0200] In some embodiments, the inducing moiety acts as a competitive antagonist of the effector domain for binding to the detection domain. In some embodiments, the inducing moiety does not act as a competitive antagonist of the effector domain.

[0201] In some embodiments, the inducing moiety acts as a non-competitive antagonist of the effector domain for binding to the detection domain. In some embodiments, the inducing moiety does not act as a non-competitive antagonist of the effector domain.

[0202] In some embodiments, the inducing moiety acts as an allosteric antagonist of the effector domain for binding to the detection domain. In some embodiments, the inducing moiety does not act as an allosteric antagonist of the effector domain.

[0203] In some embodiments, binding of the dual-binding detection domain to the inducing moiety physically blocks or obstructs binding of dual-binding detection domain to the effector domain or a functional fragment thereof (e.g., physically blocks or obstructs binding of the detection domain to a receptor-binding part of a cytokine, such as IL-2).

[0204] In some embodiments, when the dual-binding detection domain binds the inducing moiety, the inducing moiety occupies the binding site of the dual-binding detection domain (forexample, occupies the antigen-binding site or paratope of a dual-binding antibody or antigenbinding fragment thereof.

[0205] In some embodiments, the effector domain acts as a competitive antagonist of the inducing moiety for binding to the detection domain. In some embodiments, the effector domain does not act as a competitive antagonist of the inducing moiety for binding to the detection domain.

[0206] In some embodiments, the effector domain acts as a non-competitive antagonist of the inducing moiety for binding to the detection domain. In some embodiments, the effector domain does not act as a non-competitive antagonist of the inducing moiety for binding to the detection domain.

[0207] In some embodiments, the effector domain acts as an allosteric antagonist of the inducing moiety for binding to the detection domain. In some embodiments, the effector domain does not act as an allosteric antagonist of the inducing moiety for binding to the detection domain.

[0208] In some embodiments, binding of the dual-binding detection domain to the effector domain reduces binding of the effector domain or a functional fragment thereof to the effector partner via steric effects.

[0209] In some embodiments, binding of the dual-binding detection domain to the effector domain reduces binding of the dual-binding detection domain to the inducing moiety via steric effects.

[0210] In some embodiments, binding of the dual-binding detection domain to the inducing moiety reduces binding of the dual-binding detection domain to the effector domain or a functional fragment thereof via steric effects.

[0211] In some embodiments, binding of the dual-binding detection domain to the inducing moiety increases binding of the effector domain or a functional fragment thereof to the effector partner via steric effects.

[0212] A dual-binding detection domain can bind to a certain site, domain, or epitope of the effector domain. For example, where the effector domain is IL-2 or an IL-2 receptor agonist, the dual-binding detection domain can be configured such that binding of the dual-binding detection domain to the IL-2 or an IL-2 receptor agonist blocks or reduces binding to particular IL-2 receptor subunits.

[0213] For example, In some embodiments, a dual -binding detection domain binds to a domain or epitope of IL-2 that is an fL-2RPy binding site. In some embodiments, a dual-binding detection domain binds to a domain or epitope of IL-2 that is an IL-2RP binding site. In some embodiments, a dual -binding detection domain binds to a domain or epitope of IL-2 that is anIL-2Ry binding site. In some embodiments, a dual-binding detection domain binds to a domain or epitope of IL-2 that is an IL-2Ra binding site. In some embodiments, a dual-binding detection domain does not bind to a domain or epitope of IL-2 that is an IL-2Ra binding site. In some embodiments, a dual -binding detection domain does not bind to a domain or epitope of IL-2 that is an IL-2RPy binding site. In some embodiments, a dual-binding detection domain does not bind to a domain or epitope of IL-2 that is an IL-2RP binding site. In some embodiments, a dualbinding detection domain does not bind to a domain or epitope of IL-2 that is an fL-2Ry binding site.

[0214] In some embodiments, binding of the dual-binding detection domain to an effector domain that is IL-2 or an IL-2 receptor agonist blocks or reduces binding of the effector domain to TL-2RPy. In some embodiments, binding of the dual-binding detection domain to an effector domain that is IL-2 or an IL-2 receptor agonist blocks or reduces binding of the effector domain to IL-2RPy, but does not block or reduce binding to IL-2Ra. In some embodiments, binding of the dual-binding detection domain to an effector domain that is IL-2 or an IL-2 receptor agonist blocks or reduces binding of the effector domain to IL-2Rp. In some embodiments, binding of the dual-binding detection domain to an effector domain that is IL-2 or an IL-2 receptor agonist blocks or reduces binding of the effector domain to fL-2Ry. In some embodiments, binding of the dual-binding detection domain to an effector domain that is IL-2 or an IL-2 receptor agonist blocks or reduces binding of the effector domain to IL-2Ra. In some embodiments, binding of the dual-binding detection domain to an effector domain that is IL-2 or an IL-2 receptor agonist blocks or reduces binding of the effector domain to fL-2RotPy.

[0215] A dual-binding detection domain disclosed herein can be a dual-binding protein that is capable of binding the effector domain and the inducing moiety.

[0216] A dual-binding detection domain disclosed herein can be or can comprise an antibody or an antigen-binding fragment of an antibody. The antibody or antigen-binding fragment can be engineered (e.g., by affinity maturation or other suitable processes) to have desired affinities for both the effector domain and inducing moiety. An antigen-binding fragment can comprise a portion of an antibody, for example, the antigen-binding or variable region of the intact antibody. Non-limiting examples of antibody fragments include Fab, Fab', F(ab')2, dimers and trimers of Fab conjugates, Fv, scFv, single chain Fab (scFab), single variable domain (sVD), minibodies, dia-, tria-, and tetrabodies, heavy chain only antibodies (HCAbs), VHH, nanobodies, and linear antibodies. Fab and Fab’ are antigen-binding fragments that can comprise the VH and CHI domains of the heavy chain linked to the VL and CL domains of the light chain via a disulfide bond. A F(ab’)2 can comprise two Fab or Fab’ that are joined by disulfide bonds. A Fv can comprise the VH and VL domains held together by non-covalent interactions. A scFv (single-chain variable fragment) is a fusion protein that can comprise the VH and VL domains connected by a peptide linker. Manipulation of the orientation of the VH and VL domains and the linker length can be used to create different forms of molecules that can be monomeric, dimeric (diabody), trimeric (triabody), or tetrameric (tetrabody). Minibodies are scFv-CH3 fusion proteins that assemble into bivalent dimers.

[0217] The variable (V) domain(s) of an antibody can mediate antigen binding and define the specificity of a particular antibody for an antigen. The variable domain can comprise relatively invariant sequences called framework regions, and hypervariable regions, which differ considerably in sequence among antibodies of different binding specificities. The variable domain can comprise four framework regions separated by three hypervariable regions. For an antibody that comprises a heavy chain and a light chain or heavy chain variable domain and light chain variable domain, the variable domains can fold in a manner that brings the hypervariable regions together in close proximity to create an antigen binding site. The four framework regions can largely adopt an f3 -sheet configuration, while the three hypervariable regions form loops connecting, and in some cases forming part of, the f3 -sheet structure.

[0218] Within hypervariable regions are amino acid residues that primarily determine the binding specificity of the antibody in most cases. Sequences comprising these residues are known as complementarity determining regions (CDRs). One antigen binding site of an antibody with heavy and light chains or variable domains therefrom comprises six CDRs, three in the hypervariable regions of the light chain variable domain, and three in the hypervariable regions of the heavy chain variable domain. The CDRs in the light chain are designated LI, L2, and L3, while the CDRs in the heavy chain are designated Hl, H2, and H3. CDRs can also be designated LCDR1, LCDR2, LCDR3, HCDR1, HCDR2, and HCDR3, respectively. The contribution of each CDR to antigen binding varies among antibodies, but in some embodiments heavy chain CDRs, in particular HCDR3, can contribute most to antigen-specific binding. CDRs can vary in length. For example, CDRs are often 5 to 14 residues in length, but CDRs as short as 0 residues or as long as 25 residues or longer exist.

[0219] Certain antibodies or antigen-binding domains contain less than six CDRs. For example, certain antibodies lack a light chain, and can be referred to as heavy chain only antibodies (HCAbs). HCAbs have three CDRs in a variable region referred to as VHH. A single domain antibody, or nanobody, can be generated from such a VHH region of a heavy chain only antibody.

[0220] A dual-binding detection domain of the disclosure can comprise complementarity determining regions (CDRs). For example, an antibody, antigen-binding fragment thereof, or antigen-binding domain can comprise CDRs. In some embodiments, the CDRs determine orsubstantially determine binding specificity and / or affinity for the effector domain and / or the inducing moiety. For example, the CDRs can be grafted onto a different suitable framework, or the framework region can be altered (e g., via amino acid substitutions, deletions, and / or insertions), and the antigen-binding fragment or domain can retain binding for the target, and the dual-binding detection domain remains functional despite the alterations outside of the CDRs. In some embodiments, one or more framework regions or amino acid sequences therein contribute to binding specificity and / or affinity.

[0221] CDRs in or for use in a dual-binding detection domain or reversibly activatable polypeptide can be identified by various methods, including but not limited to the Kabat method, the Chothia method, the IMGT method, the AHO method, the AbM method, the contact method, and the Paratome method. For example, CDRs can be identified from any VH and VL domains disclosed herein using the Kabat method, the Chothia method, the IMGT method, the AHO method, the AbM method, the contact method, and / or the Paratome method.

[0222] Single domain antibodies can have longer CDR Hl and H3 loops compared with the respective classical CDRs, and can require different methods to identify CDRs. Single domain antibody CDRs can be identified, for example, using the single domain antibody database (SAbDab), based on common sequence elements, or based on a sequence alignment to the Chothia numbering scheme (e.g., as described by Wilton, et al. (2018). sdAb-DB: the single domain antibody database. ACS Synthetic Biology 2018 7 (11), 2480-2484 DOI: 10.1021 / acssynbio.8b00407).

[0223] A subset of residues within CDRs contacts an antigen. These residues that contact antigen can be referred to as specificity-determining residues (SDRs). However, in some embodiments residues other than SDRs can contribute to binding activity by helping to maintain the conformation of the binding site. The number of SDRs in an antibody can vary based on the size and type of antigen that is recognized, for example, between 0-14 SDRs can be found within a CDR. SDRs can be enriched in some residues, such as tyrosine, serine, tryptophan, and asparagine.

[0224] A CDR of a sequence herein can be, for example, between 0 and 91 residues in length, between 0 and 25 residues in length, between 5 and 14 residues in length, about 0 residues in length, about 1 residue in length, about 2 residues in length, about 3 residues in length, about 4 residues in length, about 5 residues in length, about 6 residues in length, about 7 residues in length, about 8 residues in length, about 9 residues in length, about 10 residues in length, about 11 residues in length, about 12 residues in length, about 13 residues in length, about 14 residues in length, about 15 residues in length, about 16 residues in length, about 17 residues in length, about 18 residues in length, about 19 residues in length, about 20 residues inlength, about 21 residues in length, about 22 residues in length, about 23 residues in length, about 24 residues in length, or about 25 residues in length.

[0225] A binding domain, such as a dual-binding detection domain or monospecific (e.g., LAG-3-specific) domain, can comprise the CDR amino acid sequence(s) of any one or more of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises an HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3 amino acid sequence selected from the CDR sequences disclosed herein. In some embodiments, a binding domain comprises an HCDR1, HCDR2, and HCDR3 amino acid sequence selected from the CDR sequences disclosed herein.

[0226] A dual-binding detection domain can comprise one or more CDR sequences each with at most one amino acid substitution, insertion, or deletion relative to any one of the CDR sequences disclosed herein. In some embodiments, a dual-binding detection domain comprises one or more CDR sequences each with at most two amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a dual-binding detection domain comprises one or more CDR sequences each with at most three amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0227] In some embodiments, a dual-binding detection domain comprises one or more CDR sequences with at least one amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0228] In some embodiments, a dual-binding detection domain comprises one or more CDR sequences each with 0-1 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a dual-binding detection domain comprises one or more CDR sequences each with 0-2 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a dual-binding detection domain comprises one or more CDR sequences each with 0-3 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a dual-binding detection domain comprises one or more CDR sequences each with 1-2 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a dualbinding detection domain comprises one or more CDR sequences each with 1-3 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0229] In some embodiments, a dual-binding detection domain comprises an HCDR3 amino acid sequence disclosed herein, and comprises HCDR1, HCDR2, LCDR1, LCDR2, and LCDR3sequences each with at most three amino acid substitutions, insertions, and / or deletions relative to sequences disclosed herein.

[0230] A binding domain (e.g., dual-binding detection domain or monospecific (e.g., LAG- 3-specific) domain) can comprise one or more variant CDR sequences, e.g., each with at most one amino acid substitution, insertion, or deletion relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with at most two amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with at most three amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with at most four amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with at most five amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0231] In some embodiments, a binding domain comprises one or more variant CDR sequences with at least one amino acid substitution, insertion, and / or deletion relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences with at least two amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences with at least three amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. The binding domain can further comprise additional CDRs with no substitutions relative to CDR sequences disclosed herein, for example, one CDR or two variant CDRs can be mutated, and the remaining CDRs can remain unchanged.

[0232] In some embodiments, a binding domain comprises one or more CDR sequences each with 0-1 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more CDR sequences each with 0-2 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more CDR sequences each with 0-3 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0233] In some embodiments, a binding domain comprises one or more variant CDR sequences each with 1-2 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises oneor more variant CDR sequences each with 1-3 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with 1-4 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with 1-5 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0234] In some embodiments, a binding domain comprises one or more variant CDR sequences each with 2-3 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with 2-4 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with 2-5 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0235] In some embodiments, a binding domain comprises one or more variant CDR sequences each with 3-4 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein. In some embodiments, a binding domain comprises one or more variant CDR sequences each with 3-5 amino acid substitutions, insertions, and / or deletions relative to any one of the CDR sequences disclosed herein.

[0236] In some embodiments, a binding domain comprises an HCDR3 amino acid sequence disclosed herein, and comprises HCDR1, HCDR2, LCDR1, LCDR2, and LCDR3 (or, e.g., HCDR1 and HCDR2) sequences each with at most one amino acid substitutions, insertions, and / or deletions relative to sequences disclosed herein. In some embodiments, a binding domain comprises an HCDR3 amino acid sequence disclosed herein, and comprises HCDR1, HCDR2, LCDR1, LCDR2, and LCDR3 (or, e.g., HCDR1 and HCDR2) sequences each with at most two amino acid substitutions, insertions, and / or deletions relative to sequences disclosed herein. In some embodiments, a binding domain comprises an HCDR3 amino acid sequence disclosed herein, and comprises HCDR1, HCDR2, LCDR1, LCDR2, and LCDR3 (or, e.g., HCDR1 and HCDR2) sequences each with at most three amino acid substitutions, insertions, and / or deletions relative to sequences disclosed herein. In some embodiments, a binding domain comprises an HCDR3 amino acid sequence disclosed herein, and comprises HCDR1, HCDR2, LCDR1, LCDR2, and LCDR3 (or, e.g., HCDR1 and HCDR2) sequences each with at most four amino acid substitutions, insertions, and / or deletions relative to sequences disclosed herein.

[0237] The dual-binding detection domain can be or can comprise a single domain antibody. The single domain antibody can be or can comprise a variable region of a heavy chain only antibody. Such a single domain antibody can also be known as a nanobody or VHH. The single domain antibody can be, for example, a variable region from or derived from a heavy chain only antibody from a camelid (e.g., camels: one-humped Camelus dromedaries and two-humped Camelus bactrianus; llamas: Lama glama, Lama guanicoe, and Lama vicugna; and alpacas: Vicugna pacos), a shark (e.g., a nurse shark), a wobbegong, or a spotted ratfish. Such animals have a special type of antibody called heavy chain Abs (HCAbs), that lack the entire light chain and the first heavy chain C region (CHI) compared to regular antibodies.

[0238] A dual-binding detection domain can comprise an antigen-binding domain or fragment of a chimeric, humanized, or fully human antibody. A dual-binding detection domain can comprise CDRs grafted onto a humanized or fully human framework sequence. A dualbinding detection domain can comprise a chimeric antibody wherein a portion of the heavy and / or light chain (e.g., variable region) is identical to or homologous to a corresponding sequence in an antibody derived from a particular species or belonging to a particular antibody class or subclass, while the remainder of the chain(s) is identical or homologous to a corresponding sequence in an antibody derived from another species or belonging to another antibody class or subclass, or an antigen-binding fragment of such an antibody. In some embodiments, a dual-binding detection domain comprises an antigen-binding domain or fragment that is not of a chimeric, humanized, or fully human antibody, for example, from a non-human mammalian antibody, a camelid, or another species disclosed herein.

[0239] For human administration, monoclonal antibodies or fragments thereof generated from non-human species that will be used in a dual-binding detection domain or reversibly activatable polypeptide can be further refined by a humanization process to reduce the likelihood of immunogenicity while preserving target specificity. Humanization processes can involve the incorporation of human DNA to the genetic sequence of the genes that produce the isolated antibodies, and / or the removal of predicted epitopes, such as T cell epitopes.

[0240] A dual-binding detection domain can bind to an epitope of an effector domain and / or inducing moiety. Non-limiting examples of epitopes include amino acids, sugars, lipids, phosphoryl, and sulfonyl groups. An epitope can have specific three-dimensional structural characteristics, and / or specific charge characteristics. Epitopes can be conformational or linear.

[0241] Additional non-limiting examples of binding elements that a dual -binding domain or dual-binding protein can comprise include ankyrin proteins, ankyrin repeat proteins, designed ankyrin repeat proteins (DARPins), affibodies, avimers, adnectins, anticalins, Fynomers, Kunitz domains, knottins, -hairpin mimetics, and receptors and derivatives thereof.

[0242] In some embodiments, a dual-binding detection domain disclosed herein comprises an IL-2-binding domain or IL-15-binding domain of a receptor subunit, for example, IL-2Ra, IL-2RP, ZL-2Ry, or IL-15Ra, that is modified to be capable of binding both IL-2 and LAG-3 or IL-15 and LAG-3 (e.g., but only bind one of the targets at a time).C. Immunoglobulin constant domain

[0243] A reversibly activatable polypeptide can comprise an immunoglobulin constant domain and / or a stabilizing domain. Linking a dual-binding detection domain and effector domain to an immunoglobulin constant domain and / or stabilizing domain can facilitate, for example, a simpler dosing regimen and / or longer half-life of the reversibly activatable polypeptide within a subject. A stabilizing domain can be or can comprise any immunoglobulin constant domain disclosed herein. In some embodiments, a stabilizing domain comprises albumin or a fragment or derivative thereof.

[0244] An immunoglobulin constant domain can be described with reference to the basic four chain antibody unit, which comprises two heavy chain (H) polypeptide sequences and two light chain (L) polypeptide sequences. Each of the heavy chains can comprise one N-terminal heavy chain variable domain (VH) and three or four C-terminal constant domains (CHI, CH2, and CH3, and in some cases CH4). Each of the light chains can comprise one N-terminal light chain variable domain (VL) and one C-terminal constant (CL) domain. The light chain variable domain is aligned with the heavy chain variable domain and the light chain constant domain is aligned with heavy chain constant domain CHI. Each light chain is linked to a heavy chain by one covalent disulfide bond. The two heavy chains are linked to each other by one or more disulfide bonds depending on the heavy chain isotype. Each heavy and light chain also comprises regularly-spaced intrachain disulfide bridges. The C-terminal constant domains of the heavy chains (e.g., CH2 and CH3, or CH2, CH3, and CH4) comprise the Fc region, Fc domain, or Fc fragment of the antibody, which can mediate effector functions, for example, through interactions with Fc receptors or complement proteins. The terms “VH” and “HV” can be used interchangeably herein to refer to a heavy chain variable domain. The terms “VL” and “LV” can be used interchangeably herein to refer to a light chain variable domain.

[0245] The light chain can be designated kappa or lambda based on the amino acid sequence of the constant region. The heavy chain can be designated alpha, delta, epsilon, gamma, or mu based on the amino acid sequence of the constant region. Antibodies can be categorized into five immunoglobulin classes, or isotypes, based on the heavy chain. IgA comprises alpha heavy chains, IgD comprises delta heavy chains, IgE comprises epsilon heavy chains, IgG comprises gamma heavy chains, and IgM comprises mu heavy chains. Antibodies of the IgG, IgD, and IgEclasses comprise monomers of the four chain unit described above (two heavy and two light chains), while the IgM and IgA classes can comprise multimers of the four chain unit. The alpha and gamma classes are further divided into subclasses on the basis of differences in the sequence and function of the heavy chain constant region. Subclasses of IgA and IgG expressed by humans include IgGl, IgG2, IgG3, IgG4, IgAl, and IgA2.

[0246] Generally / natively the constant regions of an antibody can mediate various effector functions, while the variable regions primarily mediate antigen binding. Different IgG isotypes or subclasses can be associated with different effector functions or therapeutic characteristics, for example, because of interactions with different Fc receptors and / or complement proteins. Reversibly activatable polypeptides comprising constant domains or Fc regions that engage activating Fc receptors can, for example, participate in antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cellular phagocytosis (ADCP), complementdependent cytotoxicity (CDC), induction of signaling through immunoreceptor tyrosine-based activation motifs (IT AMs), and induction of cytokine secretion. Reversibly activatable polypeptides comprising Fc regions that engage inhibitory Fc receptors can, for example, induce signaling through immunoreceptor tyrosine-based inhibitory motifs (ITIMs). In some embodiments, immunoglobulin constant domains in a reversibly activatable polypeptide are modified to reduce or modulate such effector functions.

[0247] Different antibody subclasses comprise varying abilities to elicit immune effector functions, which can be utilized or modified in reversibly activatable polypeptides disclosed herein. For example, wild type IgGl and IgG3 can effectively recruit complement to activate CDC, and IgG2 elicits minimal ADCC. IgG4 has a lesser ability to trigger immune effector functions and can be used, e.g., where reduced immune effector functions triggered by the immunoglobulin constant domain are desired.

[0248] Reversibly activatable polypeptides disclosed herein can comprise an immunoglobulin constant domain from a heavy chain and / or light chain of an antibody isotype, class, or subclass disclosed herein. As used herein, “immunoglobulin constant domain” does not necessarily refer to the full constant region of an immunoglobulin chain. Immunoglobulin constant domain can describe at least one domain from the full immunoglobulin constant region. For example, immunoglobulin constant domain can describe a CHI domain only or a variant, derivative or fragment thereof; a CH2 domain only or a variant, derivative, or fragment thereof; a CH3 domain only or a variant, derivative, or fragment thereof; a CH2 and CH3 domain without CHI; a CH2, CH3, and a hinge or fragment thereof without a CHI domain; or a CHI, CH2, and CH3, with or without a hinge. In each case, the CHI, CH2, CH3, or hinge, may be a variant, derivative, or fragment thereof. In some embodiments, an immunoglobulin constantdomain is a CH2 and CH3, for example of an IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, an immunoglobulin constant domain is a CH2, CH3, and a hinge or fragment thereof, for example of an IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, an immunoglobulin constant domain is a CHI, hinge, CH2, and CH3, for example of an IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, an immunoglobulin constant domain is a CL.

[0249] A hinge region of an antibody can be subdivided into three domains, an upper hinge domain, core hinge domain, and lower hinge domain. A core hinge domain can comprise one or more cysteine residues that can form a disulfide bond, for example, with a corresponding core hinge region of a second immunoglobulin constant region. In some embodiments, a hinge or fragment thereof in an immunoglobulin constant domain comprises, consists essentially of, or consists of the upper hinge domain, core hinge domain, and / or lower hinge domain. In some embodiments, an immunoglobulin constant domain can lack an upper hinge region, for example, to reduce susceptibility to proteolysis. In some embodiments, a hinge or a part thereof (e.g., an upper hinge region) can be replaced by a linker disclosed herein. In some embodiments, an immunoglobulin constant domain comprises a first hinge domain or hinge fragment from a first immunoglobulin isotype or subclass, and a second hinge domain or hinge fragment from a second immunoglobulin isotype or subclass. For example, in some embodiments an immunoglobulin constant domain comprises a core hinge domain of IgGl and a lower hinge domain of IgG2, for example, to enhance resistance to proteases and / or reduce effector function.

[0250] A reversibly activatable polypeptide can comprise an immunoglobulin constant domain that is a heavy chain constant domain, for example, a CHI, CH2, CH3, and / or CH4 domain, or a variant, derivative, or fragment thereof. The heavy chain constant domain can be a mammalian heavy chain constant domain. The heavy chain constant domain can be a human heavy chain constant domain.

[0251] The immunoglobulin constant domain can be or can comprise any one or more domains from any suitable immunoglobulin isotype(s), class(es), or subclass(es) (e.g., IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, an immunoglobulin heavy chain constant domain can be a CHI, CH2, and / or CH3 of IgGl or IgG4 (e.g., mammalian, human, or other IgGl or IgG4).

[0252] A reversibly activatable polypeptide can comprise an immunoglobulin constant domain of IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE, for example, mammalian or human IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, a reversibly activatable polypeptide comprises a CHI, hinge, CH2, CH3, fragment thereof, or a combination thereof, of IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2,IgM, IgD, or IgE. In some embodiments, a reversibly activatable polypeptide comprises a CH2 and CH3 of IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, a reversibly activatable polypeptide comprises a CH2, CH3, and hinge or fragment thereof of IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, a reversibly activatable polypeptide comprises an Fc domain of IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. In some embodiments, a reversibly activatable polypeptide does not include a CHI domain of IgG, IgGl, IgG2, IgG3, IgG4, IgA, IgAl, IgA2, IgM, IgD, or IgE. The immunoglobulin constant domain can comprise a modification, e.g., to induce heterodimerization, increase resistance to proteases, alter Fc receptor binding, and / or alter effector function. In some embodiments, an immunoglobulin constant domain comprises a single chain Fc domain, for example, parts of two heavy chains within the one open reading frame.

[0253] A reversibly activatable polypeptide can comprise an immunoglobulin constant domain of IgG, for example, mammalian or human IgG. In some embodiments, a reversibly activatable polypeptide comprises a CHI, hinge, CH2, CH3, fragment thereof, or a combination thereof, of IgG. In some embodiments, a reversibly activatable polypeptide comprises a CH2 and CH3 of IgG. In some embodiments, a reversibly activatable polypeptide comprises a CH2, CH3, and hinge or fragment thereof of IgG. In some embodiments, a reversibly activatable polypeptide comprises an Fc domain of IgG. In some embodiments, a reversibly activatable polypeptide does not include a CHI domain of IgG. SEQ ID NOs: 8-12 provide an illustrative sequences of full constant regions of human IgG, which comprises a CHI domain, hinge, CH2 domain, and CH3 domain. SEQ ID NOs: 18-63 provide illustrative sequences of IgG CHI, CH2, CH3, hinge, and Fc domains. The immunoglobulin constant domain can comprise a modification, e g., to induce heterodimerization, increase resistance to proteases, alter Fc receptor binding, and / or alter effector function. In some embodiments, an immunoglobulin constant domain comprises a single chain Fc domain, for example, parts of two heavy chains within the one open reading frame.

[0254] A reversibly activatable polypeptide can comprise an immunoglobulin constant domain of IgA, for example, mammalian or human IgAl or IgA2. In some embodiments, a reversibly activatable polypeptide comprises a CHI, hinge, CH2, CH3, fragment thereof, or a combination thereof, of IgA. In some embodiments, a reversibly activatable polypeptide comprises CH2 and CH3 of IgA. In some embodiments, a reversibly activatable polypeptide comprises a CH2, CH3, and hinge or fragment thereof of IgA. In some embodiments, a reversibly activatable polypeptide comprises an Fc domain of IgA. In some embodiments, a reversibly activatable polypeptide does not include a CHI domain of IgA. SEQ ID NO: 14 andSEQ ID NO: 15 provide illustrative sequences of full constant region of human IgAl and IgA2, respectively, which each comprises a CHI domain, hinge (e.g., SEQ ID NO: 64 and SEQ ID NO: 65), CH2 domain, and CH3 domain. The immunoglobulin constant domain can comprise a modification, e.g., to induce heterodimerization, increase resistance to proteases, alter Fc receptor binding, and / or alter effector function.

[0255] An immunoglobulin constant domain can comprise a light chain constant domain, for example, a CL domain. The light chain constant domain can be a mammalian light chain constant domain. The light chain constant domain can be a human light chain constant domain. In some embodiments, the light chain constant domain is a murine, rodent, canine, feline, equine, porcine, primate, or bovine light chain constant domain. In some embodiments, the light chain constant domain is a non-human light chain constant domain.

[0256] The immunoglobulin light chain constant domain can be or can comprise a domain from any suitable immunoglobulin isotype, class, or subclass. For example, an immunoglobulin light chain constant domain can be a lambda (IgL) or kappa (IgK) CL domain (e.g., mammalian, human, or other CL).

[0257] A reversibly activatable polypeptide can comprise an immunoglobulin constant domain of IgL or IgK, for example, mammalian or human IgL or IgK. In some embodiments, a reversibly activatable polypeptide comprises a CL of IgL. In some embodiments, a reversibly activatable polypeptide comprises a CL of IgK. SEQ ID NOs: 66 and 67 provide illustrative sequences of IgK and IgL constant domains, respectively.

[0258] SEQ ID NOs: 8-67 provide illustrative sequences of immunoglobulin constant domains that can be used (e.g., in full or in part) or adapted in compositions and methods disclosed herein.

[0259] An immunoglobulin constant domain disclosed herein can comprise, consist essentially of, or consist of an amino acid sequence with at least about 70%, at least about 71%, at least about 72%, at least about 73%, at least about 74%, at least about 75%, at least about 76%, at least about 77%, at least about 78%, at least about 79%, at least about 80%, at least about 81%, at least about 82%, at least about 83%, at least about 84%, at least about 85%, at least about 86%, at least about 87%, at least about 88%, at least about 89%, at least about 90%, at least about 91%, at least about 92%, at least about 93%, at least about 94%, at least about 95%, at least about 95.5%, at least about 96%, at least about 96.5%, at least about 97%, at least about 97.5%, at least about 98%, at least about 98.5%, at least about 99%, at least about 99.5%, or about 100% sequence identity or sequence similarity to any one of SEQ ID NOs: 8-67. The immunoglobulin constant domain can comprise one or more modifications relative to adisclosed sequence, for example, Fc modifications and / or heterodimerization domain modifications as disclosed herein.

[0260] An immunoglobulin constant domain disclosed herein can comprise, consist essentially of, or consist of an amino acid sequence with at most about 70%, at most about 71%, at most about 72%, at most about 73%, at most about 74%, at most about 75%, at most about 76%, at most about 77%, at most about 78%, at most about 79%, at most about 80%, at most about 81%, at most about 82%, at most about 83%, at most about 84%, at most about 85%, at most about 86%, at most about 87%, at most about 88%, at most about 89%, at most about 90%, at most about 91%, at most about 92%, at most about 93%, at most about 94%, at most about 95%, at most about 95.5%, at most about 96%, at most about 96.5%, at most about 97%, at most about 97.5%, at most about 98%, at most about 98.5%, at most about 99%, or at most about 99.5% sequence identity or sequence similarity to any one of SEQ ID NOs: 8-67.

[0261] In some embodiments, an immunoglobulin constant domain comprises, consists essentially of, or consists of an amino acid sequence with about 70%, about 71%, about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about 82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about 90%, about 91%, about 92%, about 93%, about 94%, about 95%, about95.5%, about 96%, about 96.5%, about 97%, about 97.5%, about 98%, about 98.5%, about 99%, or about 99.5% or about 100% sequence identity or sequence similarity to any one of SEQ ID NOs: 8-67.

[0262] In some embodiments, the immunoglobulin constant domain comprises, consists essentially of, or consists of the amino acid sequence of any one of SEQ ID NOs: 8-67. The immunoglobulin constant domain can comprise one or more modifications relative to a disclosed sequence, for example, Fc modifications and / or heterodimerization domain modifications as disclosed herein.

[0263] In some embodiments, the immunoglobulin constant domain comprises an amino acid sequence with one or more insertions, deletions, and / or substitutions relative to any one of SEQ ID NOs: 8-67. In some embodiments, the insertions, deletions, and / or substitutions comprise Fc modifications and / or heterodimerization domain modifications as disclosed herein.

[0264] For example, the immunoglobulin constant domain can comprise an amino acid sequence with at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, at least 20, at least 25, or at least 30 amino acid insertions relative to any one of SEQ ID NOs: 8-67.

[0265] In some embodiments, the immunoglobulin constant domain comprises an amino acid sequence with at most 1, at most 2, at most 3, at most 4, at most 5, at most 6, at most 7, at most 8, at most 9, at most 10, at most 11, at most 12, at most 13, at most 14, at most 15, at most 16, at most 17, at most 18, at most 19, at most 20, at most 25, at most 30, at most 35, at most 40, at most 45, or at most 50 amino acid insertions relative to any one of SEQ ID NOs: 8-67.

[0266] In some embodiments, the immunoglobulin constant domain comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or 50 amino acid insertions relative to any one of SEQ ID NOs: 8-67.

[0267] The one or more insertions can be at the N-terminus, the C-terminus, within the amino acid sequence, or a combination thereof. The one or more insertions can be contiguous, non-contiguous, or a combination thereof.

[0268] In some embodiments, the immunoglobulin constant domain comprises an amino acid sequence with at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, at least 20, at least 25, or at least 30 amino acid deletions relative to any one of SEQ ID NOs: 8-67.

[0269] In some embodiments, the immunoglobulin constant domain comprises an amino acid sequence with at most I, at most 2, at most 3, at most 4, at most 5, at most 6, at most 7, at most 8, at most 9, at most 10, at most 11, at most 12, at most 13, at most 14, at most 15, at most 16, at most 17, at most 18, at most 19, at most 20, at most 25, at most 30, at most 35, at most 40, at most 45, or at most 50 amino acid deletions relative to any one of SEQ ID NOs: 8-67.

[0270] In some embodiments, the immunoglobulin constant domain comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or 50 amino acid deletions relative to any one of SEQ ID NOs: 8-67

[0271] The one or more deletions can be at the N-terminus, the C-terminus, within the amino acid sequence, or a combination thereof. The one or more deletions can be contiguous, non-contiguous, or a combination thereof.

[0272] In some embodiments, the immunoglobulin constant domain comprises an amino acid sequence with at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, at least 20, at least 25, or at least 30 amino acid substitutions relative to any one of SEQ ID NOs: 8-67. In some embodiments, the substitutions comprise Fc modifications and / or heterodimerization domain modifications as disclosed herein.

[0273] In some embodiments, the immunoglobulin constant domain comprises an amino acid sequence with at most 1, at most 2, at most 3, at most 4, at most 5, at most 6, at most 7, atmost 8, at most 9, at most 10, at most 11, at most 12, at most 13, at most 14, at most 15, at most 16, at most 17, at most 18, at most 19, at most 20, at most 25, at most 30, at most 35, at most 40, at most 45, or at most 50 amino acid substitutions relative to any one of SEQ ID NOs: 8-67. In some embodiments, the substitutions comprise Fc modifications and / or heterodimerization domain modifications as disclosed herein.

[0274] In some embodiments, the immunoglobulin constant domain comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or 50 amino acid substitutions relative to any one of SEQ ID NOs: 8-67. In some embodiments, the substitutions comprise Fc modifications and / or heterodimerization domain modifications as disclosed herein.

[0275] The one or more substitutions can be at the N-terminus, the C-terminus, within the amino acid sequence, or a combination thereof. The one or more substitutions can be contiguous, non-contiguous, or a combination thereof.D. Fc modifications

[0276] Modifications to the constant domains can affect characteristics of an antibody or reversibly activatable polypeptide that comprises the constant domain, for example, for enhancement or reduction of Fc receptor ligation, enhancement or reduction of ADCC, enhancement or reduction of ADCP, enhancement or reduction of CDC, enhancement or reduction of signaling through IT AMs, enhancement or reduction of cytokine induction, enhancement or reduction of signaling through ITIMs, or enhancement or reduction of susceptibility to protease-mediated degradation, or enhancement or reduction of half-life. Modifications can include, for example, amino acid mutations, altering post-translational modifications (e.g., glycosylation), combining domains from different isotypes or subclasses, or a combination thereof. A reversibly activatable polypeptide disclosed herein can comprise an immunoglobulin constant domain or Fc region that is modified to achieve desirable characteristics, for example, reduced binding to one or more particular Fc receptors, reduced induction of immune effector functions, increased resistance to proteases, and enhanced half-life in vivo. Binding to one or more particular Fc receptors can be increased or decreased, while binding to one or more other Fc receptors is not substantially altered. Binding to one or more particular Fc receptors can be increased, while binding to one or more other Fc receptors is not substantially altered. Binding to one or more particular Fc receptors can be decreased, while binding to one or more other Fc receptors is not substantially altered. Binding to one or more particular Fc receptors can be increased, while binding to one or more other Fc receptors is decreased. A particular Fc receptor can be, for example, a single chain IgG receptor (e.g., FcyRIIA, FcyRIIB, FcyRIIC, and FcyRIIIB), IgE receptor (e.g., FceRII), IgM receptor (e.g.,FcpR), or IgA / IgM receptors (e.g., plgR and FcapR). A particular Fc receptor can be, for example, a multichain receptor of IgA (e.g., FcaRI), IgE (e.g., FcsRI), or IgG (e.g., FcyRI, FcyRIIIA, FcyRIV, and FcRn). In some embodiments, an immunoglobulin constant domain or Fc region comprises modifications that enhance recycling via the FcRn receptor.

[0277] A reversibly activatable polypeptide disclosed herein can comprise an immunoglobulin constant domain or Fc region that is selected or modified to provide suitable characteristics, for example, suitable characteristics for treating a disease or condition as disclosed herein. In some embodiments, IgGl can be used, for example, to promote immune activation effector functions (e.g., ADCC, ADCP, CDC, IT AM signaling, cytokine induction, or a combination thereof). In some embodiments, IgG4 can be used, for example, in cases where reduced immune effector functions based on antibody binding are desirable.

[0278] Non-limiting examples of immunoglobulin constant domain modifications and their effects are provided in TABLE 2. The numbering used can be EU numbering. For example, for IgGl, numbering of the constant region according to EU numbering starts with residue number 118, and accordingly, a mutation at residue “L234” in the table below will be at residue LI 17 in SEQ ID NO: 8, residue “N434” will be residue N317 in SEQ ID NO: 8, etc. Similarly, numbering can be adjusted to the EU numbering of constant region sequences of other isotypes.E. Heterodimers

[0279] Reversibly activatable polypeptides disclosed herein can optionally be heterodimers, for example, utilize heterodimeric heavy chains.

[0280] Heterodimerization can be induced using a number of methods. Various techniques can be used to promote pairing of desirable heavy chain combinations, rather than random chain associations. Heterodimerization domain(s) can be used to facilitate formation of the heterodimer. A heterodimerization domain can be, for example, an immunoglobulin constant domain or Fc chain with one or more modifications that promote heterodimer formation. A heterodimerization domain can be or can comprise, for example, one or more modifications in an immunoglobulin constant domain or Fc chain that facilitates heterodimer formation.

[0281] In some embodiments, engineering strategies are used to introduce mutations into the CH2 and / or CH3 domains to promote heterodimerization based on steric and / or electrostatic complementarity.

[0282] Non-limiting examples of heterodimerization domains and / or strategies to induce heterodimerization of polypeptides (e.g., immunoglobulin constant domains disclosed herein, such as Fc domains) include knobs-in-holes, SEEDbody, biochemical optimization andmutations identified therefrom, electrostatic optimization / steering and mutations identified therefrom, DNL (natural association of 2 antibodies or antibody fragments anchored with DDD (dimerization and docking domain) from PKA (protein kinase A) and AD (anchoring domain) from A-kinase anchor protein (AKAP), respectively), CrossMab, LUZ-Y (e.g., leucine zipper tethered at the C-termini of HC and later proteolytically removed, plus point mutation), quadroma (e.g., somatic fusion of hybridomas each encoding a monoclonal antibody), and strand exchange.

[0283] In some embodiments, knob-in-hole modifications of immunoglobulin constant (e.g., CH3 or IgG Fc) domains are used to promote formation of heterodimers between the first immunoglobulin constant domain (e.g., heavy / Fc chain) and the second immunoglobulin constant domain (e.g., heavy / Fc chain). The “knobs in holes” approach allows the generation of complementary interacting interfaces by manipulating key amino acid residues that participate in the Fc dimeric interaction. Amino acids with small side chain are replaced by ones with larger side chains, thereby creating a knob or protrusion in one chain, and vice versa to create a hole or socket in the partner chain.

[0284] The “knob” heavy chain can contain a mutation of threonine at a position equivalent to 366 in CH3 of IgG, such as a T366W or T366Y mutation. The “knob” heavy chain can also contain, for example, an F405A mutation.

[0285] The “hole” heavy chain can contain multiple mutations, e.g., T366S, L368A, T394W, F405A, and / or Y407V / T). In some embodiments, the “hole” heavy chain comprises T366S, L368A, and Y407V substitutions. In some embodiments, the “hole” heavy chain comprises T366S, L368A, and Y407V substitutions.

[0286] The residue numbering can be according to EU numbering (e.g., as described herein).

[0287] In some embodiments, an immunoglobulin constant domain (e.g., comprising a knob or hole heterodimerization domain) comprises one or more cysteine replacement residues, for example, to facilitate formation of a disulfide bond with another immunoglobulin constant domain. The immunoglobulin constant domain can comprise a cysteine replacement, for example, at residue Y349, L351, S354, E356, E357, K392, T394, V397, D399, or a combination thereof. A first immunoglobulin constant domain and a second immunoglobulin constant domain can each comprise residues replaced with cysteines, for example, the pair can comprise K392C and D399'C; S354C and Y349'C; E356C and Y349'C; or E357C and Y349'C mutations (where the ' indicates the mutation is in the second immunoglobulin constant domain). In some embodiments, a first immunoglobulin constant domain comprises an S354C substitution and a second immunoglobulin constant domain comprises a Y349'C substitution. In someembodiments, a first immunoglobulin constant domain comprises a Y349C substitution and a second immunoglobulin constant domain comprises an E356'C substitution.

[0288] An immunoglobulin constant domain or a heterodimerization domain can comprise a combination of cysteine replacement residues and knob-in-hole modifications, for example, one or more cysteine replacement residues and one or more knob-in-hole modifications in each of a pair of immunoglobulin constant domains.

[0289] In some embodiments, a first immunoglobulin constant domain comprises substitutions at positions S354 and T366, and a second immunoglobulin constant domain comprises substitutions at positions Y349, T366, L368, and Y407. In some embodiments, a first immunoglobulin constant domain comprises S354C and T366W substitutions, and a second immunoglobulin constant domain comprises Y349C, T366S, L368A, and Y407V substitutions. In some embodiments, knob and hole mutations are introduced into the Fc chain as described in or based on the disclosure in Merchant (1998) An efficient route to human bispecific IgG. Nature Biotechnology 16, 677-681.

[0290] In some embodiments, pairs of a heterodimer comprise mutations in CH2 domain residues (e.g., F241R / F243S or F241S / F243R) that remain solvent exposed in aglycosylated IgG molecules to avoid covalent association of knob / knob or hole / hole monomers.

[0291] A SEEDbody (Strand-Exchange Engineered Domain) approach can involve creating alternating human IgG and IgA fragments in CH3 to guide heavy chain heterodimerization. For example, patches of IgG and IgA CH3 can be mutually replaced in a heterodimerization domain to facilitate heterodimerization.

[0292] In some embodiments, a heterodimerization domain developed using biochemical optimization is used to facilitate heterodimer formation. An illustrative heterodimerization domain comprises mutations in CH3 domains, for example, S364H and F405A in the first chain CH3, and Y349T and T394F in the second chain CH3.

[0293] In some embodiments, an electrostatic heterodimerization domain is used to promote heterodimer formation. A reversibly activatable polypeptide can comprise an electrostatic steering modification that favors heterodimeric interaction between a first immunoglobulin constant domain (e.g., first Immunoglobulin heavy chain constant domain) and a second first immunoglobulin constant domain (e.g., second immunoglobulin heavy chain constant domain). For example, charged pair based attraction / repulsion can be tailored in different Fc chains. Mutations in certain residues of the chains can favor heterodimer formation, e.g., (K409D- D399'K, K409D-D399'R, K409E-D399'K, K409E-D399'R, D399'K-E356'K, or K392D- E356'K). In some embodiments, charged amino acids from the core of the CH3 domain are substituted with hydrophobic residues to perturb the structural symmetry, and long-rangeelectrostatic attraction is engineered at the edge of the CH3 domain to promote heterodimer formation. Illustrative examples include a K409W and D399'V / F405'T pair; and K360E and Q347'R pair. In some embodiments, heterodimer formation is promoted by mutations (T350V / L351Y / F405A / Y407V) in a first chain heterodimerization domain and (T350V / T366L / K393L / T394W) in a second chain heterodimerization domain. In some embodiments, heterodimer formation is promoted by mutations (K409D / K392D) in a first chain heterodimerization domain and (D399'K / D357'K) in a second chain heterodimerization domain. In some embodiments, heterodimer formation is promoted by mutations (K409D / K370D) in a first chain heterodimerization domain and (D399'K / D357'K) in a second chain heterodimerization domain. In some embodiments, heterodimer formation is promoted by mutations (K409D / K392D) in a first chain heterodimerization domain and (D339'K / E356'K) in a second chain heterodimerization domain. In some embodiments, heterodimer formation is promoted by mutations (K409D / K392D) in a first chain heterodimerization domain and (D399'K / E356'K) in a second chain heterodimerization domain.

[0294] In some embodiments, an oxidation-reduction methodology or heterodimerization domain can be used for chain pairing, for example, K409 and L368 can be mutated in the CH3 domain (e.g., of IgGl / IgG2), and the chains can be co-expressed, or purified monomers can be mixed under mild reducing conditions. The K409 / L368 mutations can be introduced alone or in combination with IgGl -hinge or IgG2 -hinge mutations (if the hinge domain or a fragment thereof is present).F. Additional domains and configuration

[0295] A polypeptide disclosed herein, such as reversibly activatable polypeptide can comprise one or more linkers for example, between different domains of the polypeptide. A linker can be a chemical bond, for example, a covalent bond or a non-covalent bond. A linker as described herein can include a flexible or rigid linker. A linker can be a peptide.

[0296] A linker can comprise a linker sequence, for example, a linker peptide sequence. The length and composition of a linker can be adjusted to allow for proper folding or to increase or decrease biological activity of the effector domain or reversibly activatable polypeptide, or, e.g., to facilitate low background activity in the absence of an inducing moiety and a desirable level of effector partner signaling in the presence of the inducing moiety. A preferred linker length can be between 5 and 20 amino acids in length. A linker can be, for example, about 1, about 2, about 3, about 4, about 5, about 6, about 7, about 8, about 9, about 10, about 11, about 12, about 13, about 14, about 15, about 20, about 25, about 30, about 40, about 50, about 60, or about 70 amino acid residues in length. In some cases, a linker can be, for example at least about 2, atleast about 3, at least about 4, at least about 5, at least about 6, at least about 7, at least about 8, at least about 9, at least about 10, at least about 15, at least about 20, at least about 30, at least about 40, or at least about 50 amino acids in length. In some cases, a linker can be, for example at most about 2, at most about 3, at most about 4, at most about 5, at most about 6, at most about 7, at most about 8, at most about 9, at most about 10, at most about 15, at most about 20, at most about 30, at most about 40, at most about 50, at most about 60, at most about 70, at most about 80, or at most about 100 amino acids in length. In some cases, a linker is 5-20 amino acids in length. In some cases, a linker is 10-20 amino acids in length.

[0297] A flexible linker can have a sequence containing glycine residues. The small size of the glycine residues can provide flexibility, and allow for mobility of the connected protein domains. The incorporation of serine or threonine can maintain the stability of the linker in aqueous conditions by forming hydrogen bonds with the water molecules, thereby reducing unfavorable interactions between the linker and protein moieties. In some cases flexible linkers can also contain additional amino acids, such as threonine and alanine, to maintain flexibility, and / or polar amino acids such as lysine and glutamine, to improve solubility.

[0298] A rigid linker can have, for example, an alpha helix-structure. An alpha-helical rigid linker can act as a spacer between protein domains. A rigid linker can have a proline-rich sequence, (XP)n, with X designating alanine, lysine, glutamine, or any amino acid, and n designating a number of repeats. The presence of proline in non-helical linkers can increase stiffness, and allow for effective separation of protein domains.

[0299] A linker can comprise a hinge region, for example an amino acid sequence derived from a hinge region of an antibody or immune receptor. In some embodiments, a linker comprises a hinge region from CD8a, IgGl, or IgG4.

[0300] Examples of linkers include, but are not limited to, those disclosed in SEQ ID NOs: 68-78 and 92, repeats thereof (for example, 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 repeats), and combinations thereof, which can be used to link any portion (e.g., domain) of a polypeptide to any other portion (e.g., domain).

[0301] In some embodiments, a linker used in a reversibly activatable polypeptide is a non- cleavable linker, for example, does not contain a protease cleavage site or lacks a human protease cleavage site.

[0302] A non-cleavable linker can also include a non-proteolytically cleavable peptide. A non-proteolytically cleavable peptide can be inert to proteases present in a given sample or organism. For example, a peptide may be inert to all human protease cleavage sequences, and thereby can comprise a high degree of stability within humans and human samples. Such a peptide can also comprise a secondary structure which renders a protease cleavage site inert orinaccessible to a protease. A non-cleavable linker of the present disclosure can comprise a halflife for cleavage of at least 1 hour, at least 2 hours, at least 4 hours, at least 8 hours, at least 12 hours, at least 16 hours, at least 1 day, at least 2 days, at least 3 days, at least 1 week, at least 2 weeks, or at least 1 month, e.g., in the presence of human proteases at 25 °C or 37 °C in pH 7 or 7.4 buffer.

[0303] A non-cleavable linker of the present disclosure can include a chemical linker that is stable. Examples of non-cleavable linkers can include a thioether linker, an alkyl linker, a polymeric linker. A linker can be an SMCC linker or a PEG linker.

[0304] A linker of the disclosure can include a chemical linker. For example, two compounds (e.g., amino acid sequences) of the disclosure can be connected together by a chemical linker. Each chemical linker of the disclosure can be alkylene, alkenylene, alkynylene, heteroalkylene, cycloalkylene, heterocycloalkylene, arylene, or heteroarylene, any of which is optionally substituted. In some embodiments, a chemical linker of the disclosure can be an ester, ether, amide, thioether, or polyethyleneglycol (PEG). Non-limiting examples of such linkers include diesters of dicarboxylic acids, such as oxalyl diester, malonyl diester, succinyl diester, glutaryl diester, adipyl diester, pimetyl diester, fumaryl diester, maleyl diester, phthalyl diester, isophthalyl diester, and terephthalyl diester. Non-limiting examples of such linkers include diamides of diamino linkers, such as ethylene diamine, l,2-di(methylamino)ethane, 1,3- diaminopropane, l,3-di(methylamino)propane, l,4-di(methylamino)butane, 1,5- di(methylamino)pentane, l,6-di(methylamino)hexane, and pipyrizine.

[0305] A polynucleotide encoding a reversibly activatable polypeptide of the disclosure can be designed to encode two or more components of a protein linked by one or more 2A linkers, which can be processed into separate polypeptides co-translationally or after translation. Inclusion of a 2A linker can increase the likelihood that an appropriate ratio of components are produced (e.g., a 1 : 1, 1:2, 1 :3, 1:4, or 1:5 ratio of two components), such as heavy chains and / or light chains.

[0306] Reversibly activatable polypeptides as disclosed herein are provided in various configurations. In some embodiments, a reversibly activatable polypeptide utilizes the basic four chain scaffold structure of an immunoglobulin (e.g., IgG), or a part or derivative thereof.

[0307] An illustrative structure that can be used for a reversibly activatable polypeptide is provided in FIG. 3. The reversibly activatable polypeptide can comprise a first heavy chain that comprises, consists essentially of, or consists of, from N-to-C-terminus, the effector domain, a linker, the VH of a dual-binding detection domain, and an immunoglobulin constant domain (e.g., comprising CHI, hinge, CH2, CH3, and optionally CH4). The first light chain can comprise, consist essentially of, or consist of, from N-to-C terminus, the VL of the dual-bindingdetection domain, and a CL. In some embodiments, the reversibly activatable polypeptide is symmetrical, e g., can comprise a second heavy chain and a second light chain of the same or substantially similar sequence or structure as the first heavy chain and the first light chain, respectively.

[0308] In some embodiments, the reversibly activatable polypeptide is asymmetrical, e.g., can comprise a second heavy chain and a second light chain of different sequences than the first heavy chain and the first light chain, respectively. The second heavy chain and second light chain can comprise different VH, VL, or CDR sequences, resulting in differential binding specificity and / or affinity. In some embodiments the antigen-binding site formed by the second VH and the second VL can be a second dual-binding detection domain, and the same effector domain or a different effector domain can be linked to the second heavy chain or the second light chain, and the second dual-binding detection domain can bind the same inducing moiety or a different inducing moiety. In some embodiments, different IL-2R agonists are used as effector domains linked to each arm of the reversibly activatable polypeptide, for example, IL-2 and IL- 15, IL-2 and an IL-2 mutein, or a first IL-2 mutein and a second IL-2 mutein.

[0309] In some embodiments, the reversibly activatable polypeptide is asymmetrical, for example, the second heavy chain and / or the second light chain can comprise different sequences than the first heavy and / or light chains, respectively. The second heavy chain can lack the linker and effector domain, for example, as shown in FIG. 4, and / or the second heavy chain and second light chain can comprise different VH, VL, or CDR sequences, resulting in differential binding specificity and / or affinity. In some embodiments the antigen-binding site formed by the second VH and the second VL is monospecific and binds, e g., LAG-3 or another immune checkpoint, such as PD-1 or PD-L1, or a tumor-associated antigen (TAA) or neoepitope. The antigen-binding site formed by the second VH and second VL can be antagonistic to LAG-3 or another immune checkpoint, such as PD-1 or PD-L1. In some embodiments the antigen-binding site formed by the second VH and the second VL is monospecific and binds a T cell marker, such as NKG2A, NKG2D, CD4 or CD8. In some embodiments the antigen-binding site formed by the second VH and the second VL is monospecific and binds a marker of non-T cell population that expresses LAG3. In some embodiments the antigen-binding site formed by the second VH and the second VL can be a second dual-binding detection domain.

[0310] A further illustrative structure that can be used for a reversibly activatable polypeptide is provided in FIG. 5. The reversibly activatable polypeptide can comprise a first heavy chain that comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH of a dual -binding detection domain, and an immunoglobulin constant domain (e.g., CHI, hinge, CH2, CH3, and optionally CH4). The first light chain can comprise, consists essentially of, orconsists of, from N-to-C terminus, the effector domain, a linker, the VL of the dual -binding detection domain, and an a CL. In some embodiments, the reversibly activatable polypeptide is symmetrical, e g., can comprise a second heavy chain and a second light chain of the same or substantially similar sequence or structure as the first heavy chain and the first light chain, respectively.

[0311] In some embodiments, the reversibly activatable polypeptide is asymmetrical, e.g., can comprise a second heavy chain and / or a second light chain with different sequences to those of the first heavy chain and light chain. The second heavy chain and second light chain can comprise different VH, VL, or CDR sequences, resulting in differential binding specificity and / or affinity. In some embodiments the antigen-binding site formed by the second VH and the second VL can be a second dual-binding detection domain, and the same effector domain or a different effector domain can be linked to the second heavy chain or the second light chain, and the second dual-binding detection domain can bind the same inducing moiety or a different inducing moiety.

[0312] In some embodiments, the reversibly activatable polypeptide is asymmetrical, and the second light chain can lack the linker and effector domain, for example, as shown in FIG. 6, and / or the second heavy chain and second light chain can comprise different VH, VL, or CDR sequences, resulting in differential binding specificity and / or affinity. In some embodiments the antigen-binding site formed by the second VH and the second VL is monospecific and binds, e.g., LAG-3 or another immune checkpoint, such as PD-1 or PD-L1, or a tumor-associated antigen (TAA) or neoepitope. The antigen-binding site formed by the second VH and second VL can be antagonistic to LAG-3 or another immune checkpoint, such as PD-1 or PD-L1. In some embodiments the antigen-binding site formed by the second VH and the second VL can be a second dual-binding detection domain.

[0313] In some embodiments, a reversibly activatable polypeptide is asymmetrical, and a first effector domain is linked to the N-terminus of the first heavy chain, and a second effector domain is linked the N-terminus of the second light chain, e.g., as shown in FIG. 7. The first effector domain and the second effector domain can be the same or different. In some embodiments, different IL-2R agonists are used as effector domains linked to each arm of the reversibly activatable polypeptide, for example, IL-2 and IL-15, IL-2 and an IL-2 mutein, or a first IL-2 mutein and a second IL-2 mutein.

[0314] In some embodiments, a reversibly activatable polypeptide comprises an Fc domain, and an scFv is linked to the N or C-terminus of one or both heavy chains of the Fc domain. The reversibly activatable polypeptide can be symmetrical or asymmetrical. The scFv can be appended to the N-terminus of, for example, a CHI or hinge of one both heavy chains. The scFvcan be appended to the C-terminus of, for example, a CH2, CH3, or CH4 of one both heavy chains.

[0315] In some embodiments, the scFv is further linked to an effector domain, e.g., as shown in an illustrative example in FIG. 8. For example, a second heavy chain of a reversibly activatable polypeptide can comprise, consist essentially of, or consist of, from N-to-C terminus, an effector domain, a first linker, an scFv (e.g., VH-second linker- VL, or VL-second linker- VH), optionally a third linker, and an immunoglobulin constant domain (e.g., optionally CHI, and hinge, CH2, CH3, and optionally CH4). In some embodiments the Fab of the other arm of the reversibly activatable polypeptide is monospecific and binds, e.g., LAG-3 or another immune checkpoint, such as PD-1 or PD-L1, or a tumor-associated antigen (TAA) or neoepitope. In some embodiments the Fab of the other arm of the reversibly activatable polypeptide is antagonistic to LAG-3 or another immune checkpoint, such as PD-1 or PD-L1.

[0316] In some embodiments, the scFv is not further linked to an effector domain, and, e g., can be monospecific and bind to LAG-3 or another cell surface marker, such as PD-L1 or a tumor-associated antigen (TAA) or neoepitope, as shown in an illustrative example in FIG. 9. For example, a second heavy chain of a reversibly activatable polypeptide can comprise, consist essentially of, or consist of, from N-to-C terminus, an scFv (e g., VH-linker-VL, or VL-linker- VH), optionally a second linker, and an immunoglobulin constant domain (e.g., optionally CHI, and hinge, CH2, CH3, and optionally CH4). In some embodiments the scFv is antagonistic to LAG-3 or another immune checkpoint, such as PD-1 or PD-L1.

[0317] In some embodiments, a reversibly activatable polypeptide can he heterodimeric and comprise a second heavy chain that lacks an effector domain, lacks a dual-binding detection domain, and lacks an additional binding domain. For example, a reversibly activatable polypeptide can comprise a structure in which the left hand side is similar to FIG. 6 or FIG. 9, and the right hand side comprises the heavy chain constant domain but lacks the Fab on the right hand side of FIG. 6 and / or lacks the scFv of FIG. 9. In another example, the reversibly activatable polypeptide can comprise a structure in which the right hand side is similar to FIG. 8, with an effector domain linked to dual-binding detection domain scFv, with the left hand side comprising a heavy chain constant domain but lacking the Fab of FIG. 8.

[0318] In some embodiments, the scFv is a dual-binding detection domain that is linked to the N-terminus or the C-terminus of a first heavy chain, e.g., of a heterodimer, and an effector domain is linked to the N-terminus or the C-terminus of the second heavy chain.

[0319] An illustrative example of a reversibly activatable polypeptide comprising an scFv dual-binding detection domain is provided in FIG. 10, in which the reversibly activatable polypeptide comprises first heavy and light chains that form a first Fab, second heavy and lightchains that form an Fab. The first heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, VH, CHI, hinge, CH2, CH3, optionally CH4, a linker, and an scFv that is the dual-binding detection domain. The second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, VH, CHI, hinge, CH2, CH3, optionally CH4, a linker, and the effector domain.

[0320] A further illustrative example of a reversibly activatable polypeptide comprising an scFv dual-binding detection domain is provided in FIG. 11, in which the reversibly activatable polypeptide comprises a first heavy chain and a second heavy chain that form an Fc domain. The first heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, optionally CHI, hinge, CH2, CH3, optionally CH4, a linker, and an scFv that is the dual -binding detection domain. The second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, optionally CHI, hinge, CH2, CH3, optionally CH4, a linker, and the effector domain.

[0321] In some embodiments, a reversibly activatable polypeptide lacks an immunoglobulin constant domain or lacks an Fc domain. For example, a reversibly activatable polypeptide can comprise an effector domain linked to the N-terminus or C-terminus of an scFv, single domain antibody, VHH, nanobody, or other antigen-binding fragment disclosed herein.H. POLYNUCLEOTIDES AND VECTORS

[0322] In some embodiments the disclosure provides a polynucleotide that encodes a reversibly activatable polypeptide disclosed herein. The polynucleotide can be prepared by standard molecular biology techniques. The polynucleotide can be prepared by molecular cloning. The polynucleotide can be synthesized de novo. The polynucleotide can comprise a nucleotide sequence encoding the reversibly activatable polypeptide, operably linked to transcription regulatory sequences such as a promoter, and optionally a 3' untranslated region. A constitutive, inducible, or tissue-specific promoter can be used.

[0323] The polynucleotide can be a DNA. The polynucleotide can be an RNA. The polynucleotide can comprise a modified base, for example, to enhance stability of the polynucleotide upon administration to a subject. A polynucleotide provided can include a recombinant, artificial, or synthetic polynucleotide. The polynucleotide can be single stranded. The polynucleotide can be double stranded. The polynucleotide can be recombinant and / or isolated.

[0324] The polynucleotide can be inserted into or part of a vector, such as an expression vector, such that the genes are operatively linked to transcriptional and / or translational control sequences. The vector can comprise a selectable marker for selection of a vector-carrying hostcell. The vector can lack a selectable marker. The vector can comprise an origin of replication or can lack an origin of replication. The vector can be a plasmid, for example, a nanoplasmid. The vector can be a minicircle. The vector can be a liner polynucleotide phagemid, cosmid, RNA vector, viral vector or the like. Non-limiting examples of viral vectors include a retrovirus (e.g., lentivirus), an adenovirus, and an adeno-associated virus. The vector can be a nonviral vector.

[0325] In some embodiments, a polynucleotide encoding a reversibly activatable polypeptide is packaged in a lipid-based delivery vector, such as a liposome or lipid nanoparticle.HL PHARMACEUTICAL COMPOSITIONS

[0326] Compositions disclosed herein can comprise a polypeptide, polynucleotide, or vector and a pharmaceutically-acceptable excipient (e.g., vehicle, carrier, or diluent). For example, in some embodiments the disclosure provides a pharmaceutical composition comprising a reversibly activatable polypeptide and a pharmaceutically-acceptable excipient (e.g., vehicle, carrier, or diluent). The pharmaceutical composition can be in a unit dosage form.

[0327] A pharmaceutical composition disclosed herein can comprise a saline solution. A pharmaceutical composition disclosed herein can comprise a buffered saline solution, for example, PBS, dPBS, HBSS, or the like. A pharmaceutical composition disclosed herein can comprise Ringer's solution, dextrose solution, or Hank's solution.

[0328] A pharmaceutical composition disclosed herein can comprise a buffer, for example, a citrate buffer (e.g., sodium citrate) or a phosphate buffer (e.g., sodium phosphate buffer). A pharmaceutical composition disclosed herein can comprise a pH-stabilizing agent.

[0329] A pharmaceutical composition disclosed herein can comprise an organic co-solvent, e.g., polysorbate 20, polysorbate 80, propylene glycol, or polyethylene glycol (PEG).

[0330] A pharmaceutical composition disclosed herein can comprise a stabilizing agent, e.g., sucrose, sorbitol, glycerol, trehalose, or mannitol.

[0331] A pharmaceutical composition disclosed herein can comprise a tonicity agent, e.g., a salt, such as NaCl or KC1.

[0332] A pharmaceutical composition disclosed herein can comprise a preservative. A pharmaceutical composition disclosed herein can comprise an antimicrobial agent. A pharmaceutical composition disclosed herein can comprise an antifungal agent.

[0333] In some embodiments, a polypeptide disclosed herein is in an aqueous buffer. In some embodiments, a polypeptide disclosed herein is in a powdered (e.g., lyophilized) form, e.g., for re-constitution with a suitable vehicle, e.g., sterile pyrogen-free water, before use.

[0334] Parenteral injections can be formulated for bolus injection or continuous infusion. The pharmaceutical compositions can be in a form suitable for parenteral injection as a sterile suspension, solution, or emulsion in oily or aqueous vehicles, and can contain formulatory agents such as suspending, stabilizing and / or dispersing agents.

[0335] Aqueous injection suspensions can contain substances which increase the viscosity of the suspension, such as sodium carboxymethyl cellulose, sorbitol, or dextran. The suspension can also contain suitable stabilizers or agents which increase the solubility and / or reduces aggregation.

[0336] Non-limiting examples of pharmaceutically-acceptable excipients, vehicles, carriers, and diluents can be found, for example, in Remington: The Science and Practice of Pharmacy, Nineteenth Ed (Easton, Pa.: Mack Publishing Company, 1995); Hoover, John E., Remington’s Pharmaceutical Sciences, Mack Publishing Co., Easton, Pennsylvania 1975; Liberman, H.A. and Lachman, L., Eds., Pharmaceutical Dosage Forms, Marcel Decker, New York, N.Y., 1980; and Pharmaceutical Dosage Forms and Drug Delivery Systems, Seventh Ed. (Lippincott Williams & Wilkinsl999), and updated versions thereof, each of which is incorporated by reference in its entirety.

[0337] Pharmaceutical compositions comprising a reversibly activatable polypeptide described herein can be manufactured, for example, by expressing the reversibly activatable polypeptide in a recombinant system, purifying the reversibly activatable polypeptide, lyophilizing the reversibly activatable polypeptide, mixing, or dissolving. The pharmaceutical compositions can include at least one pharmaceutically acceptable excipient (e g., vehicle, carrier, or diluent) and compounds described herein as free-base or pharmaceutically-acceptable salt form.IV. METHODS

[0338] The disclosure encompasses methods of treating a subject and compositions for use in a method of treating a subject. For example, a reversibly activatable polypeptide disclosed herein can be useful for treating a condition in a subject in need thereof.

[0339] In some embodiments, provided is a method of treating a condition in a subject in need thereof, the method comprising administering to the subject an effective amount of a reversibly activatable polypeptide disclosed herein. In some embodiments, provided is a method of treating a condition in a subject in need thereof, the method comprising administering to the subject an effective amount of a polynucleotide encoding a reversibly activatable polypeptide disclosed herein, or a vector comprising the polynucleotide. The effective amount can be a therapeutically-effective amount.

[0340] In some embodiments, the reversibly activatable polypeptide is administered in an amount sufficient to act as an immune checkpoint (e.g., LAG-3) antagonist, e.g., inhibit LAG-3 activity or signaling.

[0341] In some embodiments, a reversibly activatable polypeptide is administered as part of a combination therapy with an immune checkpoint inhibitor, for example, an anti-PDl or anti- PDL1 antibody. The reversibly activatable polypeptide and the immune checkpoint inhibitor can be administered together or separately.

[0342] The condition can be a cancer. The condition can be a solid tumor. The condition can be a hematologic tumor. In some embodiments, the cancer is metastatic. In some embodiments, the cancer is not metastatic. The cancer can be an immune cell cancer. The cancer can be a B cell cancer. The cancer can be a T cell cancer. The cancer can be a myeloid cell cancer. The cancer can be a leukemia. The cancer can be a lymphoma. The cancer can be a myeloma. The cancer can be a carcinoma. The cancer can be a sarcoma. The cancer can be an adenoma.

[0343] The condition can be an autoimmune disease. The condition can be an infectious disease, for example, a chronic or acute infectious disease. The condition can be a bacterial or viral infectious disease, for example, or an infectious disease caused by a eukaryotic parasite.

[0344] In some embodiments, the present disclosure provides a method for treating a cancer, the method comprising administering to a subject in need thereof an effective amount of a reversibly activatable polypeptide of the present disclosure, or a pharmaceutical composition comprising the reversibly activatable polypeptide. In some embodiments, the cancer, cancer microenvironment, subject, immune cells from the subject, and / or T cells from the subject comprise LAG-3+ or LAG-3 high cells, for example, as determined by a tissue biopsy, immunohistochemistry, or flow cytometry, ELISA, or another suitable assay.

[0345] The reversibly activatable polypeptide can promote proliferation, survival, and / or persistence of anti-tumor or tumor-specific immune cells (e.g., T cells), for example, in the tumor microenvironment or in proximity to cancer cells.

[0346] The reversibly activatable polypeptide can enhance activation of anti-tumor or tumor-specific immune cells (e.g., T cells or activated NK cells), for example, in the tumor microenvironment or in proximity to cancer cells.

[0347] The reversibly activatable polypeptide can enhance effector function (e.g., pro- inflammatory cytokine production, chemotaxis, and / or cancer cell killing) of anti-tumor or tumor-specific immune cells (e.g., T cells), for example, in the tumor microenvironment or in proximity to cancer cells.

[0348] In some embodiments, the reversibly activatable polypeptide reduces exhaustion of anti-tumor or tumor-specific immune cells (e.g., T cells), for example, in the tumor microenvironment or in proximity to cancer cells.

[0349] In some embodiments, the anti-tumor or tumor-specific immune cells comprise activated T cells, LAG-3+ T cells, LAG-3 high T cells, antigen-experienced T cells, effector T cells, effector / memory T cells, memory T cells, central memory T cells, resident memory T cells, CD8+ T cells, activated CD8+ T cells, antigen-experienced CD8+ T cells, CD8+ effector T cells, CD8+ effector / memory T cells, CD8+ memory T cells, CD8+ central memory T cells, CD8+ resident memory T cells, CD4+ T cells (e.g., TH1, TH2, TH9, or TH17), activated CD4+ T cells, antigen-experienced CD4+ T cells, CD4+ effector T cells, CD4+ effector / memory T cells, CD4+ memory T cells, CD4+ central memory T cells, CD4+ resident memory T cells, gamma delta T cells (e.g., LAG-3+ or LAG-3 high), NKT cells (e.g., LAG-3+ or LAG-3 high), lymphocytes (e.g., LAG-3+ or LAG-3 high), tumor-infiltrating lymphocytes (e.g., LAG-3+ or LAG-3 high), plasmacytoid dendritic cells (pDCs; e.g., LAG-3+ or LAG-3 high), B cells, activated B cells (e.g., LAG-3+ or LAG-3 high), NK cells (e.g. activated NK cells), monocytes, macrophages, dendritic cells, or a combination thereof. In some embodiments, the anti-tumor or tumor-specific immune cells are LAG-3+ or LAG-3 high (e.g., activated in cis'). In some embodiments, the anti-tumor or tumor-specific immune cells are LAG-3- or LAG-3 low (e.g., activated in trans).

[0350] In some embodiments, the reversibly activatable polypeptide inhibits tumor growth in the subject. In some embodiments, the reversibly activatable polypeptide inhibits tumor growth as compared to a control agent, such as an immune checkpoint inhibitor (e.g., PD-1, PD- Ll, LAG-3, or CTLA-4 inhibitor), combination of immune checkpoint inhibitors (e.g., PD-1 plus LAG-3 inhibitors, or PD-L1 plus LAG-3 inhibitors), agent with constitutively active effector domain, agent that is not targeted (e.g., lacks a monospecific LAG-3-binding domain), inducible agent that is irreversibly activated in response to a stimulus, or a combination thereof. In some embodiments, the reversibly activatable polypeptide inhibits tumor growth when administered as a monotherapy as compared to a control agent. In some embodiments, the reversibly activatable polypeptide inhibits tumor growth when administered as a combination therapy as compared to a control agent, for example, when the reversibly activatable polypeptide is administered in combination with an immune checkpoint inhibitor, such as an anti -PD-1 or anti-PDLl checkpoint inhibitor.

[0351] In some embodiments, a reversibly activatable polypeptide (e.g., as a monotherapy) inhibits tumor growth at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, atleast about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to an untreated control.

[0352] In some embodiments, a reversibly activatable polypeptide (e.g., as a monotherapy) inhibits tumor growth at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to a control agent, such as an immune checkpoint inhibitor (e.g., PD-1, PD-L1, LAG-3, or CTLA-4 inhibitor), combination of immune checkpoint inhibitors (e.g., PD-1 plus LAG-3 inhibitors, or PD-L1 plus LAG-3 inhibitors), agent with constitutively active effector domain, agent that is not targeted (e.g., lacks a monospecific LAG-3 -binding domain), inducible agent that is irreversibly activated in response to a stimulus, or a combination thereof.

[0353] In some embodiments, a reversibly activatable polypeptide (e.g., as a monotherapy) improves average survival duration or percentage by at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to an untreated control.

[0354] In some embodiments, a reversibly activatable polypeptide (e.g., as a monotherapy) improves average survival duration or percentage by at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500- fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to a control agent, such as an immune checkpoint inhibitor (e g., PD-1, PD-L1, LAG-3, or CTLA-4 inhibitor), combination of immune checkpoint inhibitors (e.g., PD-1 plus LAG-3 inhibitors, or PD-L1 plus LAG-3 inhibitors), agent with constitutively active effector domain, agent that is not targeted (e g., lacks a monospecific LAG-3-binding domain), inducible agent that is irreversibly activated in response to a stimulus, or a combination thereof.

[0355] In some embodiments, a reversibly activatable polypeptide (e.g., as a combination therapy, such as combined with an anti-PDl or anti-PDl checkpoint inhibitor) inhibits tumor growth at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50- fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to an untreated control.

[0356] In some embodiments, a reversibly activatable polypeptide (e.g., as a combination therapy, such as combined with an anti-PDl or anti-PDl checkpoint inhibitor) inhibits tumor growth at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3-fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50- fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to a control agent, such as the reversibly activatable polypeptide as a monotherapy, an immune checkpoint inhibitor (e.g., PD-1, PD-L1, LAG-3, or CTLA-4 inhibitor), combination of immune checkpoint inhibitors (e.g., PD-1 plus LAG-3 inhibitors, or PD-L1 plus LAG-3 inhibitors), agent with constitutively active effector domain, agent that is not targeted (e.g., lacks a monospecific LAG-3-binding domain), inducible agent that is irreversibly activated in response to a stimulus, or a combination thereof.

[0357] In some embodiments, a reversibly activatable polypeptide (e.g., as a combination therapy, such as combined with an anti-PDl or anti-PDl checkpoint inhibitor) improves average survival duration or percentage by at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3 -fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to an untreated control.

[0358] In some embodiments, a reversibly activatable polypeptide (e.g., as a combination therapy, such as combined with an anti-PDl or anti-PDl checkpoint inhibitor) improves average survival duration or percentage by at least about 10%, at least about 25%, at least about 50%, at least about 2-fold, at least about 3 -fold, at least about 5-fold, at least about 10-fold, at least about 20-fold, at least about 50-fold, at least about 100-fold, at least about 500-fold, at least about 1000-fold, at least about 5000-fold, or at least about 10,000-fold compared to a control agent, such as the reversibly activatable polypeptide as a monotherapy, an immune checkpoint inhibitor (e.g., PD-1, PD-L1, LAG-3, or CTLA-4 inhibitor), combination of immune checkpoint inhibitors (e.g., PD-1 plus LAG-3 inhibitors, or PD-L1 plus LAG-3 inhibitors), agent with constitutively active effector domain, agent that is not targeted (e.g., lacks a monospecific LAG- 3 -binding domain), inducible agent that is irreversibly activated in response to a stimulus, or a combination thereof.

[0359] Toxicity can be reduced as compared to a control compound for which the effector domain is active in the absence of the inducing moiety, e.g., at the same dose or a comparable dose. For example, in some embodiments, the reversibly activatable polypeptide elicits reduced systemic or extra-tumoral inflammation, reduced systemic or extra-tumoral immune activation, reduced systemic or extra-tumoral levels of pro-inflammatory cytokines, or a combinationthereof compared to a control agent, such as a control compound with a constitutively active effector domain, or an effector domain that is irreversibly activated in response to a stimulus.

[0360] In some embodiments, treatment with the reversibly activatable polypeptide results in reduced clinical signs of IL-2-mediated toxicity as compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence of moderate, severe, or moderate-to-severe IL-2-mediated toxicity is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0361] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of hypotension requiring fluid support compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of hypotension requiring fluid support is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0362] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of hypotension requiring vasopressor support compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of hypotension requiring vasopressor support is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0363] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of vascular leak syndrome compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of vascular leak syndrome is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0364] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of pulmonary edema compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of pulmonary edema is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0365] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of renal failure compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of renal failure is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0366] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of transaminitis compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of transaminitis is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0367] Treatment with a reversibly activatable polypeptide can result in reduced incidence or severity of altered mental status compared to a control agent, such as a control compound with a constitutively active IL-2, or an IL-2 that is irreversibly activated in response to a stimulus. In some embodiments, the incidence or severity of altered mental state is at least 5%, at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, or at least 95% lower (e.g., on average) for subjects that receive the reversibly activatable polypeptide compared to the control agent.

[0368] In some embodiments, toxicity is reduced as compared to a control compound for which the dual-binding detection domain does not bind to the effector domain.

[0369] The subject can be a mammalian subject. The subject can be a human subject. In some embodiments, the subject is a murine, rodent, canine, feline, equine, porcine, primate, or bovine subject. In some embodiments, the subject is a non-human and / or non-rodent mammalian subject.

[0370] In practicing the methods of treatment or use provided herein, therapeutically- effective amounts of the reversibly activatable polypeptide described herein are administered in pharmaceutical compositions to a subject suffering from a condition. In some instances the pharmaceutical composition will affect the physiology of the animal, such as the immune system, inflammatory response, or other physiologic affect. In some embodiments, the subject is a mammal such as a human. A therapeutically-effective amount can vary widely depending on the severity of the disease, the age and relative health of the subject, the potency of the compounds used, and other factors. In some embodiments the methods comprise administering the reversibly activatable polypeptide to the subject once. In some embodiments the methods comprise administering the reversibly activatable polypeptide to the subject two times, or more.

[0371] A pharmaceutical composition can be administered in therapeutically-effective amounts by various forms and routes including, for example, intravenous, intratumoral, subcutaneous, intramuscular, rectal, aerosol, parenteral, ophthalmic, pulmonary, transdermal, vaginal, optic, nasal, oral, inhalation, dermal, intradermal, intra-articular, intrathecal, intranasal, and topical administration. A pharmaceutical composition can be administered in a local or systemic manner, for example, via injection of the reversibly activatable polypeptide described herein directly into an organ, optionally in a depot. The administering can be local. The administering can be systemic. The administering can be parenteral.

[0372] Certain methods described herein comprise administering to the subject an intravenous pharmaceutical composition comprising a reversibly activatable polypeptide of the present disclosure. Intravenous pharmaceutical compositions of reversibly activatable polypeptides include any formulation suitable for administration to a subject via any intravenous method, including a bolus, an infusion which occurs over time or any other intravenous method.

[0373] The disclosure also provides a method of selectively inducing signaling by an effector partner (e.g., IL-2RaP7, IL-2RPy, or IL-15Ra / IL-2RP'y). The method can comprise contacting or treating a population of cells with a reversibly activatable polypeptide disclosed herein. The population of cells can be treated or contacted under suitable conditions to induce signaling by the effector partner, for example, in the presence of the inducing moiety (e.g., LAG-3), or in cells with surface expression of the inducing moiety. The treating or contactingcan be in vitro. The treating or contacting can be ex vivo. The treating or contacting can be in vivo (e.g., the reversibly activatable polypeptide, a polynucleotide encoding the reversibly activatable polypeptide, a vector comprising the polynucleotide, or a pharmaceutical composition comprising the reversibly activatable polypeptide, polynucleotide, or vector can be administered to a subject).V. EXAMPLESEXAMPLE 1: Antibody discovery using phage and yeast display

[0374] This example describes the isolation of LAG3 antibodies and the isolation of a set of dual binding antibodies that bind human LAG3 and human IL-2.

[0375] Monospecific anti-LAG3 antibodies were isolated from two antibody phage display libraries, the Superhuman 2.0 antibody library from Distributed Bio, Inc. and the Generation 3 libraries Lib IE, Lib4E, Lib6A and Lib8A from Specifica, a Q2 Solutions company. The libraries were subjected to one or two rounds of selection for binding to biotinylated human LAG3 with standard protocols for phage display.

[0376] DNA was isolated from the final round of phage selection and used to prepare a library in a yeast display vector equivalent to pCTcon2 that was synthesized from the publicly available sequence. This vector was used to express the scFv libraries in yeast as Myc-tagged fusions to the Aga2p yeast mating protein, as described in Chao et al. "Isolating and engineering human antibodies using yeast surface display." Nature protocols 1.2 (2006): 755-768. The yeast library was subjected to several rounds of selection for binding to LAG3. Yeast cells were labelled with either biotinylated LAG3 at concentrations of 1-100 nM followed by incubation with streptavidin-PE and anti-Myc-FITC. Each round of selection was carried out using a Sony MA900 cell sorter, with gates drawn to capture PE and FITC double positive events. After the final round of yeast selections, plasmid DNA was recovered from yeast and transfected into E. coli to produce individual scFv clones.

[0377] Anti-LAG3 and anti-IL-2 DBAs were isolated from Tumbler antibody phage display libraries based on the Superhuman 2.0 antibody library from Distributed Bio. Tumbler antibody phage display libraries were constructed to incorporate the heavy chain CDR1, heavy chain CDR2, and light chain diversity of the Superhuman library combined with (16) heavy chain CDR3 sequences from LAG3 binding antibodies. The libraries were subjected to one or two rounds of selection for binding to biotinylated human IL-2 and / or human LAG3 at a concentration of lOOnM with standard protocols for phage display. A yeast display library was prepared from DNA isolated from the final round of phage selection.

[0378] The yeast libraries were expressed as described above and subjected to several rounds of selection for binding to LAG3 and IL-2. In some cases the selections were carried out in the presence of IL-2 receptor proteins. Each round of selection was carried out using a Sony MA900 cell sorter, with gates drawn to capture PE and FITC double positive events. After the final round of yeast selections, plasmid DNA was recovered from yeast and transfected into E. coli to produce individual scFv clones.EXAMPLE 2: Monospecific LAG3 Antibodies

[0379] This example describes the characterization of novel LAG3 monospecific antibodies. The antibodies in TABLE 3 were discovered from an SH2.0 Superhuman antibody phage display library acquired from Distributed Bio, Inc. as a starting point, as described in Example 1. The antibodies in TABLE 4 were discovered from a set of Generation 3 antibody phage display libraries (Lib IE, -4E, -6A, -8A) acquired from Specifica as a starting point, as described in Example 1.

[0380] After selection, individual scFv clones were screened for LAG3 binding. DNA was prepared from the final round of yeast selections and transformed into E. coli. Individual colonies were subjected to DNA sequencing and unique clones were selected based on the sequence information. The clones were amplified by PCR using primers that incorporated a T7 promoter and a translation initiation site in the forward primer, and cMyc and V5 tags, and a T7 terminator sequence in the reverse primer such that the coding sequence for the scFv sequence was in frame with the tags in the PCR fragment. Protein was expressed from each of the PCR fragments using a cell-free transcription / translation system (Cosmo Bio USA, Inc., PUREfrex2.1, Product # GFK-PF213 with DS Supplement, Prod. # GFK-PF005). Two control scFvs were included in each experiment. In TABLE 3 the control scFvs were expressed from the sequence of favezelimab, a published LAG3 antibody and AF388, a human PD-L1 antibody. In TABLE 3 the control scFvs were expressed from the sequence of relatlimab, a published LAG3 antibody and AF327, a human PD-1 antibody. The scFv samples were subjected to ELISA analysis to detect LAG3 (human, cynomolgus, and mouse) binding. In these experiments, wells of a 384-well plate were coated with an anti-V5 antibody (Sv5-Pkl, BioRad) at Ipg / ml overnight at 4 degrees. After washing, wells were blocked with SuperBlock (ThermoFisher, 37515) followed by addition of saturating levels of scFvs in SuperBlock. After washing, antigens were added and plates incubated for one hour. To detect LAG3 binding, a biotinylated recombinant LAG3 was used (human LAG3 His Avi, LA3-H82E9, Aero; cynomolgus LAG3, LA3-C82E5, Aero; Mouse LAG3, LA3-M82E5, Aero). Biotinylated antigens were detected using streptavidin HRP using standard methods. Varying amounts oflabelled test antigen were added to show binding and to estimate relative affinities of the different scFvs.

[0381] TABLE 3 shows the ELISA screening results on (20) scFvs from the Distributed Bio library, including EC50s for human LAG3 binding and qualitative binding results for cynomolgus LAG3, mouse LAG3 and a negative control protein, human PDL1. This set of antibodies yielded EC50s between 0.03 and 2.5 nanomolar for binding to human LAG3 protein, with high binding to cynomolgus LAG3 protein and no detectable binding to human PD-L1 protein. (19) of the (20) showed no binding to mouse LAG3 protein, while 42j_D09 showed high binding to mouse LAG3 protein.

[0382] TABLE 3: Binding of PUREfrex expressed scFvs from DBio Library to LAG3 byELISA.

[0383] TABLE 4 shows the ELISA screening results on (9) scFvs from the Specifica libraries, including EC50s for human LAG3 binding and qualitative binding results forcynomolgus LAG3, mouse LAG3, and a negative control protein, human PD-1. The antibodies yielded EC50s between 0.03 and 1.8 nanomolar for binding to human LAG3 protein, with high binding to cynomolgus LAG3 protein and no detectable binding to mouse LAG3 protein or human PD-1 protein.

[0384] TABLE 4: Binding of PUREfrex expressed scFvs from Specifica libraries to LAG3 by ELISA.

[0385] Additional LAG3 antibodies were discovered using next generation sequencing of outputs from phage and yeast display selection campaigns described above. The Distributed Bio Superhuman 2.0 and Specifica Gen 3 libraries described above were subjected to two rounds of selection for binding to biotinylated or Fc-tagged human and cynomolgus LAG3 recombinant protein at concentrations of 50-100nM. Yeast display libraries were prepared from DNA isolated from the final round of phage selection. Yeast display libraries were subjected to several rounds of selection for binding to LAG3. Yeast cells were labeled with biotinylated human and cynomolgus monkey LAG3 at concentrations of l-100nM, as described previously. Plasmid DNA was recovered from the selected rounds of yeast display and full length scFv sequences were amplified by PCR. Amplicon libraries were prepared for PacBio sequencing with the SMRTbell library prep kit and sequenced on a PacBio Sequel II sequencer.

[0386] Full length scFv sequences were processed to identify heavy and light chain variable domains and their frameworks and CDRs. Heavy CDRs and the first two light CDRs were concatenated and used as unique sequence identifiers (“FiveCDR”) and were stratified by theirIGHV assignment and yeast display library. Amino acids from HCDR3s were mapped to more general versions using physicochemical properties in pseudoHCDR3s. After filtering out those with low representation, unique pseudoHCDR3s within each IGHV classification and library were clustered based on Levenshtein distance using an agglomerative hierarchical method and ranked by the number of unique FiveCDR sequences with those having more than 20 further clustered based on the Levenstein distance of the FiveCDR sequence. This collection of clusters and subclusters were categorized based on the frequency of the pseudoHCDR3, the total number of reads represented by the cluster, the number of subclusters within a pseudoHCDR3 cluster and sequence similarity to clones that already been observed and assayed. Candidates were selected such that most categories were sampled at least once but the selection was biased towards clusters that did not include an scFv that had previously been identified in the selections described above. For each selected candidate FiveCDR sequence, the most frequent scFv DNA sequence in the PacBio dataset was determined. Where there was not a dominant scFv DNA sequence, the one that most closely matched its framework amino acid sequence was selected. These DNA sequences were corrected to match the expected linker and heavy framework 1 sequence but were otherwise left as observed and synthesized to produce individual scFv clones for further testing.

[0387] TABLE 5 shows the ELISA screening results on (10) scFvs from the NGS sequence analysis described above, including EC50s for human LAG3 binding and qualitative binding results for cynomolgus LAG3, mouse LAG3, and a negative control protein, human PD-1. These antibodies yielded EC50s between 0.3 and 4.0 nanomolar for binding to human LAG3 protein, with high or medium binding to cynomolgus LAG3 protein and no detectable binding to mouse LAG3 protein or human PD-1 protein.

[0388] TABLE 5 Binding of PUREfrex expressed scFvs identified by NGS methods to LAG3 by ELISA

[0389] The NGS-derived antibody sequences were produced by gene synthesis, cloned into mammalian expression vectors, produced as scFv-Fcs from mammalian cells, purified using standard methods and characterized for binding to human LAG3 on mammalian cells. HEK293 cells were transiently transfected with a LAG3 expression vector and grown under standard conditions. The cells were harvested, washed, counted, resuspended at 2,000,000 cells / ml and 100 microliter per well (200,000 cells) added to individual wells in a 96-well plate. The cells were washed twice with 200 microliters of flow buffer (PBS, 2%FBS, 2 mM EDTA). A 5x- dilution series in flow buffer was prepared for each antibody, ranging from 100 nM to 0.032 nM, and 100 microliters of each dilution step added to a well in the plate. After 30 minutes incubation at 4C, the wells were washed twice with flow buffer and then incubated with a secondary antibody for 30 minutes at 4C. The cells were washed twice in flow buffer and resuspended in 100 microliters per well of flow buffer. Mean fluorescence intensity for each sample was measured by flow. Two control antibodies were included in the experiment, Relatlimab (a published LAG3 antibody) and a PSMA-specific isotype control. EC50s are shown in TABLE 6. The antibodies showed binding to LAG3 expressing HEK293 cells, while the control relatlimab showed an EC50 of 04 nM and the isotype control did not bind. None of the antibodies bound to HEK293 cells that did not express LAG3.

[0390] TABLE 6. Binding of scFv-Fcs to LAG3 -expressing HEK293 cells.EXAMPLE 3: LAG3-IL2 Dual-Binding Antibodies (DBAs)

[0391] This example describes the discovery and characterization of dual binding detection domains of the present disclosure. The dual binding detection domains are antibodies & antigen-binding fragments thereof that are capable of specifically binding LAG3 and are capable of specifically binding IL-2.

[0392] LAG3-IL2 DBAs were selected as described in EXAMPLE 1. Selected scFvs were expressed using cell free expression and screened for LAG3 binding using ELISA assays as described in EXAMPLE 2.

[0393] TABLE 7 shows the screening results for (6) antibodies from two yeast selection campaigns, including EC50s for human LAG3 binding and qualitative binding results for cynomolgus LAG3, mouse LAG3, and a negative control protein, human PD-L1. This set of antibodies yielded EC50s between 0.23 and 1.3 nanomolar for binding to human LAG3 protein, with medium or high binding to cynomolgus LAG3 protein and no detectable binding to human PD-L1 protein. The positive control, an scFv based on the sequence of relatlimab (a published LAG3 antibody), showed an EC50 of 0.57 nM on human LAG3 protein and bound to cynomolgus LAG3 protein. None of the antibodies tested here bound mouse LAG3 and only the negative control, an antibody to human PD-L1, bound to PD-L1.

[0394] TABLE 7 : Binding of PUREfrex expressed DBA scFvs to LAG3 by ELISA.

[0395] TABLE 8 shows the screening results for (7) antibodies from a third yeast selection campaign, including EC50s for human LAG3 binding and qualitative binding results for cynomolgus LAG3, mouse LAG3 and a negative control protein, human PD-L1. This set of antibodies yielded EC50s between 0.12 and 1.5 nanomolar for binding to human LAG3 protein. All (7) antibodies showed binding to cynomolgus LAG3 protein and no detectable binding mouse LAG3 protein or to human PD-L1 protein. The positive control, an scFv based on the sequence of relatlimab (a published LAG3 antibody), showed an EC50 of 0.16 nM on human LAG3 protein and bound to cynomolgus LAG3 protein. Only the negative control, an antibody to human PD-L1, bound to PD-L1.

[0396] TABLE 8: Binding of PUREfrex expressed DBA scFvs to LAG3 by ELISA.

[0397] A total of (13) LAG3-IL2 dual binding antibodies were discovered during the experiments described in this example.EXAMPLE 4: Monospecific anti-LAG3 antibodies using common light chains

[0398] The example describes characterization of monospecific LAG3 antibodies that share a light chain with illustrative LAG3-IL2 DBAs. The goal of this example is to demonstrate reversibly activatable polypeptides of the present disclosure in a common light chain bispecific format, with (i) a DBA and an effector on one arm, and (ii) a monospecific LAG3 antibody Fab on the other arm, as shown in FIG. 12B. The monospecific anti-LAG3 Fabs comprise a light chain from a LAG3-IL2 DBA paired with the heavy chain from a parental LAG3 monospecific antibody. The light chains were derived from five DBA, 7w2_A02c, 7w2_C01, 7w2_Bl 1c,7w2_C06c and sn6_Cl 1c. The heavy chains were derived from three different parental LAG3 antibodies, cjp_A01, 42j_D06 and zqm_B05. Certain (e.g., preferred) common light chain LAG3 antibodies (cjp_A01_L_7w2_A02c, cjp_A01_L_7w2_C01, cjp_A01_L_7w2_Bl 1c, cjp A01 L 7w2 C06c, 42j D06 L sn6 Cl lc and zqm B05 L 7w2 C06c) were characterized for LAG3 binding. Each antibody was produced as a mAb using transient expression in HEK293 and purified with standard methods.

[0399] Four common light chain LAG3 antibodies were characterized for binding to human LAG3 on mammalian cells and compared to two of the parental LAG3 antibodies. HEK293 cells were transiently transfected with a LAG3 expression vector and grown under standard conditions. The cells were harvested, washed, counted, resuspended at 2,000,000 cells / ml and 100 microliter per well (200,000 cells) added to individual wells in a 96-well plate. The cells were washed twice with 200 microliters of flow buffer (PBS, 2%FBS, 2 mM EDTA). A 5x- dilution series in flow buffer was prepared for each antibody, ranging from 100 nM to 0.032 nM, and 100 microliters of each dilution step added to a well in the plate. After 30 minutes incubation at 4C, the wells were washed twice with flow buffer and then incubated with a secondary antibody for 30 minutes at 4°C. The cells were washed twice in flow buffer and resuspended in 100 microliters per well of flow buffer. Mean fluorescence intensity for each sample was measured by flow cytometry. Two control antibodies were included in the experiment, relatlimab (a published LAG3 antibody) and an isotype control. FIG. 13 shows a plot of the MFI in relation to antibody concentration for the four antibodies and EC50s are shown in TABLE 9. The antibodies tested here showed binding to LAG3 expressing HEK293 cells with EC50s ranging from 0.47 nM to 1.7 nM, while the parental antibodies showed EC50s of 0.92 nM and 1.6 nM. Relatlimab showed an EC50 of 0.8 nM and the isotype control did not bind. None of the antibodies bound to HEK293 cells that did not express LAG3.

[0400] TABLE 9. Binding of anti-LAG3 antibodies to LAG3 -Expressing HEK293 cells.

[0401] Additional common light chain LAG3 antibodies were selected by phage and yeast display. Common light chain Tumbler phage display libraries were generated from (8) parental anti-LAG3 binding antibodies, (6) derived from the Distributed Bio Superhuman 2.0 antibody library and (2) from the Specifica Gen3 antibody library. The Tumbler phage display libraries incorporated the heavy chain CDR3 sequence from parent monospecific antibodies and the heavy chain CDR1 and CDR2 diversity of their respective parent libraries paired with a common light chain sequence. Common light chain sequences were derived from (3) anti- LAG3-IL2 DBAs

[0402] The libraries were subjected to two rounds of selection for binding to biotinylated or Fc-tagged human and cynomolgus LAG3 at concentrations ranging from 25nM to lOOnM with standard protocols for phage display. A yeast display library was prepared from DNA isolated from the final round of phage selection as described in EXAMPLE 1. The yeast library was subjected to several rounds of selection for binding to LAG3. Yeast cells were labelled with biotinylated human or cynomolgus LAG3 recombinant protein (InM- lOOnM) followed by incubation with streptavidin-PE and anti-Myc-FITC. Each round of selection was carried out using a Sony MA900 cell sorter, with gates drawn to capture PE and FITC double positive events. After the final round of selection DNA was prepared from the yeast, transformed into E. coli and the individual scFv clones were subjected to DNA sequencing. Selected clones were screened using cell free protein expression and ELISA assays as described in EXAMPLE 2 for binding to human LAG3, cynomolgus LAG3 and a negative control protein (human PD-L1) as described herein. Results from (28) clones are shown in TABLE 10, including identifiers of the parent anti-LAG3 monospecific and anti-LAG3 / anti-IL-2 DBA from which they were derived. This set of antibody scFvs bound human LAG3 with EC50s ranging from 0.04 nanomolar to 35 nanomolar All (28) bound cynomolgus LAG3 but did not bind human PD-L1.

[0403] TABLE 10. Binding of PUREfrex expressed scFvs to LAG3 by ELISAEXAMPLE 5: LAG3 antibody activity in LAG3 / MHC-II T cell activation assay

[0404] In this example, LAG3 binding domains were assessed for their ability to functionally block LAG3 inhibitory activity using a LAG3 / MHC-II Blockade Bioassay (Promega, Catalog JA1115). MHC-II-expressing antigen presenting cells (APCs) were thawed and cultured overnight in the presence of a TCR-activating peptide antigen. The following day, Jurkat T cells expressing human LAG3 and a luciferase reporter driven by T-cell activation pathway-dependent response elements were co-cultured with the peptide-pulsed APCs in the presence of titrating concentrations of anti-LAG3 antibodies. Following five hours of co-culture, the bioluminescent signal was quantified using the Bio-Gio Luciferase Assay System (Promega) and a luminometer (Perkin Elmer).

[0405] Functional anti-LAG3 blocking antibodies release LAG3 -mediated inhibition, which results in increased luminescence (e.g., indicative of increased T cell activation in response to MHC-presented antigen). Accordingly, increased luminescence can demonstrate the ability of the anti-LAG3 binding domains to functionally block LAG3 inhibitory activity on T cells.

[0406] The (17) LAG3 antibodies in TABLE 11 were expressed as mAbs in mammalian culture, purified using standard methods and assayed as described above for functional blocking of LAG3 mediated suppression of T cell activation. Twelve of the (17) LAG3 antibodies functionally blocked LAG3 activity in this assay.

[0407] TABLE 11. Activity of anti-LAG3 antibodies in LAG3 -mediated Suppression.

[0408] The (7) LAG3 antibodies in TABLE 12 were expressed as monomeric scFvs (i.e., as the scFv arm shown in FIG. 12A). The antibody constructs were transiently expressed mammalian culture, purified using standard methods and assayed as described above for functional blocking of LAG3 mediated suppression of T cell activation. Four of the (7) LAG3 antibodies functionally blocked LAG3 activity in this assay.

[0409] TABLE 12. Activity of anti-LAG3 antibodies LAG3-mediated Suppression

[0410] Five of the common light chain antibodies described in EXAMPLE 4 were expressed as mAbs and assayed as described above for functional blocking of LAG3 mediated suppression of T cell activation. All five of the common light chain LAG3 antibodies functionally blocked LAG3 activity in this assay (TABLE 13).EXAMPLE 6: LAG3 antibody binding on primary human T cells

[0411] Anti-LAG3 antibodies were characterized for binding to LAG3 -expressing primary human T cells. CD8+ T cells were isolated from human PBMCs using immunomagnetic negative selection (STEMCELL) and stimulated with plate-bound anti-CD3 and soluble anti- CD28 for 72 hours to induce expression of LAG3. In some instances, total PBMCs were stimulated with Concanavalin A (Invivogen) and IL-2 for 72 hours. Cells were stained with Live / Dead Viability Dye (Invitrogen) and fluor ophore-conjugated antibodies directed against CD3, CD4, and CD8 (Biolegend). Cells were then incubated with titrating concentrations of anti-LAG3 antibodies followed by detection with goat anti-human Fc PE-labeled polyclonal antibody (lackson ImmunoResearch). Samples were fixed with 2% PFA and analyzed by flow cytometry.

[0412] The anti-LAG3 antibodies in TABLE 14 were expressed as mAbs from mammalian cell culture, purified using standard methods and assayed as described above for binding to CD8+ primary human T cells. FIG. 14 shows a plot of mean fluorescence intensity (MFI) forPE labeling of the CD8+ T cells for a dilution series of each antibody. TABLE 14 includes EC50s for binding of each antibody.

[0413] TABLE 14. Binding of LAG3 antibodies to Primary CD8+ T cells

[0414] The LAG3 antibodies in TABLE 15 were expressed as heterodimeric antibodies in the format of FIG. 12A, with a Fab of an anti-PSMA antibody on one arm and the anti-LAG3 scFv on the other arm of the antibody. Proteins were expressed in mammalian cells, purified using standard methods and assayed as described above for binding to CD8+ primary human T cells. FIG. 15 shows a plot of percent of PE positive CD8+ T cells for a dilution series of each antibody. TABLE 15 includes EC50s for binding for each antibody.

[0415] TABLE 15. Binding of LAG3 antibodies to Primary CD8+ T cells.

[0416] Five common light chain LAG3 antibodies in TABLE 16 were expressed as mAbs from mammalian cell culture, purified using standard methods and assayed as described above for binding to CD8+ primary human T cells. FIG. 16 shows a plot of percent of PE positive CD8+ T cells for a dilution series of each antibody. TABLE 16 includes EC50s for binding of each antibody. All five common light chain LAG3 antibodies demonstrated similar binding to LAG3 expressing primary human CD8+ cells.

[0417] TABLE 16. Binding of LAG3 antibodies to Primary CD8+ T cells

[0418] Collectively this example shows binding of (29) different LAG3 antibodies to LAG3 on the surface of CD8+ primary human T cells.

[0419] Additional constructs were tested in a similar assay characterizing binding to LAG-3- expressing primary human T cells. Total T cells were isolated from human PBMCs using immunomagnetic negative selection (STEMCELL) and stimulated with Concanavalin A for 48 hours to induce expression of LAG3. Cells were stained with Live / Dead Viability Dye (Invitrogen) and fluorophore-conjugated antibodies directed against CD3, CD4, and CD8 (Biolegend). Cells were then incubated with titrating concentrations of cLAG3-ZL2 constructs, an anti -LAG-3 monoclonal antibody (relatlimab), or an isotype control antibody. The cLAG3- IL2 constructs were heterodimeric comprising (i) on a first arm, an IL-2 variant linked to DBA with binding specificity for IL-2 and LAG3, and (ii) on a second arm, a monospecific anti- LAG3 scFv (e.g., based on anti-LAG3 antibodies disclosed herein, such as AB005297, AB005298, or AB005299). Cells were then incubated with goat anti-human Fc PE-labeled polyclonal antibody (Jackson ImmunoResearch). Samples were fixed with 2% paraformaldehyde (PF A) and analyzed by flow cytometry. The cLAG3-IL2 constructs showed comparable binding to the stimulated CD8+ T cells as the anti-LAG-3 monoclonal antibody control (FIGs. 35A-F).EXAMPLE 7 : LAG3-dependent IL-2 activity

[0420] This example describes the activity of reversibly activatable polypeptides of the present disclosure in the presence and absence of LAG3. The LAG3-IL2 DBAs described in EXAMPLE 3 were expressed as heterodimeric antibodies with the DBA linked to an IL-2 variant on one arm and an scFv based on the relatlimab or favezelimab (published LAG3 antibodies) sequences on the other arm as shown in FIG. 17H & FIG. 181. The IL-2 variant (SEQ ID NO: 79) has three mutations that reduce binding to IL-2Ra (CD25) and two mutations to facilitate production and characterization. When tested by surface plasmon resonance (SPR), CD25 did not bind to the variant at IpM CD25.

[0421] In one experiment, DBA constructs that incorporated DBAs 4bv_B01, 62d_A04, mi4_B08, kp2_A03, 8ea_B12, and 8ea_D02 were tested on HEK-Blue™ IL-2 reporter cells and on HEK-Blue™ IL-2 reporter cells expressing LAG3. A construct with an isotype-matched control antibody in place of the DBA was used as a control. Assays were carried out using a HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb-il2). Each DBA construct was assayed on HEK-Blue™ IL-2 reporter cells (which do not express LAG3) and on HEK-Blue™ IL-2 reporter cells that had been transfected with a LAG3 expression vector using Lipofectamine 3000 (ThermoFisher cat# L3000001) according to the manufacturer’s instructions. Cell surface expression of LAG3 was confirmed by flow cytometry 2 days after transfection. In 384-well ELISA plates (Coming 3701) the antibody-cytokine complexes were serially diluted 1 :4 for 8 points in complete DMEM (+10% FBS, 2 mM L-glutamine, sodium pymvate) from a starting concentration of 6 nM. HEK-Blue™ IL-2 reporter cells (12,500 cells) were added to each well and incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution (InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision. The (6) constructs showed markedly higher IL-2 activity on the LAG3 expressing cells (solid lines) than on the LAG3 negative cells (dotted lines) (FIGs. 17A-F), while the control construct showed similar activity in the presence and absence of LAG3 (FIG. 17G). The constructs were active (e.g., induced IL-2R signaling) at much lower concentrations on LAG3 positive cells than on LAG3 negative cells.

[0422] In a second experiment, DBA constructs that incorporated DBAs 7w2_A02, 7w2_Bl 1, 7w2_C02, 7w2_C06, sn6_C02, sn6_E01 and 9wg_A05 were tested on HEK-Blue™ IL-2 reporter cells and on HEK-Blue™ IL-2 reporter cells expressing LAG3. A construct with an isotype antibody in place of the DBA was tested as a control. Assays were carried out using a HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb-il2). Each DBA construct was assayed on HEK-Blue™ IL-2 reporter cells (which do not express LAG3) and on HEK-Blue™IL-2 reporter cells that had been transfected with a LAG3 expression vector using Lipofectamine 3000 (ThermoFisher cat# L3000001) according to the manufacturer’s instructions. Cell surface expression of LAG3 was confirmed by flow cytometry 2 days after transfection. In 384-well ELISA plates (Coming 3701) the antibody-cytokine complexes were serially diluted 1 :4 for 8 points in complete DMEM (+10% FBS, 2 mM L-glutamine, sodium pyruvate) from a starting concentration of 6 nM. HEK-Blue™ IL-2 reporter cells (12,500 cells) were added to each well and incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution (InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision. As shown in FIGs. 18A-G, the (7) constructs showed markedly higher IL-2 activity on the LAG3 expressing cells (solid lines) than on the LAG3 negative cells (dotted lines). The control construct showed activity in the absence of LAG3, with slightly increased activity in presence of LAG3 (FIG. 18H). The constructs are active at much lower concentrations on LAG3 positive cells than they are on LAG3 negative cells.

[0423] In further experiments, sequences of the DBA (e.g., VH, VL, and / or CDR) parts of constructs were modified to potentially improve stability of the constructs. The LAG3-IL2 DBAs were expressed as heterodimeric antibodies with the DBA linked to an IL-2 variant on one arm and an scFv based on the relatlimab sequence on the other arm as shown in FIG. 181. The IL-2 variant has three mutations that reduce binding to IL-2Rot (CD25) and two mutations to facilitate production and characterization. When tested by surface plasmon resonance (SPR), CD25 did not bind to the variant at I LIM of CD25.

[0424] In one experiment, DBA constructs that incorporated variants in the DBA kp2_A03 were tested on HEK-Blue™ IL-2 reporter cells and on HEK-Blue™ IL-2 reporter cells expressing LAG3 The variants comprised, for example, an N59S substitution in VH relative to kp2_A03 per IMGT numbering, an N36S substitution in VL relative to kp2_A03 per IMGT numbering, or both (e.g., an N54S substitution relative to SEQ ID NO: 256 and an N30S substitution relative to SEQ ID NO: 276). Assays were carried out using a HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb-il2). Each DBA construct was assayed on HEK- Blue™ IL-2 reporter cells (which do not express LAG3) and on HEK-Blue™ IL-2 reporter cells that had been transfected with a LAG3 expression vector using Lipofectamine 3000 (ThermoFisher cat# L3000001) according to the manufacturer’s instructions. Cell surface expression of LAG3 was confirmed by flow cytometry two days after transfection. In 384-well ELISA plates (Coming 3701) the antibody-cytokine complexes were serially diluted 1 :4 for 8 points in complete DMEM (+10% FBS, 2 mM L-glutamine, sodium pyruvate) from a starting concentration of 6 nM. HEK-Blue™ IL-2 reporter cells (12,500 cells) were added to each welland incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution (InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision. As shown in FIGs. 32A-D, the (3) constructs showed comparable activity to the parent kp2_A03, and markedly higher IL-2 activity on the LAG3 expressing cells (filled squares) than on the LAG3 negative cells (open circles). The constructs were active at much lower concentrations on LAG3 positive cells than on LAG3 negative cells.

[0425] In another experiment, DBA constructs that incorporated variants in the DBA 7w2_A02 were tested on HEK-Blue™ IL-2 reporter cells and on HEK-Blue™ IL-2 reporter cells expressing LAG3. The variants comprised, for example, (i) M39L and S70A substitutions in VH and G28S in VL; (ii) M39L, S70A, and M125L substations in VH; or (iii) M39L, S70A, and M125L substations in VH and G28S in VL, relative to 7w2_A02c, per IMGT numbering. Assays were carried out using a HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb- il2). Each DBA construct was assayed on HEK-Blue™ IL-2 reporter cells (which do not express LAG3) and on HEK-Blue™ IL-2 reporter cells that had been transfected with a LAG3 expression vector using Lipofectamine 3000 (ThermoFisher cat# L3000001) according to the manufacturer’s instructions. Cell surface expression of LAG3 was confirmed by flow cytometry two days after transfection. In 384-well ELISA plates (Coming 3701) the antibody-cytokine complexes were serially diluted 1 :3 for 9 points in complete DMEM (+10% FBS, 2 mM L- glutamine, sodium pyruvate) from a starting concentration of 6 nM. HEK-Blue™ IL-2 reporter cells (12,500 cells) were added to each well and incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution (InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision. As shown in FIGs. 33A-D, the (3) constructs showed comparable activity to the parent 7w2_A02, and markedly higher IL-2 activity on the LAG3 expressing cells (filled squares) than on the LAG3 negative cells (open circles). The constmcts were active at much lower concentrations on LAG3 positive cells than on LAG3 negative cells.

[0426] Additional constructs were tested in similar assays. The LAG3-IL2 DBAs were expressed as heterodimeric antibodies with the DBA linked to an IL-2 variant on one arm and an scFv that binds LAG-3 on the other arm. The IL-2 variant included three mutations that reduce binding to CD25 and two mutations to facilitate production and characterization. When tested by SPR, CD25 did not bind to the variant at IpM CD25.

[0427] In this experiment, DBA constmcts that incorporated on DBAs kp2_A03_SS, 4bv_B01, or 7w2_A02, paired with the monospecific LAG3 antibodies Y37_cl0-c, Y37_c02b,or Y37_c05, were tested on HEK-Blue™ IL-2 reporter cells and on HEK-Blue™ IL-2 reporter cells expressing LAG3. Constructs with an isotype-matched control antibody in place of the DBA with either a control non-binding antibody or a LAG3 -targeting antibody were used as controls. Assays were carried out using a HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb-il2). Each DBA construct was assayed on HEK-Blue™ IL-2 reporter cells (which do not express LAG3) and on HEK-Blue™ IL-2 reporter cells that had been transfected with a LAG3 expression vector using Lipofectamine 3000 (ThermoFisher cat# L3000001) according to the manufacturer’s instructions. Cell surface expression of LAG3 was confirmed by flow cytometry two days after transfection. In 384-well ELISA plates (Coming 3701) the antibodycytokine complexes were serially diluted 1:4 for 8 points in complete DMEM (+10% FBS, 2 mM L-glutamine, sodium pyruvate) from a starting concentration of 15 nM. HEK-Blue™ IL-2 reporter cells (12,500 cells) were added to each well and incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution (InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision. The six constructs showed markedly higher IL-2 activity on the LAG3 expressing cells (filled squares) than on the LAG3 negative cells (open circles) as shown in FIGs. 34A-34F, while the control construct showed similar activity in the presence and absence of LAG3 (FIGs. 34G and 34H). The reversibly activatable constructs were active at much lower concentrations on LAG3 positive cells than on LAG3 negative cells.

[0428] Four common light chain LAG3 antibodies were used to produce reversibly activatable polypeptides of the present disclosure in the asymmetric format shown in FIG. 19F. LAG3 -dependent IL-2 activity was assayed by comparing the activity of the activatable polypeptides on HEK-Blue™ IL-2 reporter cells and on HEK-Blue™ IL-2 reporter cells expressing LAG3 A construct with an isotype anti-PSMA antibody in place on both arms of the construct was used as a control (with IL-2 attached to the N-terminus of one heavy chain). Assays were carried out using a HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb- il2). Each DBA construct was assayed on HEK-Blue™ IL-2 reporter cells (which do not express LAG3) and on HEK-Blue™ IL-2 reporter cells that had been transfected with a LAG3 expression vector using Lipofectamine 3000 (ThermoFisher cat# L3000001) according to the manufacturer’s instructions. Cell surface expression of LAG3 was confirmed by flow cytometry two days after transfection. In 384-well ELISA plates (Coming 3701) the antibody-cytokine complexes were serially diluted 1 :4 for 8 points in complete DMEM (+10% FBS, 2 mM L- glutamine, sodium pyruvate) from a starting concentration of 6 nM. HEK-Blue™ IL-2 reporter cells (12,500 cells) were added to each well and incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution(InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision. As shown in FIGs. 19A-D, the (4) constructs showed markedly higher IL-2 activity on the LAG3 expressing cells (solid lines) than on the LAG3 negative cells (dotted lines), while the control construct showed similar activity in the presence and absence of LAG3 (FIG. 19E).

[0429] DBA-based reversibly activatable polypeptide constructs were tested for the ability to induce STAT5 phosphorylation in primary human CD8+ T cells in the presence or absence of LAG-3 binding (STAT5 signaling can be downstream of IL-2R). Total T cells were isolated from human PBMCs using immunomagnetic negative selection (STEMCELL) and stimulated with plate-bound anti-CD3 and soluble anti-CD28 for 72 hours to induce LAG-3 expression, and rested in the absence of stimulation overnight. The stimulated T cells were incubated (e.g., pretreated) for 15 minutes with either a LAG-3 blocking antibody or an isotype control antibody. Next, titrating concentrations of cLAG3-IL2 or a non-conditional IL2-containing isotype control (e.g., lacking an anti-IL-2 binding specificity, thus constitutively active) were added, and cells were incubated at 37°C for 20 minutes. The T cells were fixed with Perm Buffer III (BD Biosciences), washed, and stained with antibodies directed against CD3, CD4, CD8, CD45RA, CD45RO, and pSTAT5. STAT5 phosphorylation within the T cell populations was assessed by flow cytometry. As shown in FIG. 20, in the CD8+CD45RO+ T cells, which are largely LAG-3 positive, the cLAG3-IL2 treatment induced an increased frequency of STAT5 phosphorylationpositive CD8+ T cells compared to the non-conditional IL-2-containing isotype control.However, CD8+CD45RO+ T cells that were pre-treated with a LAG-3 blocking antibody, and to which cLAG3-IL2 could not bind, had a lower frequency of STAT5 phosphorylation-positive cells following treatment with cLAG3-IL2 compared to the non-conditional IL-2-containing isotype control, which exhibited similar activity with or without anti-LAG3 pre-treatment Taken together, these data show that cLAG3-IL2 has decreased activity in the absence of LAG-3 binding and increased activity in the presence of LAG-3 binding compared to a non-conditional IL-2 control.

[0430] A similar STAT5 phosphorylation assay was performed for additional constructs, also using primary human CD8+ T cells. Total T cells were isolated from human peripheral blood mononuclear cells (PBMCs) using immunomagnetic negative selection (STEMCELL) and stimulated with plate-bound anti-CD3 and soluble anti-CD28 for 72 hours to induce LAG-3 expression. Stimulated T cells were incubated for 15 minutes at 37°C with either a LAG-3 blocking antibody or an isotype control antibody. Next, titrating concentrations of either cLAG3-IL2 constructs, non-conditional LAG-3 -targeted IL-2 (Rela_PSMA_IL2), or non- conditional non-targeted IL-2 (PSMA PSMA IL2) were added to the cells. The T cells werethen incubated at 37°C for 20 minutes, fixed with Perm Buffer III (BD Biosciences), washed, and stained with antibodies directed against CD3, CD4, CD8, CD45RA, CD45RO, and pSTAT5. STAT5 phosphorylation within the T cell populations was assessed by flow cytometry. In the CD8+CD45RO+ T cells, which are largely LAG-3 positive, the cLAG3-IL2 treatment induced an increased frequency of STAT5 phosphorylation-positive CD8+ T cells compared to the non-conditional non-targeted IL-2 control (TABLE 17). However, CD8+CD45RO+ T cells that were pre-treated with a LAG-3 blocking antibody, and to which cLAG3-IL2 could not bind, had a lower frequency of STAT5 phosphorylation-positive cells following treatment with cLAG3-IL2 compared to either of the non-conditional controls. Taken together, these data demonstrate the LAG-3 binding-dependent pSTAT5 activity of cLAG3-IL2 constructs compared to non-conditional IL-2 controls.

[0431] TABLE 17: STAT5 phosphorylation EC50 values for primary human CD8+ T cells after pre-treatment with a LAG-3 blocking antibody or isotype control antibody, followed by treatment with titrating concentrations of either cLAG3-IL2 constructs, non-conditional LAG-3 - targeted IL-2 (Rela_PSMA_IL2), or non-conditional non-targeted IL-2 (PSMA_PSMA_IL2).EXAMPLE 8: cLAG3-IL2 Inhibition of Tumor Growth as a Monotherapy in a Syngeneic Tumor Model

[0432] This example describes in vivo effects of reversibly activatable polypeptides of the disclosure (cLAG3-IL2) including inhibition of tumor growth in the MC38 syngeneic mouse tumor model, decreased toxicity compared to non-conditional LAG-3 -targeted IL-2, and effects on intratumoral CD8+ T cell activation. Two cLAG3-IL2 constructs (AF008825 and AF008451), a non-conditional LAG3-targeted IL-2 (AF008522), and an isotype control were tested (FIG. 21).

[0433] MC38 tumor cells were implanted subcutaneously in genetically modified knock-in mice (GenOway) that express chimeric versions of PD-1 and LAG-3 comprising the extracellular domains of human PD-1 and LAG-3 and the transmembrane / intracellular domains of mouse PD-1 and LAG-3 (hPD-l / hLAG-3 mice). Tumors were measured twice weekly, and volumes calculated as (Length x Width x Height / 2). When tumors reached a volume of ~75 mm3 on day 8 post-implantation, mice were separated into treatment groups and treated intravenously (IV) with cLAG3-IL2 constructs (5 mg / kg), an isotype control antibody (5 mg / kg), or the maximum tolerated dose of a non- conditional LAG-3 -targeted IL-2 (2.5 mg / kg). Mice received a second dose of treatments on day 11 post-implantation. Body weights were measured twice weekly. On day 13, a subset of mice was euthanized, and blood, spleens, and tumors were collected and a portion of each sample used for phenotypic characterization by flow cytometry using fluorophore-conjugated antibodies directed against cell surface and intracellular markers including NKp46, NK1.1, CD62L, CD44, CD19, TCRb, CD4, Ki-67, GrzB, TCF1 (BD Biosciences), CD8, CD45, Foxp3, fixable viability dye, (Thermo Fisher) and M8 tetramer (MBL Life Science). Cell numbers were quantitated using CountBright Absolute Counting Beads (Thermo Fisher). A second portion of the blood, spleens, and tumors samples collected on Day 13 were restimulated using Cell Activation Cocktail with Brefeldin A (Biolegend) for 4 hours at 37°C. Cells were then characterized for cytokine production by flow cytometry using fluorophore-conjugated antibodies directed against phenotypic and functional markers including NKp46, NK1.1, CD44, TCRb, CD4, GrzB, IFN-y, TNF-a (BD Biosciences), CD8, CD45, Foxp3, fixable viability dye, (Thermo Fisher) and M8 tetramer (stains cells with antigen-specific TCR; MBL Life Science).

[0434] Tumor growth inhibition. FIG. 22A shows average tumor size for the four treatment groups and FIG. 22B shows the same data plotted for the individual mice in each group. The conditional cLAG3-IL2 constructs inhibited tumor growth similarly to the nonconditional LAG-3 -targeted IL-2, with one tumor-free mouse in the cLAG3-IL2 AF008825 group and one tumor-free mouse in the non-conditional LAG-3 -targeted IL-2 group. These data demonstrate the ability of cLAG3-IL2 to inhibit tumor growth when administered as a monotherapy.

[0435] Weight loss and signaling / systemic expansion of LAG 3 negative cells. In contrast to the non-conditional LAG-3 -targeted IL-2 (AF008522), the cLAG3-IL2 constructs did not induce body weight loss (FIG. 23A), or increase circulating CD8+ T cell, NK cell, or Treg cell numbers (FIGs. 23B-D, respectively). These data demonstrate the decreased toxicity of cLAG3- IL2 compared to non-conditional LAG-3 -targeted IL-2 and the failure of cLAG3-IL2 to expand LAG-3 -negative IL-2R-positive immune cells in circulation.

[0436] Fluid accumulation in the lungs. A major side effect for patients treated with high- dose IL-2 is vascular leak syndrome, which leads to accumulation of intravascular fluid in organs such as lungs and liver. The treatment group that received the non-conditional LAG3- targeted IL-2 (AF008522) showed large increases in lung weight due to fluid accumulation, while lung weights for the cLAG3-IL2 groups were identical to the isotype control (FIG. 24). This data demonstrates that at a dose of 5 mg / kg treatment with cLAG3-IL2 does not lead to fluid accumulation in the lungs, indicating lower toxicity.

[0437] Immunomodulation. Treatment with cLAG3-IL2 constructs resulted in increased intratumoral CD8+ T cell and stem-like CD8+ T cell density (FIG. 25A), increased frequency of intratumoral Ki-67+ and GrzB+ cells amongst CD8+ T cells (FIG. 25B), an increase in the intratumoral CD8+ / Treg ratio (FIG. 25C), changes in the composition of the intratumoral CD8+ T cell population (FIG. 25D), and increased frequency of M8 tetramer-positive cells amongst CD8+ T cells in the spleen and tumor (FIG. 25E). These data demonstrate the ability of cLAG3- IL2 constructs to drive intratumoral CD8+ T cell activation and to induce the expansion of tumor antigen-specific CD8+ T cell clones.

[0438] Cytokine production by Tumor-Infiltrating CD8+ T Cells. Treatment with cLAG3- IL2 constructs resulted in an increased frequency of intratumoral IFN-y+ TNF-a+ cells amongst CD8+ T cells (FIG. 26A) and an increase in the proportion of intratumoral polyfunctional CD8+ T cells producing IFN-y, TNF-a, and GrzB (FIG. 26B). These data demonstrate the ability of cLAG3-IL2 constructs to drive effector function, effector cytokine production, and cytolytic activity by intratumoral CD8+ T cells.EXAMPLE 9: cLAG3-IL2 Induction of LAG-3 and PD-1 Expression on Intratumoral CD8+ T Cells

[0439] This example describes the ability of a reversibly activatable polypeptide (cLAG3- IL2) to induce the expression of LAG-3 and PD-1 on intratumoral CD8+ T cells in the MC38 syngeneic mouse tumor model. MC38 tumor cells were implanted subcutaneously in genetically modified knock-in mice (GenOway) that express chimeric versions of PD-1 and LAG-3 comprising the extracellular domains of human PD-1 and LAG-3 and the transmembrane / intracellular domains of mouse PD-1 and LAG-3 (hPD-l / hLAG-3 mice). Tumors were measured twice weekly, and volumes calculated as (Length * Width * Height / 2).

[0440] When tumors reached a volume of ~90 mm3 on day 9 post-implantation, mice were separated into treatment groups and treated intravenously (IV) with cLAG3-IL2 construct (5 mg / kg) or an isotype control antibody (5 mg / kg). Mice received a second dose of treatments on day 12 post-implantation. On day 14, a subset of mice was euthanized, and tumors werecollected for assessment of PD-1 and LAG-3 expression by flow cytometry using fluorophore- conjugated antibodies directed against human LAG-3 (Miltenyi Biotec), human PD-1 (inhouse), TCRb, CD4 (BD Biosciences), CD8, CD45, and fixable viability dye (Thermo Fisher).

[0441] Treatment with cLAG3-IL2 resulted in an increase in the frequency of PD-1 and LAG-3 double-positive cells amongst CD8+ T cells within the tumor (FIG. 27). These data demonstrate the ability of cLAG3-IL2 to drive intratumoral T cell activation and induce expression of LAG-3 and PD-1 on intratumoral T cells.EXAMPLE 10: Combination treatment with cLAG3-IL2 and anti-PD-1 drives inhibition of tumor growth

[0442] This example describes tumor growth inhibition by a reversibly activatable polypeptide (cLAG3-IL2) when administered in combination with anti-PD-1 in the MC38 syngeneic mouse tumor model. MC38 tumor cells were implanted subcutaneously in genetically modified knock-in mice (GenOway) that express chimeric versions of PD-1 and LAG-3 comprised of the extracellular domains of human PD-1 and LAG-3 and the transmembrane / intracellular domains of mouse PD-1 and LAG-3 (hPD-l / hLAG-3 mice). Tumors were measured twice weekly, and volumes calculated as (Length x Width x Height / 2). When tumors reached a volume of -130 mm3 on day 9 post-implantation, mice were separated into treatment groups and treated intravenously (IV) with cLAG3-IL2 (5 mg / kg), cLAG3-IL2 (5 mg / kg) + anti-PD-1 (10 mg / kg), anti-LAG-3 (5 mg / kg) + anti-PD-1 (10 mg / kg), non-conditional LAG-3 -targeted IL-2 (2 mg / kg) or an isotype control (5 mg / kg).

[0443] Average tumor volumes for each treatment group are plotted in FIG. 28A. Tumor volumes for individual mice in each treatment group are plotted in FIG. 28B. cLAG3-IL2 treatment in combination with anti-PD-1 resulted in increased tumor growth inhibition compared to cLAG3-IL2 monotherapy, anti-LAG-3 + anti-PD-1, non-conditional LAG-3 -targeted IL-2, or the isotype control. These data demonstrate the ability of cLAG3-IL2 to drive increased tumor growth inhibition in combination with anti-PD-1 treatment.EXAMPLE 11: cLAG3-IL2 avoidance of IL-2-mediated toxicity in vivo at high dose levels

[0444] This example describes the decreased toxicity of a reversibly activatable polypeptide (cLAG3-IL2) compared to non-conditional LAG3-targeted IL-2 when administered to nontumor bearing mice. Genetically modified knock-in mice (GenOway) expressing chimeric versions of PD-1 and LAG3 comprised of the extracellular domains of human PD-1 and LAG3 and the transmembrane / intracellular domains of mouse PD-1 and LAG3 (hPD-l / hLAG3 mice) were dosed intravenously (IV) with cLAG3-IL2 constructs (30 mg / kg), an isotype controlantibody (30 mg / kg), or the maximum tolerated dose of a non-conditional LAG3 -targeted IL-2 (2.5 mg / kg). Mice received a second dose 72 hours following the first. Body weight was monitored at least twice weekly (FIG. 29A). 5 days after the initial dose, mice were euthanized and lung weights (FIG. 29B) as well as NK and CD8+T cell counts in the blood (FIG. 29C) were assessed.

[0445] At 30 mg / kg, the cLAG3-IL2 induced less toxicity than the non-conditional LAG3- targeted IL-2 at 2.5 mg / kg. These data show that even at high doses, cLAG3-IL2 shows minimal toxicity in mice compared to a non-regulated IL-2 construct.EXAMPLE 12: Lack of IFN-y production by human PBMCs treated with CLAG3-IL2

[0446] This example describes the activity of cLAG3-IL2 on total human peripheral blood mononuclear cells (PBMCs). PBMCs were isolated from the whole human blood of six healthy donors and treated with titrating concentrations of cLAG3-IL2 constructs or a non-conditional IL-2 control. After 24 hours, supernatants were collected, and IFN-y concentration was assessed using an MSD Immunoassay (Meso Scale Diagnostics). The dose response of IFN-y induction is plotted in FIG. 30A as an average of all donors and the IFN-y release at lOOnM for individual donors is shown in FIG. 30B. The cLAG3-IL2 constructs induced minimal IFN-y production from the PBMCs compared to the non-conditional IL-2 control.EXAMPLE 13: Pharmacokinetics of cLAG3-IL2 in C57BL / 6 mice

[0447] This example describes the persistence of reversibly activatable polypeptides (cLAG3-IL2) in circulation in mice. C57BL / 6 mice were intravenously (IV) dosed with 1 mg / kg of cLAG3-IL2, non-conditional LAG3-targeted IL-2, or an PSMA-directed isotype control. At various time points mice were bleed, and serum was frozen at -80°C. Treatment concentrations in serum were quantified using an MSD capture immunoassay (Meso Scale Diagnostics). MSD plates were coated for 1 hour at room temperature with biotinylated LAG3 or biotinylated PSMA, washed, and then incubated with diluted serum samples for 1 hour at room temperature. Plates were washed again, and the captured treatments were detected using an anti-mouse IgG2a SULFO-TAG antibody. Plates were washed again, MSD GOLD Read Buffer B was added to each well, and the ECL signal was read using a MESO SECTOR S plate reader (Meso Scale Diagnostics) As shown in FIG. 31, the cLAG3-IL2 constructs demonstrate pharmacokinetic properties similar to those of the anti-PSMA isotype control antibody.EXAMPLE 14: cLAG3-IL2 inhibition of tumor growth in a syngeneic tumor model

[0448] This example describes cLAG3-IL2 inhibition of tumor growth in the MC38 syngeneic mouse tumor model. MC38 tumor cells were implanted subcutaneously in genetically modified knock-in mice (Biocytogen) that express a chimeric LAG-3 comprising the extracellular domain of human LAG-3 and the transmembrane / intracellular domain of mouse LAG-3 (hLAG-3 mice). Tumors were measured twice weekly, and volumes calculated as (Length x Width x Height / 2). When tumors reached a volume of ~ 125 mm3on day 8 postimplantation, mice were separated into treatment groups and treated intravenously (IV) with cLAG3-IL2 (0.5 mg / kg), cLAG3-IL2 (0.5 mg / kg) in combination with anti-mouse PD-1 (clone 29F.1A12, Bio X Cell; 10 mg / kg), anti-human LAG-3 (10 mg / kg) in combination anti-mouse PD-1 (10 mg / kg), an isotype control antibody (PSMA; 10 mg / kg), or a non-conditional nontargeted IL-2 (1 mg / kg). Mice received a second dose of treatments on day 11 post-implantation.

[0449] FIG. 36 shows average tumor size for the treatment groups over time. The isotype control antibody, the anti-PD-1 antibody, and the non-conditional non-targeted IL-2 showed little to no tumor growth inhibition. The anti-LAG-3 and anti-PD-1 combination showed intermediate tumor growth inhibition with 2 / 9 mice tumor-free. cLAG3-IL2 administered as a monotherapy demonstrated superior tumor growth inhibition with 5 / 9 mice tumor-free. cLAG3- IL2 activity was further enhanced when combined with anti-PD-1, with 8 / 9 mice tumor-free. These data demonstrate the ability of cLAG3-IL2 to inhibit tumor growth when administered as a monotherapy and enhancement of cLAG3-IL2 activity when combined with PD-1 blockade.EXAMPLE 15: IL-2 variants

[0450] This example describes the binding properties and activity of IL-2 variants. The variants were designed to reduce the binding of the IL-2 to the IL2Ralpha without altering binding to the IL2Rbeta / gamma.

[0451] The variants were tested by surface plasmon resonance (SPR) for binding to IL- 2Ralpha (Aero ILA-H52H9) and IL-2Rbeta / gamma (Aero ILG-H5283). In a separate experiment, the variants were tested for activity using HEK-Blue™ IL-2 reporter cells (InvivoGen catalog code hkb-il2). In 384-well ELISA plates (Corning 3701) the IL-2 variants were diluted in complete DMEM (+10% FBS, 2 mM L-glutamine, sodium pyruvate). HEK- Blue™ IL-2 reporter cells (12,500 cells) were added to each well and incubated overnight. Five microliters from each well were transferred to a new plate containing 45 microliters of QuantiBlue solution (InvivoGen Product # rep-qbs). After 30 to 60 minutes the absorbance at 630 nm was determined using a Perkin-Elmer Envision.

[0452] In a first example, IL-2 variants were produced as monovalent IL-2 by expression as heterodimeric antibodies with the IL-2 variant linked to Fab on one arm and an scFv on the other arm, where neither of the arms have any antigen-binding affinity for IL-2. Several variants exhibited reduced IL2Ra binding while maintaining IL-2Rbeta / gamma binding and activity in an IL-2 reporter assay (TABLE 18).

[0453] TABLE 18: Fab-linked IL-2 variants binding to IL-2Ra, binding to IL-2Rbg, and activity in IL-2 reporter cell assay.

[0454] In a second example, IL-2 variants were linked to an IL-2-binding DBA on the Fab arm with no scFv on the second arm. The IL-2 variants were produced by a mammalian expression system and purified using standard methods.

[0455] TABLE 19: DBA-linked IL-2 variants binding to IL-2Ra

[0456] This example demonstrates a set of IL-2 variants that reduce or eliminate the binding of IL-2 to the IL2Ralpha which can be used in reversibly activatable polypeptides.EXAMPLE 16: Stability testing

[0457] This example describes testing stability of reversibly activatable polypeptides of the present disclosure. In this example, conditions of stress are used to simulate conditions that may be present during manufacture, storage, and use of a protein therapeutic.

[0458] In one example, the LAG3-IL2 DBAs are expressed as standard antibodies using mammalian expression systems and standard purification methods. The antibodies are then be placed in stress conditions. These conditions can include acidic or basic buffers, oxidative conditions, elevated temperature, extremes of ionic strength, and / or freeze / thaw cycles. For example, the antibodies are exchanged into buffers ranging in pH from pH 3 to pH 10 and then incubated at 37°C for Ihr to one week. Control antibodies are kept in stable formulation buffer at 4°C for the matched period. Resulting samples are analyzed for chemical changes by mass spectroscopy using standard procedures. For example, the conversion of an amide group to a carboxylic acid group in deamidation results in 0.984 Da mass shift that can be detected by MS methods. The antibodies are also tested for changes in physical properties. For example, the antibodies are tested for the presence of aggregations using methods such as size exclusion chromatography, dynamic light scattering, or the binding of fluorescent dyes. The antibodies are tested for a decrease in function, such as binding, using methods such as surface plasmon resonance (SPR), enzyme-linked immunoassays, and / or binding to cells using flow cytometry.

[0459] In a second example, the LAG3-IL2 DBAs are expressed as heterodimeric antibodies with the DBA linked to an IL-2 variant on one arm and an scFv targeting LAG-3 on the other arm (e.g., as shown in FIG. 9, FIG. 17H & FIG. 181). The constructs are subjected to stress conditions as above. In addition to the testing described above, the constructs are tested for the regulated function. For example, the constructs can be tested using HEK-Blue™ FL-2 reporter cells and HEK-Blue™ IL-2 reporter cells that have been transfected to express LAG3 (InvivoGen catalog code hkb-il2), as described herein or using standard procedures.

[0460] Antibodies and constructs are evaluated for suitable and / or improved stability properties.VI. SEQUENCES OF REVERSIBLY ACTIVATABLE POLYPEPTIDES, ANTIBODIES, AND ANTIGEN-BINDING FRAGMENTS THEREOF

[0461] A number of naming conventions are used throughout this document. The general class of reversibly activated polypeptides of the present invention are referred to as “cLAG3- IL2” constructs. Every antibody can be assigned a seven character name of the form (3) characters, underscore, (3) characters (e.g., kp2_A03). In some cases, the same antibody may be referred to with a descriptor such as “LAG3_” appended to the beginning or end of the name(e.g., LAG3_kp2_A03 or kp2_A03 are the same antibody). In some cases, antibody names contain additional letters appended to the end, such as kp2_A03c or Y37-c02c, which can be the same as kp2_A03 and Y37-c02, respectively. Additional letters in some embodiments indicate mutations, such as kp2_A03_N59S or kp2 A03 SS, which can refer to variants of the kp2 A03 DBA. Symmetric antibodies with Fab or scFv arms can be given an antibody number in the form ABOOnnnn (e.g., AB004668). Asymmetric constructs (i.e., antibodies with a heterodimeric Fc) can be given a construct number in the form AFOOnnnn. Some construct numbers refer to a reversibly activated polypeptide of the present invention (e.g., AF008451, which has LAG3 (e.g., monospecific) scFv on one arm and a LAG3-IL2 DBA Fab linked to IL-2 through the heavy chain on the other arm, see FIG. 21). Some construct numbers refer to unregulated control constructs (e.g., AF008522 which has LAG3 scFv on one arm and a PSMA antibody linked to IL-2 through the heavy chain on the other arm, see FIG. 21).

[0462] In some embodiments, 7w2_A02c, 7w2_C01, 7w2_Bl lc, 7w2_C06c, and sn6_Cl lc (e.g., and variants thereof) are preferred dual-binding antibodies (DBAs) disclosed herein. In some embodiments, kp2_A03, 4bv_B01 and 7w2_C06 (e.g., and variants thereof) are preferred DBAs.

[0463] TABLE 20 provides sequences of heavy chain variable domains (VH / HV) of LAG-3 binding antibodies and a control anti-PSMA antibody.-Ill-

[0464] TABLE 21 provides light chain variable domains (VL / LV) of LAG-3 binding antibodies and a control anti-PSMA antibody, e g., corresponding to the heavy chain variable domains in TABLE 20.

[0465] TABLE 22 provides heavy chain variable domains (VH / HV) of illustrative LAG3 / IL2 dual -binding antibodies disclosed herein.

[0466] TABLE 23 provides light chain variable domains (VL / LV) of illustrative LAG3 / IL2 dual-binding antibodies disclosed herein, for example, corresponding to the heavy chain variable domains in TABLE 22.

[0467] TABLE 24 provides CDR sequences of illustrative LAG3 -binding antibodies disclosed herein.

[0468] TABLE 25 provides CDR sequences of illustrative LAG3 / IL2 dual-binding antibodies disclosed herein.

[0469] TABLE 26 provides an index of combinations of sequences used in illustrative constructs in disclosed herein (e.g., reversibly activatable polypeptides and control constructs). The illustrative constructs use human or mouse IgG heavy chain constant domains and IgK light chain constant domains.

[0470] TABLE 27 provides the sequences of components of constructs disclosed herein (e.g., heavy chains and light chains of reversibly activatable polypeptides and control constructs)VII. ADDITIONAL SEQUENCES AND ASPECTS

[0471] A reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) disclosed herein can comprise, consist essentially of, or consist of an amino acid sequence with at least about 70%, at least about 71%, at least about 72%, at least about 73%, at least about 74%, at least about 75%, at least about 76%, at least about 77%, at least about 78%, at least about 79%, at least about 80%, at least about 81%, at least about 82%, at least about 83%, at least about 84%, at least about 85%, at least about 86%, at least about 87%, at least about 88%, at least about 89%, at least about 90%, at least about 91%, at least about 92%, at least about 93%, at least about 94%, at least about 95%, at least about 95.5%, at least about 96%, at least about 96.5%, at least about 97%, at least about 97.5%, at least about 98%, at least about 98.5%, at least about 99%, at least about 99.5%, or about 100% sequence identity or sequence similarity to any one of SEQ ID NOs: 1-1894.

[0472] A reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domaindisclosed herein, or combination thereof) disclosed herein can comprise, consist essentially of, or consist of an amino acid sequence with at least about 70%, at least about 71%, at least about 72%, at least about 73%, at least about 74%, at least about 75%, at least about 76%, at least about 77%, at least about 78%, at least about 79%, at least about 80%, at least about 81%, at least about 82%, at least about 83%, at least about 84%, at least about 85%, at least about 86%, at least about 87%, at least about 88%, at least about 89%, at least about 90%, at least about 91%, at least about 92%, at least about 93%, at least about 94%, at least about 95%, at least about 95.5%, at least about 96%, at least about 96.5%, at least about 97%, at least about 97.5%, at least about 98%, at least about 98.5%, at least about 99%, at least about 99.5%, or about 100% sequence identity or sequence similarity to at least 50 consecutive amino acids of any one of SEQ ID NOs: 1-1894

[0473] A reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) disclosed herein can comprise, consist essentially of, or consist of an amino acid sequence with at least about 70%, at least about 71%, at least about 72%, at least about 73%, at least about 74%, at least about 75%, at least about 76%, at least about 77%, at least about 78%, at least about 79%, at least about 80%, at least about 81%, at least about 82%, at least about 83%, at least about 84%, at least about 85%, at least about 86%, at least about 87%, at least about 88%, at least about 89%, at least about 90%, at least about 91%, at least about 92%, at least about 93%, at least about 94%, at least about 95%, at least about 95.5%, at least about 96%, at least about 96.5%, at least about 97%, at least about 97.5%, at least about 98%, at least about 98.5%, at least about 99%, at least about 99.5%, or about 100% sequence identity or sequence similarity to at least 100 consecutive amino acids of any one of SEQ ID NOs: 1-1894

[0474] A reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) disclosed herein can comprise, consist essentially of, or consist of an amino acid sequence with at most about 70%, at most about 71%, at most about 72%, at most about 73%, at most about 74%, at most about 75%, at most about 76%, at most about 77%, at most about 78%, at most about 79%, at most about 80%, at most about 81%, at most about 82%, at most about 83%, at most about 84%, at most about 85%, at most about 86%, at most about 87%, at most about 88%, at most about 89%, at most about 90%, at most about 91%, at most about 92%, at most about 93%, at most about 94%, at most about 95%, at most about 95.5%, at most about 96%, at most about 96.5%, at most about 97%, at most about 97.5%,at most about 98%, at most about 98.5%, at most about 99%, or at most about 99.5% sequence identity or sequence similarity to any one of SEQ ID NOs: 1-1894.

[0475] In some embodiments, a reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises, consists essentially of, or consists of an amino acid sequence with about 70%, about 71%, about 72%, about 73%, about 74%, about 75%, about 76%, about 77%, about 78%, about 79%, about 80%, about 81%, about82%, about 83%, about 84%, about 85%, about 86%, about 87%, about 88%, about 89%, about90%, about 91%, about 92%, about 93%, about 94%, about 95%, about 95.5%, about 96%, about 96.5%, about 97%, about 97.5%, about 98%, about 98.5%, about 99%, or about 99.5% or about 100% sequence identity or sequence similarity to any one of SEQ ID NOs: 1-1894.

[0476] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises, consists essentially of, or consists of the amino acid sequence of any one of SEQ ID NOs: 1-1894.

[0477] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises an amino acid sequence with one or more insertions, deletions, and / or substitutions relative to any one of SEQ ID NOs: 1-1894

[0478] For example, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) can comprise an amino acid sequence with at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, at least 20, at least 25, or at least 30 amino acid insertions relative to any one of SEQ ID NOs: 1-1894

[0479] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises an amino acid sequence with at most 1, at most 2, at most 3, at most 4, at most 5, at most 6, at most 7, at most 8, at most 9, at most 10, at most 11, at most 12, at most 13, at most 14, at most 15, at most 16, at most 17, at most 18, at most 19, at most 20, at most 25, at most 30, at most 35, at most 40, at most 45, or at most 50 amino acid insertions relative to any one of SEQ ID NOs: 1-1894.

[0480] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or 50 amino acid insertions relative to any one of SEQ ID NOs: 1-1894.

[0481] The one or more insertions can be at the N-terminus, the C-terminus, within the amino acid sequence, or a combination thereof. The one or more insertions can be contiguous, non-contiguous, or a combination thereof.

[0482] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises an amino acid sequence with at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, at least 20, at least 25, or at least 30 amino acid deletions relative to any one of SEQ ID NOs: 1-1894.

[0483] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises an amino acid sequence with at most 1, at most 2, at most 3, at most 4, at most 5, at most 6, at most 7, at most 8, at most 9, at most 10, at most 11, at most 12, at most 13, at most 14, at most 15, at most 16, at most 17, at most 18, at most 19, at most 20, at most 25, at most 30, at most 35, at most 40, at most 45, or at most 50 amino acid deletions relative to any one of SEQ ID NOs: 1-1894.

[0484] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or 50 amino acid deletions relative to any one of SEQ ID NOs: 1-1894.

[0485] The one or more deletions can be at the N-terminus, the C-terminus, within the amino acid sequence, or a combination thereof. The one or more deletions can be contiguous, non-contiguous, or a combination thereof.

[0486] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises an amino acid sequence with at least 1, at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least17, at least 18, at least 19, at least 20, at least 25, or at least 30 amino acid substitutions relative to any one of SEQ ID NOs: 1-1894.

[0487] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises an amino acid sequence with at most 1, at most 2, at most 3, at most 4, at most 5, at most 6, at most 7, at most 8, at most 9, at most 10, at most 11, at most 12, at most 13, at most 14, at most 15, at most 16, at most 17, at most 18, at most 19, at most 20, at most 25, at most 30, at most 35, at most 40, at most 45, or at most 50 amino acid substitutions relative to any one of SEQ ID NOs: 1-1894.

[0488] In some embodiments, the reversibly activatable polypeptide or a domain thereof (e.g., an effector domain, dual-binding detection domain, immunoglobulin constant domain, linker, other domain disclosed herein, or combination thereof) comprises 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, or 50 amino acid substitutions relative to any one of SEQ ID NOs: 1-1894.

[0489] The one or more substitutions can be at the N-terminus, the C-terminus, within the amino acid sequence, or a combination thereof. The one or more substitutions can be contiguous, non-contiguous, or a combination thereof.

[0490] The degree of sequence identity between two sequences can be determined, for example, by comparing the two sequences using computer programs designed for this purpose, such as global or local alignment algorithms. Non-limiting examples include BLASTp, BLASTn, Clustal W, MAFFT, Clustal Omega, AlignMe, Praline, GAP, BESTFIT, Needle (EMBOSS), Stretcher (EMBOSS), GGEARCH2SEQ, Water (EMBOSS), Matcher (EMBOSS), LALIGN, SSEARCH2SEQ, or another suitable method or algorithm. A global alignment algorithm, such as a Needleman and Wunsch algorithm, can be used to align two sequences over their entire length, maximizing the number of matches and minimizes the number of gaps. Default settings can be used.

[0491] To generate similarity scores for two amino acid sequences, scoring matrices can be used that assign positive scores for some non-identical amino acids (e.g., amino acids with similar physio-chemical properties and / or amino acids that exhibit frequent substitutions in orthologs, homologs, or paralogs), Non -limiting examples of scoring matrices include PAM30, PAM70, PAM250, BLOSUM45, BLOSUM50, BLOUM62, BLOSUM80, and BLOSUM90.

[0492] Amino acids can include genetically encoded and non-genetically encoded occurring amino acids. Amino acids can include naturally occurring and non-naturally occurring amino acids. Amino acids can be L forms or D forms. Substitutions disclosed herein can include conservative and / or non-conservative amino acid substitutions. A conservative amino acidsubstitution can be a substitution of one amino acid for another amino acid of similar biochemical properties (e.g., charge, size, and / or hydrophobicity). A non-conservative amino acid substitution can be a substitution of one amino acid for another amino acid with different biochemical properties (e.g., charge, size, and / or hydrophobicity). A conservative amino acid change can be, for example, a substitution that has minimal effect on the secondary or tertiary structure of a polypeptide. A conservative amino acid change can be an amino acid change from one hydrophilic amino acid to another hydrophilic amino acid. Hydrophilic amino acids can include Thr (T), Ser (S), His (H), Glu (E), Asn (N), Gin (Q), Asp (D), Lys (K) and Arg (R). A conservative amino acid change can be an amino acid change from one hydrophobic amino acid to another hydrophilic amino acid. Hydrophobic amino acids can include He (I), Phe (F), Vai (V), Leu (L), Trp (W), Met (M), Ala (A), Gly (G), Tyr (Y), and Pro (P). A conservative amino acid change can be an amino acid change from one acidic amino acid to another acidic amino acid. Acidic amino acids can include Glu (E) and Asp (D). A conservative amino acid change can be an amino acid change from one basic amino acid to another basic amino acid. Basic amino acids can include His (H), Arg (R) and Lys (K). A conservative amino acid change can be an amino acid change from one polar amino acid to another polar amino acid. Polar amino acids can include Asn (N), Gin (Q), Ser (S) and Thr (T). A conservative amino acid change can be an amino acid change from one nonpolar amino acid to another nonpolar amino acid. Nonpolar amino acids can include Leu (...

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A reversibly activatable polypeptide comprising:(a) an effector domain; and(b) a dual -binding detection domain that is configured to bind the effector domain and an inducing moiety; wherein:(i) in an absence of the inducing moiety, the dual-binding detection domain binds to the effector domain and the effector domain is sterically hindered from binding to an effector partner; and(ii) upon contacting the reversibly activatable polypeptide to the inducing moiety, the dual-binding detection domain binds to the inducing moiety, the dualbinding detection domain is sterically hindered from binding to the effector domain, and the effector domain is activated; wherein the inducing moiety comprises LAG-3.

2. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a dual -binding antibody or antigen-binding fragment thereof comprising a heavy chain complementarity determining region (HCDR)3 with the amino acid sequence of SEQ ID NO: 1663.

3. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a dual -binding antibody or antigen-binding fragment thereof comprising:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662; and(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663.

4. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a dual -binding antibody or antigen-binding fragment thereof comprising:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662;(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663;(d) a light chain complementarity determining region (LCDR)l with the amino acid sequence of SEQ ID NO: 1664;(e) an LCDR2 with the amino acid sequence of SEQ ID NO: 1665 (AAS); and(f) an LCDR3 with the amino acid sequence of SEQ ID NO: 1666.

5. The reversibly activatable polypeptide of claim 1, wherein the effector domain is an IL-2 receptor agonist.

6. The reversibly activatable polypeptide of claim 1, wherein when activated, the effector domain is not sterically hindered from binding to the effector partner.

7. The reversibly activatable polypeptide of claim 1, wherein the effector domain comprises IL-2.

8. The reversibly activatable polypeptide of claim 1, wherein the effector domain comprises IL-15.

9. The reversibly activatable polypeptide of claim 1, wherein the effector domain comprises at least 80% sequence identity to any one of SEQ ID NOs: 1-4 and 79-91.

10. The reversibly activatable polypeptide of claim 1, wherein the effector domain comprises at least 95% sequence identity to any one of SEQ ID NOs: 1-4 and 79-91.

11. The reversibly activatable polypeptide of claim 1, wherein the effector domain comprises at least 98% sequence identity to any one of SEQ ID NOs: 1-4 and 79-91.

12. The reversibly activatable polypeptide of claim 1, wherein the effector domain comprises the amino acid sequence of any one of SEQ ID NOs: 1-4 and 79-91.

13. The reversibly activatable polypeptide of claim 1, wherein the effector partner comprises an IL-2 receptor.

14. The reversibly activatable polypeptide of claim 1, wherein the effector partner comprises IL-2Rp.

15. The reversibly activatable polypeptide of claim 1, wherein the effector partner comprises IL-2Ry.

16. The reversibly activatable polypeptide of claim 1, wherein the effector partner comprises IL-2RP and rL-2Ry.

17. The reversibly activatable polypeptide of claim 1, wherein the effector partner comprises IL-2Ra, IL-2RP and fL-2Ry.

18. The reversibly activatable polypeptide of claim 1, wherein the dual-binding detection domain binds to a domain or epitope of IL-2 that is an fL-2RPy binding site.

19. The reversibly activatable polypeptide of claim 18, wherein the dual-binding detection domain does not bind to a domain or epitope of IL-2 that is an IL-2Ra binding site.

20. The reversibly activatable polypeptide of claim 19, wherein the effector domain comprises an IL-2 mutein with modification(s) that reduce or eliminate binding to IL- 2Ra.

21. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a dual -binding antibody or antigen-binding fragment thereof.

22. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises an HCDR3 sequence disclosed in TABLE 25.

23. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 25.

24. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 25.

25. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a heavy chain variable domain (VH) and a light chain variable domain (VL).

26. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 282.

27. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a VH amino acid sequence with at least 95% sequence identity toSEQ ID NO: 262 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 282.

28. The reversibly activatable polypeptide of claim 1, wherein the dual-binding detection domain comprises a VH with the amino acid sequence of SEQ ID NO: 262 and a VL with the amino acid sequence of SEQ ID NO: 282.

29. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLES 22 & 23.

30. The reversibly activatable polypeptide of claim 1, wherein the dual-binding detection domain comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLES 22 & 23.

31. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLES 22 & 23.

32. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises, consists essentially of, or consists of a Fab.

33. The reversibly activatable polypeptide of claim 1, wherein the dual-binding detection domain comprises, consists essentially of, or consists of an scFv.

34. The reversibly activatable polypeptide of claim 1, wherein the dual -binding detection domain comprises, consists essentially of, or consists of a single domain antibody.

35. The reversibly activatable polypeptide of claim 1, wherein the effector domain is activated when the reversibly activatable polypeptide is contacted to a LAG-3+ cell.

36. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide induces IL-2R signaling, expansion, or activation of LAG-3+ IL-2RPy+ or LAG-3+ IL-2Rupy+ cells.

37. The reversibly activatable polypeptide of claim 1, wherein the effector domain is inactive when the reversibly activatable polypeptide is contacted to a LAG-3 -negative (LAG-3-) cell.

38. The reversibly activatable polypeptide of claim 1, wherein the effector domain is inactive when the reversibly activatable polypeptide is contacted to a LAG-3- JL-2RPy+ cell.

39. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide does not induce or substantially does not induce expansion or activation of LAG-3- IL-2RPy+ cells.

40. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide induces at least 10-fold higher IL-2 signaling by LAG-3+ IL-2R y+ or LAG-3 + fL-2RaPy+ cells as compared to LAG-3- IL-2RPy l or LAG-3- ZL-2R(XPY+ cells.

41. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide is a LAG-3 antagonist.

42. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide inhibits LAG-3 activity and induces IL-2R signaling.

43. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide further comprises a monospecific antibody or antigen-binding fragment thereof that binds to LAG-3.

44. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises an HCDR3 with the amino acid sequence of SEQ ID NO: 817.

45. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816; and(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817.

46. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816;(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817;(d) an LCDR1 with the amino acid sequence of SEQ ID NO: 818;(e) an LCDR2 with the amino acid sequence of SEQ ID NO: 819 (DVS); and(f) an LCDR3 with the amino acid sequence of SEQ ID NO: 820.

47. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises an HCDR3 sequence disclosed in TABLE 24.

48. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 24.

49. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 24.

50. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 214.

51. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 214.

52. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a VH with the amino acid sequence of SEQ ID NO: 137 and a VL with the amino acid sequence of SEQ ID NO: 214.

53. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a VH and VL pair, wherein the VH andthe VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLES 20 & 21.

54. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLES 20 & 21.

55. The reversibly activatable polypeptide of claim 43, wherein the monospecific antibody or antigen-binding fragment thereof comprises a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLES 20 & 21.

56. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises an Fc domain.

57. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises an IgG isotype antibody.

58. The reversibly activatable polypeptide of claim 57, wherein the dual-binding detection domain is a Fab of the IgG isotype antibody.

59. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises a first heavy chain and a first light chain, wherein the first heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the effector domain, a linker, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the first light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the VL and a CL.

60. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises a first heavy chain and a first light chain, wherein the first heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the first light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the effector domain, a linker, the VL, and a CL.

61. The reversibly activatable polypeptide of claim 59, wherein the reversibly activatable polypeptide further comprises a second heavy chain.

62. The reversibly activatable polypeptide of claim 61, wherein the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, an scFv, optionally CHI, hinge, CH2, CH3, and optionally CH4.

63. The reversibly activatable polypeptide of claim 61, wherein the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, (i) a single domain antibody, VHH, or nanobody, (ii) optionally CHI, (iii) hinge, (iv) CH2, (v) CH3, and (vi) optionally CH4.

64. The reversibly activatable polypeptide of claim 59, wherein the reversibly activatable polypeptide further comprises a second heavy chain and a second light chain.

65. The reversibly activatable polypeptide of claim 64, wherein the second heavy chain comprises a VH and the second light chain comprises a VL.

66. The reversibly activatable polypeptide of claim 59, wherein the reversibly activatable polypeptide is heterodimeric and the first heavy chain, and the second heavy chain each comprise a modification that promotes heterodimerization of the first heavy chain with the second heavy chain.

67. The reversibly activatable polypeptide of claim 64, wherein the VH of the second heavy chain and the VL of the second light chain form a second dual-binding detection domain that is configured to bind the effector domain and the inducing moiety.

68. The reversibly activatable polypeptide of claim 64, wherein the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the effector domain, a linker, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the second light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the VL and a CL.

69. The reversibly activatable polypeptide of claim 64, wherein the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH, CHI, hinge, CH2, CH3, and optionally CH4, and the second light chain comprises, consistsessentially of, or consists of, from N-to-C terminus, the effector domain, a linker, the VL, and a CL.

70. The reversibly activatable polypeptide of claim 65, wherein the VH of the second heavy chain and the VL of the second light chain form an antigen-binding fragment, optionally wherein the antigen-binding fragment is monospecific.

71. The reversibly activatable polypeptide of claim 65, wherein the second heavy chain comprises, consists essentially of, or consists of, from N-to-C-terminus, the VH of the second heavy chain, CHI, hinge, CH2, CH3, and optionally CH4, and the second light chain comprises, consists essentially of, or consists of, from N-to-C terminus, the VL of the second light chain and a CL.

72. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises a set of polypeptide chains, wherein each of the polypeptide chains comprises an amino acid sequence with at least 80% sequence identity to a polypeptide chain disclosed in TABLE 27.

73. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises a set of polypeptide chains, wherein each of the polypeptide chains comprises an amino acid sequence with at least 90% sequence identity to a polypeptide chain disclosed in TABLE 27.

74. The reversibly activatable polypeptide of claim 1, wherein the reversibly activatable polypeptide comprises a set of polypeptide chains disclosed in TABLE 26, wherein each of the polypeptide chains comprises the amino acid sequence of a polypeptide chain disclosed in TABLE 27.

75. A polynucleotide comprising a nucleic acid sequence that encodes the reversibly activatable polypeptide of claim 1.

76. A vector comprising the polynucleotide of claim 75.

77. The vector of claim 76, wherein the vector is a non-viral vector.

78. The vector of claim 76, wherein the vector is a viral vector.

79. A pharmaceutical composition comprising the reversibly activatable polypeptide of claim 1 and a pharmaceutically-acceptable excipient, vehicle, carrier, or diluent.

80. A method of selectively inducing signaling by an effector partner, the method comprising contacting a population of cells with the reversibly activatable polypeptide of claim 1, wherein signaling by the effector partner is preferentially induced in cells with surface expression of the inducing moiety.

81. A method of selectively inducing signaling by an effector partner, the method comprising administering the reversibly activatable polypeptide of claim 1 to the subject.

82. A method of treating a condition in a subject in need thereof, the method comprising administering the reversibly activatable polypeptide of claim 1 to the subject.

83. The method of claim 82, wherein the condition is a cancer.

84. The method of claim 83, wherein the reversibly activatable polypeptide drives expansion and activation of tumor-specific T cells.

85. The method of claim 83, wherein the reversibly activatable polypeptide inhibits tumor growth.

86. The method of claim 82, wherein toxicity is reduced as compared to a control compound for which the effector domain is active in the absence of the inducing moiety.

87. The method of claim 82, wherein toxicity is reduced as compared to a control compound for which the dual -binding detection domain does not bind to the effector domain.

88. A combination therapy comprising the reversibly activatable polypeptide of claim 1 and an immune checkpoint inhibitor.

89. The combination therapy of claim 88, wherein the immune checkpoint inhibitor is a PD1 inhibitor.

90. The combination therapy of claim 88, wherein the immune checkpoint inhibitor is a PDL1 inhibitor.

91. A dual-binding antibody or antigen-binding fragment thereof comprising an HCDR3 with the amino acid sequence of SEQ ID NO: 1663.

92. A dual-binding antibody or antigen-binding fragment thereof comprising:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662; and(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663.

93. A dual-binding antibody or antigen-binding fragment thereof comprising:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 1661;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 1662;(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 1663;(d) a light chain complementarity determining region (LCDR)1 with the amino acid sequence of SEQ ID NO: 1664;(e) an LCDR2 with the amino acid sequence of SEQ ID NO: 1665 (AAS); and(f) an LCDR3 with the amino acid sequence of SEQ ID NO: 1666.

94. A dual-binding antibody or antigen-binding fragment thereof comprising a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 282.

95. A dual-binding antibody or antigen-binding fragment thereof comprising a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 262 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 282.

96. A dual-binding antibody or antigen-binding fragment thereof comprising a VH with the amino acid sequence of SEQ ID NO: 262 and a VL with the amino acid sequence of SEQ ID NO: 28297. A dual-binding antibody or antigen-binding fragment thereof comprising an HCDR3 sequence disclosed in TABLE 25.

98. The dual -binding antibody or antigen-binding fragment thereof of claim 97, wherein the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

99. A dual-binding antibody or antigen-binding fragment thereof comprising a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 25.

100. The dual-binding antibody or antigen-binding fragment thereof of claim 99, wherein the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

101. A dual-binding antibody or antigen-binding fragment thereof comprising a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 25.

102. The dual-binding antibody or antigen-binding fragment thereof of claim 101, wherein the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

103. A dual-binding antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLES 22 & 23.

104. The dual-binding antibody or antigen-binding fragment thereof of claim 103, wherein the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

105. A dual-binding antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLES 22 & 23.

106. The dual-binding antibody or antigen-binding fragment thereof of claim 105, wherein the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

107. A dual-binding antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLES 22 & 23.

108. The dual -binding antibody or antigen-binding fragment thereof of claim 107, wherein the dual -binding antibody or antigen-binding fragment thereof binds to IL-2 and LAG3.

109. An antibody or antigen-binding fragment thereof that binds to LAG-3, comprising an HCDR3 with the amino acid sequence of SEQ ID NO: 817.

110. An antibody or antigen-binding fragment thereof that binds to LAG-3, comprising:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816; and(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817.

111. An antibody or antigen-binding fragment thereof that binds to LAG-3, comprising:(a) an HCDR1 with the amino acid sequence of SEQ ID NO: 815;(b) an HCDR2 with the amino acid sequence of SEQ ID NO: 816;(c) an HCDR3 with the amino acid sequence of SEQ ID NO: 817;(d) an LCDR1 with the amino acid sequence of SEQ ID NO: 818;(e) an LCDR2 with the amino acid sequence of SEQ ID NO: 819 (DVS); and(f) an LCDR3 with the amino acid sequence of SEQ ID NO: 820.

112. An antibody or antigen-binding fragment thereof that binds to LAG-3, comprising a VH amino acid sequence with at least 90% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 90% sequence identity to SEQ ID NO: 214.

113. An antibody or antigen-binding fragment thereof that binds to LAG-3, comprising a VH amino acid sequence with at least 95% sequence identity to SEQ ID NO: 137 and a VL amino acid sequence with at least 95% sequence identity to SEQ ID NO: 214.

114. An antibody or antigen-binding fragment thereof that binds to LAG-3, comprising a VH with the amino acid sequence of SEQ ID NO: 137 and a VL with the amino acid sequence of SEQ ID NO: 214.

115. An antibody or antigen-binding fragment thereof comprising an HCDR3 sequence disclosed in TABLE 24.

116. The antibody or antigen-binding fragment thereof of claim 115, wherein the antibody or antigen-binding fragment thereof binds to LAG-3.

117. An antibody or antigen-binding fragment thereof comprising a set of heavy chain CDR sequences (HCDR1, HCDR2, and HCDR3) disclosed in TABLE 24.

118. The antibody or antigen-binding fragment thereof of claim 117, wherein the antibody or antigen-binding fragment thereof binds to LAG-3.

119. An antibody or antigen-binding fragment thereof comprising a set of heavy chain and light chain CDR sequences (HCDR1, HCDR2, HCDR3, LCDR1, LCDR2, and LCDR3) disclosed in TABLE 24.

120. The antibody or antigen-binding fragment thereof of claim 119, wherein the antibody or antigen-binding fragment thereof binds to LAG-3.

121. An antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 80% sequence identity to a VH and VL pair disclosed in TABLEs 20 & 21.

122. The antibody or antigen-binding fragment thereof of claim 121, wherein the antibody or antigen-binding fragment thereof binds to LAG-3.

123. An antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL each comprise amino acid sequences with at least 95% sequence identity to a VH and VL pair disclosed in TABLEs 20 & 21.

124. The antibody or antigen-binding fragment thereof of claim 123, wherein the antibody or antigen-binding fragment thereof binds to LAG-3.

125. An antibody or antigen-binding fragment thereof comprising a VH and VL pair, wherein the VH and the VL comprise the amino acid sequences of a VH and VL pair disclosed in TABLES 20 & 21.

126. The antibody or antigen-binding fragment thereof of claim 125, wherein the antibody or antigen-binding fragment thereof binds to LAG-3.

Images