WO2017015141A1

WO2017015141A1 - Humanized anti-glycophorin a antibodies and uses thereof

Info

Publication number: WO2017015141A1
Application number: PCT/US2016/042573
Authority: WO
Inventors: Jeffrey Charles Way; Devin R. BURRILL; Kevin Sebastiaan HOF
Original assignee: President And Fellows Of Harvard College
Priority date: 2015-07-17
Filing date: 2016-07-15
Publication date: 2017-01-26

Abstract

Aspects of the invention relate to Glycophorin A-binding proteins and uses thereof.

Description

HUMANIZED ANTI-GLYCOPHORIN A ANTIBODIES AND USES THEREOF

RELATED APPLICATIONS

This application claims the benefit under 35 U.S.C. § 119(e) of U.S. Provisional Application Serial No. 62/194,169, filed on July 17, 2015, entitled "HUMANIZED ANTI- GLYCOPHORIN A ANTIBODIES AND USES THEREOF," and U.S. Provisional

Application Serial No. 62/292,790, filed on February 8, 2016, entitled "HUMANIZED ANTI-GLYCOPHORIN A ANTIBODIES AND USES THEREOF," the entire contents of each of which is incorporated by reference herein in its entirety.

GOVERNMENT FUNDING

This invention was made with government support under W91 lNF-11-2-0056 awarded by U.S. Department of Defense - DARPA and under GM036373 and

5F32HL122007-02 awarded by National Institutes of Health. The government has certain rights in the invention.

FIELD OF INVENTION

The disclosure relates, at least in part, to humanized anti-Glycophorin A antibodies and their use in targeting fusion proteins.

BACKGROUND OF INVENTION

Glycophorin A (GYPA) and other red blood cell surface proteins are useful in a number of therapeutic contexts. For example, the serum half-life of therapeutic proteins can be extended by fusion of the therapeutic element to another element, such as an antibody that binds to Glycophorin A. Such a fusion protein then equilibrates on and off of the red blood cell, and since the turnover of red blood cells is much slower than serum proteins, the half- life of the therapeutic fusion protein in the body is extended (Singhal and Gupta (1986) FEBS Lett. 201(2):321-6; Kontos and Hubbell (2010) Molecular Pharmaceutics 7(6):2141-2147). Additionally, attachment of a Glycophorin A-binding element to an antigen can cause a deletion of T cells that react to the antigen, thereby increasing immunotolerance (US Patent Publication No. 2012/0178139). Glycophorin A-binding elements may also be used to target erythropoietin (EPO) activity specifically to red blood cell precursors, and away from other cell types (Taylor et al. (2010) Protein Engineering Design & Selection 23(4):251-60). SUMMARY OF INVENTION

Aspects of the invention relate to a Glycophorin A-binding protein comprising a sequence selected from the group consisting of SEQ ID NOs: 2-5, 7-9, and 52-60. In some embodiments, the protein comprises an antibody or a Glycophorin A-binding fragment thereof.

In some embodiments, the protein comprises an scFv. In some embodiments, the protein comprises a protein domain that does not bind to Glycophorin A. In some

embodiments, the protein comprises a cytokine or hormone. In some embodiments, the cytokine or hormone is a four-helix-bundle protein. In some embodiments, the protein comprises an antigen.

Aspects of the invention relate to nucleic acid sequences encoding a Glycophorin A- binding protein, wherein the nucleic acid sequence comprises a sequence selected from the group consisting of SEQ ID NOs: 41-49.

Aspects of the invention relate to a Glycophorin A-binding protein comprising a sequence selected from the group consisting of the sequences in Table 1, wherein the sequence is not one of the murine sequences listed in Table 1.

In some embodiments, the protein comprises a cytokine or hormone. In some embodiments, the cytokine or hormone is Erythropoietin (EPO). In some embodiments, EPO contains one or more mutations relative to wild type human EPO. In some

embodiments, the one or more mutations relative to wild type human EPO are at an amino acid selected from R150, A30, H32, P87, W88, P90, R53, and E55. In some embodiments, EPO contains the mutation R150A. In some embodiments, the binding protein further comprises one or more additional mutations selected from A30N, H32T, P87V, W88N, P90T, R53N, and E55T.

In some embodiments, the Glycophorin A-binding protein comprises a linker. In some embodiments, the linker is located between the EPO and the Glycophorin A-bindin protein that binds to Glycophorin A.

Aspects of the invention relate to an antibody that selectively binds to Glycophorin A, wherein the antibody comprises a sequence selected from the group consisting of SEQ ID NOs: 2-5, 7-9, ad 52-70. In some embodiments, the antibody is an scFv.

Aspects of the invention relate to methods of administering a therapeutically effective amount of a Glycophorin A-binding protein described herein to a subject in need thereof. In some embodiments, the subject has anemia or kidney failure. In some embodiments, the therapeutically effective amount of the Glycophorin A-binding protein is administered to the subject more than once.

Aspects of the invention relate to a Glycophorin A-binding protein comprising SEQ ID NO:39, 63-65, or 68-72.

Other aspects of the invention relate to engineered proteins comprising an scFv that binds to glycophorin A and an erythropoietin moiety that contains at least four N-linked glycosylation sites. Yet other aspects of the invention relate to engineered proteins comprising an scFv that binds to glycophorin A, and an erythropoietin (EPO) moiety that contains one or more mutations that, in the absence of the scFv, reduces binding to a homodimeric EPO receptor by at least about two-fold but has less than a two-fold effect on binding to an EPO receptor/CD 131 heterodimer. Other aspects of the invention relate to engineered proteins comprising an scFv that binds to glycophorin A, and an erythropoietin moiety comprising (i) at least four N-linked glycosylation sites, and (ii) one or more mutations relative to wild type human EPO that, in the absence of the scFv, reduces binding to a homodimer of erythropoietin receptors by at least about two-fold but has less than a twofold effect on binding to a heterodimer of an erythropoietin receptor and CD 131. In some embodiments, the one or more mutations relative to wild type human EPO are at an amino acid selected from the group consisting of L5, R14, L93, D96, K97, S lOO, R103, S 104, T107, and L108. In some embodiments, the one or more mutations are selected from the group consisting of L5V, L5A, R14K, R14M, R14I, R14A, R14S, L93A, D96R, D96K, K97A, K97S, S 100R, R103S, R103M, R103I, R103K, R103E, S 104A, S 104G, T107S, T107S, T107A, L108V, L108A, and L108S.

Each of the limitations of the invention can encompass various embodiments of the invention. It is, therefore, anticipated that each of the limitations of the invention involving any one element or combinations of elements can be included in each aspect of the invention. This invention is not limited in its application to the details of construction and the arrangement of components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments and of being practiced or of being carried out in various ways.

BRIEF DESCRIPTION OF DRAWINGS

The accompanying drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component that is illustrated in various figures is represented by a like numeral. For purposes of clarity, not every component may be labeled in every drawing. In the drawings:

FIGs. 1A-1C show schematic illustrations of a chimeric activator. FIG. 1A depicts a targeting element fused to an activity element with reduced affinity for its receptor. The fusion protein initially binds to a target cell receptor via the targeting element. FIG. IB shows that increased local concentration of the activity element overcomes its receptor binding deficit and facilitates ligand-receptor interaction. FIG. 1C shows that on non-target cells lacking a receptor for the targeting element, the activity element produces little effect due to weak affinity for its own receptor. FIGs. 1A-1C are adapted from Figs. 1C-1E of Taylor et al. (2010). Protein Engineering, Design & Selection 23(4):251-260.

FIGs. 2A-2B depict a non-limiting example of a 10F7-mutEpo chimeric activator. FIG. 2A shows a schematic illustration of 10F7-mutEpo targeted protein structure. The protein is a single polypeptide chain with a short {e.g., approximately 35 amino acid long) linker attaching 10F7 antibody light and heavy chain variable ("V") regions to a mutEpo (mutated EPO) molecule that has reduced affinity for the EPO receptor (EPO-R). FIG. 2B shows a molecular model of a 10F7-mutEpo molecule binding to a red blood cell ("RBC") precursor cell via the GYPA receptor and the EPO receptor. FIG. 2B is adapted from Fig. IB of Taylor et al. (2010). Protein Engineering, Design & Selection 23(4):251-260.

FIG. 3 shows examples of humanized 10F7 antibody heavy chain variable regions ("VH"). Humanized VH regions are aligned against murine 10F7 VH (referred to as "Parent"). The sequences listed in FIG. 3 correspond to SEQ ID NOs: l-5. Sequences in bold indicate CDR1, 2 and 3 in the original mouse sequence.

FIG. 4 shows examples of humanized 10F7 antibody light chain variable regions ("VL"). Humanized VL regions are aligned against murine 10F7 VL (referred to as "Original"). The sequences listed in FIG. 4 correspond to SEQ ID NOs:6-9. Sequences in bold indicate CDR1, 2 and 3 in the original mouse sequence.

FIG. 5 depicts a homology model of 10F7 V regions. The homology model was generated using the Rosetta Online Server ("ROSIE") available through the website of the Gray Lab at Johns Hopkins University (rosie.graylab.jhu.edu/antibody). FIGs. 6A-6D show schematic illustrations of a chimeric activator targeting EPO to red blood cell (RBC) precursors. FIG. 6A shows a schematic illustration of a chimeric activator binding to a targeting receptor and activity receptor. FIG. 6 A top panel depicts a targeting element tethered by a linker to a mutated activity element. The black X indicates an amino acid mutation. The chimeric activator initially binds (bold arrow) a target cell receptor via the targeting element, while the mutated activity element has a low affinity (dashed arrow) for its corresponding receptor. In FIG. 6A middle panel, increased local concentration of the activity element overcomes the receptor-binding deficit of the amino acid mutation and facilitates interaction with the activity receptor (bold arrow). FIG. 6A bottom panel shows that this interaction results in signal activation via the activity receptor (bold arrow). FIG. 6B shows that the activity element has little effect (dashed arrow) on off-target cells lacking a receptor for the targeting element due to the amino acid mutation in the activity element. FIG. 6C presents a schematic molecular model of chimeric activator 10F7-EPO_RI_SOA showing the 10F7 scFv binding to glycophorin (huGYPA) and EPO binding to the receptor EPO-R.

FIG. 6D shows a predicted expression pattern of EPO-R (middle column) and huGYPA (right column) on RBCs during erythropoiesis, illustrating a period of overlap the bone marrow, during which both EPO-R and huGYPA are predicted to be expressed. FIGs. 7A-7C present in vitro characterization of chimeric activator variants. FIG. 7A shows SDS-PAGE analysis of the depicted fusion proteins to determine purity, presence of full-length protein (72 kDa), potential degradation products, and release of N-linked carbohydrate chains upon treatment with (+) or without (-) PNGase F enzyme. PNGase F runs at 32 kDa. FIG. 7B presents results of in vitro kinetic analyses of the interaction between EPO-R and unfused EPO, 10F7-EPO, or 10F7-EPO_RI₅₀A- FIG. 7C shows in vitro proliferation of EPO-R-positive MCF-7 cells versus EPO-R-negative HeLa cells, following a 3-day incubation with saline, or 50 nM of 10F7-EPO or 10F7-EPO_RI_SOA- The graph displays the mean + SEM (n = 3). Comparisons between treatments were made using a Student's t test (*P < 0.1).

FIGs. 8A-8Q show the pharmacodynamic effects of chimeric activator variants on reticulocytes and reticulated platelets. FIG. 8A presents a schematic illustration of the experimental procedure. Transgenic mice expressing huGYPA received a single IP injection of darbepoetin, saline, or 10F7-EPO variants at the indicated concentration. Blood samples were taken by tail-nick on days 0, 4, 7, and 11. FIGs. 8B-8F show the percent change in the reticulocyte fraction of total RBCs, as measured by flow cytometry following the indicated treatment. FIGs. 8G-8K show the percent change in the reticulated platelet fraction of total platelets, as measured by flow cytometry following the indicated treatment. FIGs. 8L-8M show the percent change in reticulocytes in transgenic mice expressing huGYA or non- transgenic mice following a single intraperitoneal injection of 10F7-EPOR150A. FIGs. 8N- 80 show the percent change in the reticulated platelets in transgenic mice expressing huGYA or non-transgenic mice following a single intraperitoneal injection of 10F7-EPOR150A. FIG. 8P shows the percent change in reticulocytes in transgenic mice expressing huGYA following a single intraperitoneal injection of 10F7-EPO_K4SD- FIG. 8Q shows the percent change in reticulated platelets in transgenic mice expressing huGYPA following a single intraperitoneal injection of 10F7-EPO_K4SD- Measurements were baseline-subtracted relative to day 0. Graphs display the mean + SEM (n = 4). Comparisons between treatments were made using a Student's t test (N.S. = not significant, **P < 0.05, ***P < 0.005). Chimeric activators were produced from transient transfections.

FIGs. 9A-9P show the pharmacodynamic effects of chimeric activator variants reticulocytes, hematocrit, and total platelets. FIG. 9A presents a schematic illustration of the experimental procedure. Transgenic mice expressing huGYPA received a single IP injection of darbepoetin, saline, or 10F7-EPO variants at the indicated concentration. Blood samples were taken by tail-nick on days 0, 4, 7, and 11. FIGs. 9B-9F show the percent change in the reticulocyte fraction of whole blood, as measured by flow cytometry following the indicated treatment. FIGs. 9G-9K show the percent change in the hematocrit, as measured by flow cytometry following the indicated treatment. FIGs. 9L-9P show the percent change in the total platelet counts by hematology analyzer following the indicated treatment.

Measurements were baseline- subtracted relative to day 0. Graphs display the mean + SEM (n = 4) of the day 4 measurements. Comparisons between treatments were made using a Student's t test (*P < 0.1, **P < 0.05, ***P < 0.005). Chimeric activators were produced from transient transfections.

FIGs. 10A-10D show the pharmacokinetics of chimeric activator 10F7-EPO_RI_SOA- FIG. 10A presents a compartment model illustrating the expected bio-distribution and elimination of 10F7-EPO_RI_SOA following a single intravenous injection. The chimeric activator enters the plasma, where it immediately binds to mature RBCs (left box with bold outline) that express huGYPA and act as a drug sink. Free drug in the plasma can enter other tissues to stimulate expansion of late RBC precursors (right box with bold outline) or other cell types that express EPO-R. FIG. 10B presents a schematic illustration of the experimental procedure. Transgenic mice expressing huGYPA received a single 100 μg dose of 10F7- EPO_RI5_OA intravenously, and blood was collected in a 120 h time-course. FIG. IOC presents a graph showing the percent of the injected drug (chimeric activator) in the plasma as measured by ELISA. FIG. 10D presents a graph showing the percent of the RBC-bound drug by flow cytometry. Measurements are relative to amount of drug detected at T = 0 (100%). Graphs display the mean + SEM (n = 2). The terminal plasma half-life of 10F7-EPO_RI_SOA in huGYPA transgenic or non-transgenic mice is indicated by "t^" in each graph.

FIGs. 11A-11D show in vitro verification of activity of examples of chimeric activator variants. Erythroleukemia TF-1 cells were treated with 10F7-EPO variants or unfused control proteins and allowed to proliferate for 72 h before addition of WST-1 reagent. Cell proliferation was plotted against protein concentration. FIG. 11A shows the chimeric activator 10F7-EPO compared to EPO. FIG. 11B shows the chimeric activator 10F7-EPO_RI5OA compared to EPO_RISOA- FIG. 11C shows the chimeric activator 10F7_W99G- EPO_RI5OA compared to EPO_RISOA- FIG. 11D shows the chimeric activator 10F7-EPO_K4S_D compared to EPO_K4SD- Plots were fitted by nonlinear regression to determine logEC50 values. Graphs display the mean + SEM (n = 3). N/A - not applicable; it was not possible to calculate a logEC50 for proteins containing the K45D mutation because saturation was not achieved.

FIGs. 12A-12B show huGYPA expression on human RBCs compared to transgenic mouse RBCs. FIG. 12A shows huGYPA expression on human RBCs or transgenic mouse RBCs as determined by flow cytometry using a PE-conjugated anti-huGYPA antibody and BD Quantibrite™ PE beads (BD Biosciences; San Jose, CA). The graph displays mean +SEM (n = 2). FIG. 12B presents the predicted percentage of 10F7-EPOR150A bound to RBCs upon injection calculated based on experimental and known values.

FIGs. 13A-13K show the pharmacodynamic effects of chimeric activator variants on reticulocytes and reticulated platelets in huGYPA transgenic and non-transgenic mice.

Chimeric activators used in these experiments were made from drug-selected stable pools of transfected cells. FIG. 13A presents a schematic illustration of the experimental procedure. Transgenic mice expressing huGYPA received a single IP injection of darbepoetin, saline, or 10F7-EPO variants at the indicated concentration. Blood samples were taken by tail-nick on days 0, 4, 7, and 11 post-dosing. FIGs. 13B-13F show the percent change in the

reticulocytes, as measured by flow cytometry following the indicated treatment. FIGs. 13G- 13K show the percent change in the reticulated platelets, as measured by flow cytometry following the indicated treatment. Measurements were background- subtracted relative to day 0. Graphs display the mean + SEM (n = 3).

FIG. 14A-14P show the pharmacodynamic effects of chimeric activator variants on reticulocytes, hematocrit, and total platelets in huGYPA transgenic and non-transgenic mice. This figure presents the complete data from Figure 4, showing all experimental time-points. FIG. 14A presents a schematic illustration of the experimental procedure. Transgenic mice expressing huGYPA received a single IP injection of darbepoetin, saline, or 10F7-EPO variants at the indicated concentration. Blood samples were taken by tail-nick on days 0, 4, 7, and 11 post-dosing. FIGs. 14B-14F show the percent change in the reticulocytes, as measured by flow cytometry following the indicated treatment. FIGs. 14G-14K show the percent change in hematocrit, as measured by flow cytometry following the indicated treatment. FIGs. 14L-14P show the percent change in the total platelets, as measured by hematology analyzer. Measurements were background- subtracted relative to day 0. Graphs display mean + SEM (n = 4).

FIGs. 15 A- 15P show the pharmacodynamic effects of chimeric activator variants on reticulocytes, hematocrit, and total platelets in huGYPA transgenic and non-transgenic mice. Chimeric activators used in this experiment were produced using drug-selected stable pools of transfected cells. FIG. 15A presents a schematic illustration of the experimental procedure. Transgenic mice expressing huGYPA received a single IP injection of

darbepoetin, saline, or 10F7-EPO variants at the indicated concentration. Blood samples were taken by tail-nick on days 0, 4, 7, and 11 post-dosing. FIGs. 15B-15F show the percent change in the reticulocytes, as measured by flow cytometry following the indicated treatment. FIGs. 15G-15K show the percent change in hematocrit, as measured by flow cytometry following the indicated treatment. FIGs. 15L-15P show the percent change in the total platelets, as by hematology analyzer. Measurements were background-subtracted relative to day 0. Graphs display mean + SEM (n = 3). FIG. 16 presents the ratios of RBC-bound 10F7-EPO_RI5_OA (geometric mean fluorescence) to drug in plasma (% of T = 0 measurement) at time-points post-dosing.

huGYPA transgenic mice were injected intravenously with 100 μg of 10F7-EPO_RI5_OA- The drug bound to RBCs or free in plasma was measured using flow cytometry or ELISA, respectively (n = 2).

FIG. 17 presents the levels of RBC-bound 10F7-EPO_RI_SOA at time-points post-doing. huGYPA transgenic and wild-type mice were injected intravenously with 100 μg of 10F7- EPO_RI5_OA- The drug bound to RBCs was measured using flow cytometry (n = 2).

FIGs. 18A-18F present schematic illustrations of an EPO targeted to red blood cell precursors and cells at risk for hypoxia, but not to cells that enhance thrombosis. FIG. 18A shows a schematic of EPO (1) binding to an EPO receptor dimer with a strongly interacting EPO receptor (2) and an weakly interacting EPO receptor (3). FIG. 18B shows a mutant EPO (4) that lacks a contact with the weakly interacting EPO receptor: (5) indicates the absence of the contact. Examples of mutations that may be useful in this context are described in Example 11. FIG. 18C shows the same mutant EPO of FIG. 18B binding to a heterodimer of EPO receptor making a strong contact and CD131 (6, black). Binding to CD131 is not affected by the mutation, because the mutated amino acid does not contact CD131. In FIGs. 18A-18C, the cell membrane (7) is at the bottom of the figure. FIG. 18D shows a chimeric activator that includes an scFv (8) that binds to glycophorin A (9, thick wavy line) attached via a linker (10) to the mutant EPO. The binding takes place on the surface of a red blood cell precursor (11). FIG. 18E shows that on the surface of a cell that contributes to thrombosis (12), such as a platelet precursor, the mutated EPO of the chimeric activator does not bind to its homodimeric EPO receptor because the mutated EPO is not tethered to the cell surface. FIG. 18F shows that on the surface of a cell at risk of hypoxia (13), the EPO receptor and CD 131 are co-expressed. The mutated EPO of the chimeric activator can still bind to the EPO-R/CD131 heterodimer because the mutation does not affect CD 131 binding.

DETAILED DESCRIPTION

The invention relates, at least in part, to the development of novel antibodies, and variable ("V") region elements of antibodies, that bind to GYPA. Antibodies and V regions described herein are suitable for human therapeutics based on their reduced immunogenicity and increased stability relative to previous antibodies generated against GYPA. Antibodies and V regions associated with the invention can be used to target molecules, for example hormones and cytokines, such as erythropoietin (EPO), to red blood cell precursors that express GYPA, creating improved delivery of EPO, referred to herein as "targeted EPO." Currently, EPO is used as an anemia drug. However, the currently available form of EPO has significant side effects. Such side effects may be caused by EPO binding to cells other than red blood cell precursors. For example, EPO also binds to platelet precursors, endothelial cells, and other cells in the blood system, which promotes thrombosis. Targeted EPO, described herein, represents an engineered fusion protein comprising a humanized antibody that selectively binds to GYPA and a mutated form of EPO that has reduced activity.

Targeted EPO is targeted to red blood cell precursors and away from platelet precursors, endothelial cells, and other cells in the blood system, avoiding side effects caused by the currently available form of EPO.

Accordingly, aspects of the invention relate to humanized antibodies that selectively bind to GYPA and fusion proteins comprising such antibodies. Proteins described herein that bind to GYPA are referred to as Glycophorin A-binding proteins. Glycophorin A-binding proteins can be chimeric activators, which are described in and incorporated by reference from, WO 2008/124086, filed on April 5, 2008, and entitled "Chimeric Activators:

Quantitatively Designed Protein Therapeutics and Uses Thereof." In some aspects of the invention, the chimeric activators described herein induce the proliferation of reticulocytes but do not substantially affect hematocrits or reticulated platelets.

Chimeric Activators

Aspects of the invention relate to chimeric proteins that include a targeting element connected to an activity element. In some embodiments, the targeting element binds to molecules that are selectively present on target cells. In some embodiments, the activity element is a variant of a naturally occurring protein that activates cells by binding to one or more cell surface receptors. The variant is selected such that it has reduced or no cell activating properties in the absence of the targeting element. The targeting element is selected such that it selectively binds to target cells thereby increasing the local concentration of the variant activity element on the target cells to a level that results in activation of those cells. In some embodiments, cell activation results in stimulation of red blood cell production.

An activity element may be a receptor binding protein (or functional portion or domain thereof) that binds to one or more naturally-occurring receptors on a cell surface, thereby mediating signaling to the cell (e.g., via signal transduction). A variant activity element may include one or more naturally occurring and/or engineered mutations that result in reduced binding to one or more (e.g., all) natural receptors that are bound by the wild-type activity element. For example, the binding affinity of the variant activity element for one or more of its receptors may be at least 2-fold lower and preferably at least 5-fold lower or 8- fold lower (e.g., at least 10-fold lower, about 10-50-fold lower, about 50- 100-fold lower, about 100- 150-fold lower, about 150-200-fold lower, or more than 200-fold lower) than the binding affinity of the wild-type activity element for its natural receptor(s). As a result, in some embodiments the activity element by itself (e.g., not part of a fusion or chimeric protein comprising a targeting element) is significantly less active (e.g., substantially inactive) because it cannot bind or has reduced binding to its receptor(s). In some embodiments, the variant activity element stimulates less signaling to the cell as compared to a wild type activity element (e.g., an activity element that does not have the variation/mutation). In some embodiments, the variant activity element induces at least 2-fold, at least 3-fold, at least 4- fold, at least 5-fold, at least 10-fold, at least 20-fold, at least 50-fold, at least 100-fold or less signaling to the cell as compared to a wild type activity element. Accordingly, chimeric proteins of the invention are useful to avoid unwanted side effects caused by the activity element binding to its natural receptor on non-target cells. However, the activity element is active on target cells because the targeting element provides the missing binding affinity required for activation.

According to some embodiments of the invention, the affinity of the targeting element for its target molecule (e.g., receptor) may be similar to the binding affinity of the variant activity element for its receptor(s). However, the affinity of the targeting element for its target molecule is typically higher (e.g., at least 2-fold, at least 5-fold, about 5- 10-fold, about 10-50-fold, about 50- 100-fold, about 100-500-fold, about 500- 1 ,000-fold, at least 1 ,000-fold, at least 10,000-fold, or at least 100,000-fold higher) than the binding affinity of the variant activity element for its receptor(s). Also encompassed within the present invention are mutations or variations that may enhance or reduce the binding affinity of the targeting element for its targeting molecule (e.g., receptor).

As used herein, "binding affinity" refers to the apparent association constant or KA-

The KA is the reciprocal of the dissociation constant (KD). In some embodiments, the targeting element or the activity element described herein may have a binding affinity (KD) of at least 10^"5, 10^"6, 10^"7, 10^"8, 10^"9, 10^"10 M, 10^"11 M or lower for the corresponding receptor or targeting molecule. In some embodiments, the activity and targeting elements are connected via a linker. The linker may be a natural or artificial peptide that provides sufficient flexibility to allow both the targeting element and the activity element of a chimeric activator to bind

simultaneously to their targets and/or receptors on a target cell. The linker length may be optimized for the sizes of the target molecules and/or receptors being bound by targeting and activity elements, the sizes of the targeting and activity elements themselves, and other factors described herein in more detail. Any linker known in the art may be compatible with connecting the activity and targeting elements of the chimeric activators described herein. In some embodiments, the linker is a peptide linker. In some embodiments, the peptide linker comprises glycine and serine residues. In some embodiments, the linker comprises amino acids selected from the group consisting of glycine, serine, glutamate, and aspartate.

However, in some embodiments, a linker may not be required (or only a short linker may be required) if one or both of the targeting and activity elements are flexible or are connected to each other by flexible domains.

Accordingly, aspects of the invention relate to chimeric activators— proteins that include a targeting element and an activity element that are capable of simultaneously binding to multiple (two or more) unrelated receptors on the same target cell. In some embodiments, a chimeric activator has an activity element, a linker, and a targeting element. (FIG. 1A) It should be appreciated that the linker, if present, is located between the targeting and activity elements. In some embodiments, the targeting element also comprises a linker. However, aspects of the invention are not limited by the relative positions of the targeting and activity elements. For example, the targeting element may be N-terminal to the activity element (regardless of whether a linker separates the two elements). Alternatively, the activity element may be N-terminal to the targeting element (regardless of whether a linker separates the two elements). In addition, proteins of the invention may include one or more additional peptide elements (at the N-terminus, the C-terminus, or within the protein) for purification and/or detection (e.g., peptide tags), stability, function (e.g., secretion and or other function), etc., or any combination thereof. It should be appreciated that targeting and activity elements may bind to overlapping types of molecules (e.g., receptors, substrates, enzymes, or other molecules) that are present on a cell surface. However, in any particular configuration, the element that is being used as a targeting element has sufficient affinity to target a chimeric protein to a cell-surface molecule that is preferentially (e.g., uniquely or selectively or specifically) present or exposed on a target cell of interest. In contrast, an element that is being used as an activity element may have relatively low affinity for a cell surface molecule. However, the element that is being used as an activity element may interact equally with a molecule that is present on all cell types, but only activates the target cell types due to its linkage to the targeting element. However, it should be appreciated that in some embodiments, both the targeting and activity elements interact with different cell surface molecules that all are preferentially present or exposed on the target cells.

Targeting Elements

In some embodiments, a targeting element comprises a protein component that selectively directs the binding of the chimeric protein to the surface of a desired cell or cells. Accordingly, a targeting element binds specifically to a cell-surface moiety (e.g., an antigen, epitope, protein, glycoprotein, lipid, carbohydrate, or other molecule or portion thereof) that is either only present or exposed on the target cell or is present or exposed in higher amounts on the target cell relative to non-target cells {e.g., about 2-fold, about 5-fold, about 10-fold, about 50-fold, about 100-fold, about 200-fold, about 500-fold, about 1,000-fold, about

10,000-fold, or more on the target cell). The targeting element can be any protein that binds to a cell-surface molecule. Targeting elements typically target to a subset of cells.

In some embodiments, the targeting element specifically targets the chimeric protein to a particular cell or subset of cells expressing a cell-surface moiety {e.g., a receptor for the targeting element. As used herein, a targeting element "specifically" targets a chimeric protein to a particular cell or subset of cells expressing a cell-surface moiety if it binds or interacts more frequently, more rapidly, with greater duration and/or with greater affinity with a particular cell-surface moiety than it does with alternative cell-surface moieties.

In some embodiments, the targeting element targets the chimeric protein to a cell that expresses glycophorin A (GYPA). In some embodiments, the targeting element target the chimeric protein to a cell that specifically expresses human GYPA (rather than mouse GYPA). In some embodiments, the targeting elements target the chimeric protein to erythrocytes, or cells of the erythrocyte lineage.

In some embodiments, the targeting element is an antibody. As used herein, the term "antibody" includes a full-length antibody, chimeric antibody, Fab', Fab, F(ab')2, single domain antibody (DAB), Fv, single chain Fv (scFv), minibody, diabody, triabody, or a mixture thereof, for example. In particular embodiments, the antibody binds to GYPA. In certain embodiments, the antibody is a humanized antibody that binds to GYPA. In some embodiments, the antibody is a scFv that binds to GYPA. In some embodiments, the antibody is a scFv that binds to human GYPA (huGYPA). In some embodiments, the antibody is a scFv from the antibody 10F7, which has been described in Bigbee et al Mol Immunol(l9S3) 20(12): 1353- 1362 and Catimel et al. J. Immunol. Methods (1993) 165(2): 183-192, herein incorporated by reference in their entirety.

Significantly, as is well known in the art, only a small portion of an antibody molecule, the paratope, is involved in the binding of the antibody to its epitope (see, in general, Clark, W.R. (1986) The Experimental Foundations of Modern Immunology Wiley & Sons, Inc., New York; Roitt, I. (1991) Essential Immunology, 7th Ed., Blackwell Scientific Publications, Oxford). The pFc' and Fc regions, for example, are effectors of the complement cascade but are not involved in antigen binding. An antibody from which the pFc' region has been enzymatically cleaved, or which has been produced without the pFc' region, designated an F(ab')2 fragment, retains both of the antigen binding sites of an intact antibody. Similarly, an antibody from which the Fc region has been enzymatically cleaved, or which has been produced without the Fc region, designated an Fab fragment, retains one of the antigen binding sites of an intact antibody molecule. Proceeding further, Fab fragments consist of a covalently bound antibody light chain and a portion of the antibody heavy chain denoted Fd. The Fd fragments are the major determinant of antibody specificity (a single Fd fragment may be associated with up to ten different light chains without altering antibody specificity) and Fd fragments retain epitope-binding ability in isolation.

Within the antigen-binding portion of an antibody, as is well-known in the art, there are complementarity determining regions (CDRs), which directly interact with the epitope of the antigen, and framework regions (FRs), which maintain the tertiary structure of the paratope (see, in general, Clark, 1986; Roitt, 1991). In both the heavy chain Fd fragment and the light chain of IgG immunoglobulins, there are four framework regions (FRl through FR4) separated respectively by three complementarity determining regions (CDR1 through CDR3). The CDRs, and in particular the CDR3 regions, and more particularly the heavy chain CDR3, are largely responsible for antibody specificity.

It is now well established in the art that the non-CDR regions of a mammalian antibody may be replaced with similar regions of conspecific or heterospecific antibodies while retaining the epitopic specificity of the original antibody. This is most clearly manifested in the development and use of "humanized" antibodies in which non-human CDRs are covalently joined to human FR and/or Fc/pFc' regions to produce a functional antibody. See, e.g., U.S. patents 4,816,567, 5,225,539, 5,585,089, 5,693,762 and 5,859,205, incorporated by reference herein. Fully human monoclonal antibodies also can be prepared by immunizing mice transgenic for large portions of human immunoglobulin heavy and light chain loci.

Following immunization of these mice (e.g., XenoMouse (Abgenix), HuMAb mice

(Medarex/GenPharm)), monoclonal antibodies can be prepared according to standard hybridoma technology. These monoclonal antibodies will have human immunoglobulin amino acid sequences and therefore will not provoke human anti-mouse antibody (HAMA) responses when administered to humans.

Thus, as will be apparent to one of ordinary skill in the art, the present invention also provides for F(ab')2, Fab, Fv and Fd fragments; chimeric antibodies in which the Fc and/or FR and/or CDRl and/or CDR2 and/or light chain CDR3 regions have been replaced by homologous human or non-human sequences; chimeric F(ab')2 fragment antibodies in which the FR and/or CDRl and/or CDR2 and/or light chain CDR3 regions have been replaced by homologous human or non-human sequences; chimeric Fab fragment antibodies in which the FR and/or CDRl and/or CDR2 and/or light chain CDR3 regions have been replaced by homologous human or non-human sequences; and chimeric Fd fragment antibodies in which the FR and/or CDRl and/or CDR2 regions have been replaced by homologous human or non- human sequences.

The present invention also includes so-called single chain antibodies. In some embodiments, the targeting element of a chimeric protein is an antibody moiety that has been expressed from a single linear nucleic acid as a single peptide and that can fold into a fully functional antibody. In some embodiments, the chimeric protein is expressed from a nucleic acid that encodes a portion of an antibody (e.g., a heavy chain, a light chain, or a portion thereof as described herein) fused to the activity element (via the optional linker). The remainder of the antibody required for specific binding can be provided in trans (e.g., expressed from a different coding sequence in the same cell or expressed in a different cells and mixed in vitro, etc., or any combination thereof). Accordingly, it should be appreciated that a chimeric protein of the invention may be a single linear polypeptide chain or it may include additional polypeptide chains of an antibody or other binding protein required for specific binding to a target molecule.

Also within the scope of the present invention are targeting elements comprising one or more mutations or variations that enhance the binding affinity of the targeting element for its targeting molecule by at least 2-fold, at least 5-fold, about 5- 10-fold, about 10-50-fold, about 50- 100-fold, about 100-500-fold, about 500- 1,000-fold, at least 1,000-fold, at least 10,000-fold, or at least 100,000-fold higher than the binding affinity of the targeting element for its targeting molecule in the absence of the mutation(s) or variation(s). In some embodiments, the targeting element comprises one or more mutations or variations that reduce the binding affinity of the targeting element for its targeting molecule by at least 2- fold, at least 5-fold, about 5-10-fold, about 10-50-fold, about 50-100-fold, about 100-500- fold, about 500-1,000-fold, at least 1,000-fold, at least 10,000-fold, or at least 100,000-fold lower than the binding affinity of the targeting element for its targeting molecule in the absence of the mutation(s) or variation(s).

In some embodiments, the targeting element is an antibody or fragment thereof and comprises one or more (e.g., at least 2, 3, 4, 5, or more) mutations. In some embodiments, the mutations may be within the variable region of the antibody. In some embodiments, the mutations may be within a complementarity determining region (CDR) of the antibody.

Humanized Glycophorin A ( GYPA ) Antibodies

Aspects of the invention relate to humanized antibodies that bind to GYPA. Such antibodies can serve as targeting elements in engineered proteins such as chimeric activators. In some embodiments, humanized anti-GYPA antibodies have reduced immunogenicity relative to previous anti-GYPA antibodies.

A previous GYPA-binding sequence used by Kontos and Hubbell (2010) Molecular Pharmaceutics 7(6):2141-2147, and in US2012/0178139, was a random peptide selected for GYPA binding from a phage display library. The GYPA-binding sequence used by Taylor et al. (2010) Protein Engineering Design & Selection 23(4):251-60, consists of variable ("V") regions from a mouse anti-GYPA antibody. These are both foreign proteins in a human patient. By contrast, described herein are improved GYPA-binding sequences that, compared to the parental antibody V regions, contain more human sequences, fewer mouse specific sequences and when folded have more surface area that is found on the surface of natural human antibodies expected to be in the serum of treated patients. Aspects of the invention relate to modifications of the antibody V region sequences of the anti-GYPA antibody known as 10F7 (also referred to as 10F7MN), which is described in Bigbee et al. (Molecular Immunology (1983) 20(12)1353-1362), as binding to both the M and N alleles of GYPA. These alleles correspond to differences in the first 5 amino acids of mature GYPA and define the M/N blood group system. The sequences of the murine V regions of 10F7 are provided by GenBank accession number AAK85297 and in FIGs. 3 (SEQ ID NO: l) and 4 (SEQ ID NO:6). Examples of humanized VH and VL regions that may be used in the engineered proteins such as chimeric activators described herein are also provided in FIGs. 3 (SEQ ID NO: 2-5) and 5 (SEQ ID -NO: 79), as well as in the Examples provided by SEQ ID NO: 52- 60. a. Enhancement of Protein Stability

A stabilized folded protein has several advantages, including enhanced production during expression in mammalian or other cells, enhanced stability during formulation, and reduced immunogenicity. Without wishing to be bound by any theory, the reduction in immunogenicity may be due to the following mechanism. Generally, when protein pharmaceuticals are injected into a patient, some fraction of the protein is taken up by antigen-presenting cells. The protein transits from the early endosome to the middle endosome and then to the late endosome, and finally to the lysosome where it is completely degraded. In the endosomes, proteolysis of the protein is initiated and a portion of the protein is converted to small peptides, such as 9-mer peptides, that are bound to MHC molecules such as MHC Class II and Class I. The MHC-peptide complexes are recycled from the endosomes to the cell surface and presented to T cells. The 9-mer peptides constitute T cell epitopes. Depending on other signals that may be present, T cells such as CD4+ T-helper cells that recognize these epitopes are stimulated to divide and also promote survival and proliferation of B cells that recognize the same protein.

B cells with rearranged antibody genes are constantly being generated. A small fraction of these will express an antibody capable of recognizing a specific protein such as an injected pharmaceutical protein. A B cell expressing an anti-drug antibody initially expresses this antibody on its surface as part of the B cell receptor complex. Upon antigen binding and crosslinking of B cell receptor complexes, the antibody- antigen complex is internalized and the antigen is processed into peptides and presented via MHC molecules by essentially the same mechanism as for other antigen-presenting cells such as dendritic cells and

macrophages. To survive and proliferate, the B cells are stimulated by previously activated T cells that recognize the same MHC-peptide complex involved in their initial activation.

The extent of proteolysis in the endosome is a function of the structural integrity of the protein. Proteins that are tightly folded undergo less proteolysis in the endosome. Such tightly folded proteins thus are sent, essentially intact, to the lysosome, where they are completely degraded without giving rise to peptides that can be part of MHC-peptide complexes. As a result, when proteins are tightly folded they are less immunogenic because the extent to which they can be processed into peptide antigens is reduced. Aspects of the invention relate to humanized antibodies. As used herein, "humanized antibodies" refers to antibodies from non-human animal species whose sequences have been changed to increase similarity with human sequences.

As discussed in the Examples section, antibody V region sequences of antibodies described herein were designed based on closely related human VH sequences and also based on more distantly related sequences that form more stable protein domains. Humanization of antibody V regions generally involves, as a first step, identifying human framework regions that are most similar to the mouse-derived V regions. A subset of the heavy chain designs described herein were generated in this manner using human VH1 sequences as a template because they are most closely related to the 10F7 VH (SEQ ID NOs: 4, 5). Another subset (SEQ ID NOs: 2, 3) were generated by using human VH3 sequences as a template for humanization because human VH3-based V regions can in some embodiments exhibit increased stability.

As used herein, a "protein domain" or "domain" refers to a distinct globular unit that can be identified as such by a structure determination method such as X-ray crystallography or NMR, by other biophysical methods such as scanning calorimetry according to which a protein domain melts as a distinct unit (e.g., Pabo et al. Proc Natl Acad Sci U SA. (1979) 76(4): 1608-12), or by sequence similarity to protein domains whose structure has been determined. The SCOP database (Murzin et al. J Mol Biol. (1995) 247(4):536-40) provides the identification of protein domains so that the domain organization of a new protein can be identified by sequence comparison. Protein domains comprise an amino acid sequence that is sufficient to drive folding of such a polypeptide into a discrete structure, in which essentially all of the rotatable bonds along the main chain of the polypeptide are constrained to within about 10 degrees. In contrast, linkers, short peptides, molten globules, and unstructured segments are examples of polypeptides that are not domains and do not have these characteristics. b. Modified Sequences

Aspects of the invention relate to modifying the V regions of the 10F7 anti-GYPA antibody. In some embodiments, it is desirable to modify portions of the V region that are distant from the canonical CDRs to create additional binding surfaces that mediate binding to other targets. In particular, antibody V regions may be considered as protein domains consisting of a set of beta strands that are stabilized by a disulfide bond and beta-sheet hydrogen bonding patterns, with loops at the ends. The CDRs constitute one set of loops, but there are loops at the opposite side of the protein that may also be varied. In some embodiments, certain sequences that include the CDRs plus adjacent sequences are used, in order to maintain GYPA binding. Several non-limiting examples of sequences associated with the invention are presented in Table 1 :

Table 1.

Heavy chain CDR1 and adjacent sequences: Light chain CDR1 and adjacent sequences:

SVKLSCKAS GYTFNSYF HW KQRP (murine VTMTCRAS SNVKY MYWYQQ (murine sequence) seq. ) (SEQ ID NO: 10) (SEQ ID NO:29)

CKAS GYTFNSYF HWV (SEQ ID NO: 11) VTITCRAS SNVKY LAWYQQ (SEQ ID NO: 30)

CKAS GYTFNSYF HWVRQ (SEQ ID NO:12) TCRAS SNVKY LAWYQQ (SEQ ID NO: 31)

CKAS GYTFNSYF HWVRQAP (SEQ ID VTITCRAS SNVKY (SEQ ID NO:32)

NO:13) VTITCRAS SNVKY LYWYQQ (SEQ ID NO: 33)

SLRLSCKAS GYTFNSYF HW (SEQ ID NO: 14)

SVRLSCKAS GYTFNSYF HW K (SEQ ID NO: 15)

SVKVSCKAS GYTFNSYF MHW (SEQ ID NO: 16) Light chain CDR2

PKLWIY YTS NLASGVP (murine sequence) (SEQ ID NO:34)

Heavy chain CDR2 PKLLIY YTS NLQSGVP (SEQ ID NO:35)

LEWIGM IRPNGGTTD YNEK (Murine seque: PKLLIY YTS NLASGVP (SEQ ID NO:36) (SEQ ID NO:17)

WVSM IRPNGGTTD Y (SEQ ID NO: 18)

LEWVSM IRPNGGTTD Y (SEQ ID NO: 19) Light chain CDR3

LEWMGM IRPNGGTTD Y (SEQ ID NO: 20) AATYYC QQFTSSPYTFGGGT (SEQ ID NO: 37) LEWMGM IRPNGGTTD YAQ (SEQ ID NO: 21) VATYYC QQFTSSPYTFGQGT (SEQ ID NO: 38) M IRPNGGTTD YAQ (SEQ ID NO:22)

Heavy chain CDR3

DSAVYY CARWEGSYYA LDYWGQG (murine

sequence) (SEQ ID NO:23)

DTAVYY CARWEGSYYA LD (SEQ ID NO: 24)

DTAVYY CARWEGSYYA LDVWGQG (SEQ ID NO: 25)

AVYY CARWEGSYYA LDVWGQG (SEQ ID NO: 26)

DTAVYY CARWEGSYYA LDYWGQG (SEQ ID NO: 27)

DTWYY CARWEGSYYA LDYWGQG (SEQ ID NO: 28)

The sequences listed in Table 1 are unique among GYPA-binding antibodies. The murine sequences listed in Table 1 correspond to sequences from the 10 F7 antibody.

In some embodiments, the antibody comprises one or more of the CDR sequences (VH CDR1, VH CDR2, VH CDR3, VL CDR1, VL CDR2, and/or VH CDR3) from the antibody 10F7 and binds or interacts with huGYPA. In some embodiments, the antibody is a humanized antibody derived from antibody 10F7. In some embodiments, the antibody comprises any of heavy chain variable regions provided by SEQ ID NO: 2-5, and 52-56. In some embodiments, the antibody comprises any of the light chain variable regions provided by SEQ ID NO: 7-9, and 57-60. In some embodiments, the antibody comprises one or more of the CDR sequences provided in Table 1. c. Uses of V Regions Antibody V regions that bind to GYPA and that have reduced immunogenicity are particularly useful in multiple applications. Several non-limiting examples of uses for such antibodies include targeting of EPO to red blood cell precursors, targeting of proteins with a short serum half-life to mature red blood cells to extend serum half-life, and targeting of immunogenic proteins to red blood cells to reduce their immunogenicity.

Aspects of the invention relate to targeting of fusion proteins to GYPA. It should be appreciated that methods and compositions described herein could also be used to target fusion proteins to other red blood cell surface proteins. In some embodiments, binding of a targeting element to GYPA has advantages relative to binding to other red blood cell surface markers. For example, GYPA does not have a known biological function that can be inhibited by antibody binding. Humans and transgenic animals lacking GYPA are pheno typically normal. Further, GYPA is more abundant than Glycophorins B, C, D and E. The only other protein in red blood cell membranes with comparable expression levels to GYPA is Band 3, which is an anion transporter. Since modulating anion transport may have undesirable effects, GYPA is a more attractive targeting receptor.

GYPA exists as a dimer. The copy number of GYPA is about 1 million/cell, and the number of 10F7 binding sites on a red blood cell is about 500,000. In some embodiments, binding of one V region pair to a GYPA dimer inhibits a second binding event.

Silver et al. (US Patent Publication No. 2011/0274658, incorporated by reference herein) and Taylor et al. (2010) Protein Engineering Design & Selection 23(4):251-60 describe the use of GYPA-targeting antibody V regions such as V regions from the 10F7 antibody fused to mutated EPO for the purpose of targeting erythropoietin activity to red blood cell precursors.

Antibodies described herein have improved V regions for more effective targeting. For example, the following sequence is a non-limiting example of a fusion of humanized VH and VL segments in an scFv format, followed by a long linker and a mutated (Argl50Ala) erythropoietin moiety (EPO_RI_SOA)- It should be appreciated that any of the humanized VH regions described herein can be combined with any of the humanized VL regions described herein. For example, any one of the VH segments corresponding to SEQ ID NOs: 2-5 could be combined with any one of the VL segments corresponding to SEQ ID NOs: 7-9.

VH-A - VL-A - linker - EPO(R150A) (SEQ ID NO:39): EVQLVESGGGLVQPGGSLRLSCKASGYTFNSYFMHWVRQAPGKGLVWVSMIRPNGGTTDYAD SVKGRFTI SVDNSKNTLYLQMNSLRAEDTAVYYCARWEGSYYALDVWGQGTTVTVSS GGGGSGGGGSSGGG GSS

DIQMTQSPSSLSASVGDRVTITCRASSNVKYLAWYQQKPGKAPKLLIYYTSNLQS GVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIK GGSSGGGGSGGGSSGGGSSSGGGGSGGGGSSGGGSGGGS

APPRLICDSRVLERYLLEAKEAEKITTGCAEHCSLNEKITVPDTKVNFYAWKRMEVG QQAVEVWQGLALLSEAVLRGQALLVKSSQPWEPLQLHVDKAVSGLRSLTTLLRALGAQKEAI SPPDAAS AAPLRT I TADTFRKLFRVYSNFLAGKLKL YTGEACRTGDR

The Argl50Ala mutation in EPO is indicated in boldface in the above amino acid sequence.

Singhal and Gupta (1986) FEBS Lett.201(2):321-6, showed that covalent attachment of liposomes to an antibody directed against red blood cells had the effect of increasing the concentration of liposomal material in the blood and thus extending the serum half-life of the antibody-conjugated material. Subsequently, Kontos and Hubbell (2010) Molecular

Pharmaceutics 7(6):2141-2147, identified short peptides that bind to GYPA, and further showed that when covalently attached to maltose binding protein (MBP) these peptides had the effect of extending the circulation half-life of MBP in animals, similarly to the results of Singhal and Gupta.

The V regions described herein are useful for extending the serum half-life of proteins that must otherwise be injected frequently. Human growth hormone, also known as somatotropin, illustrates this point. In pediatric patients with growth hormone deficiency, this protein is injected daily or every other day. For example, the following sequence is a fusion of humanized VH and VL segments in an scFv format, followed by a long linker and a human growth hormone moiety.

VH-A - VL-A - linker - human growth hormone (SEQ ID NO:40):

EVQLVESGGGLVQPGGSLRLSCKASGYTFNSYFMHWVRQAPGKGLVWVSMIRPNGGTTDYAD SVKGRFTI SVDNSKNTLYLQMNSLRAEDTAVYYCARWEGSYYALDVWGQGTTVTVSS GGGGSGGGGSSGGG GSS

DIQMTQSPSSLSASVGDRVTITCRASSNVKYLAWYQQKPGKAPKLLIYYTSNLQS GVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLE IK GGSSGGGGSGGGSSGGGSSSGGGGSGGGGSSGGGSGGGS FPTIPLSRLFDNAMLRAHRLHQLAFDTYQEFEEAYIPKEQKYSFLQNPQTSLCFSES IPTPS NREETQQKSNLELLRI SLLLIQSWLEPVQFLRSVFANSLVYGASDSNVYDLLKDLEEGIQTL MGRLEDGSPRTGQIFKQTYSKFDTNSHNDDALLKNYGLLYCFRKDMDKVETFLRIVQCRSVE GSCGF Multiple non-limiting examples of humanized 10F7 heavy and light chain V regions are provided in Example 3 and correspond to SEQ ID NOs: 2-5 and 7-9. Thus, aspects of the invention relate to Glycophorin A-binding proteins comprising a sequence selected from the group consisting of SEQ ID NOs: 2-5 and 7-9. In some embodiments, the Glycophorin A- binding protein comprises SEQ ID NO:39 , 40, 63-65, or 68-72. In some embodiments, the Glycophorin A-binding protein comprises an scFv. In some embodiments, the Glycophorin A-binding protein comprises a sequence selected from the group consisting of the sequences in Table 1.

Aspects of the invention relate to Glycophorin A-binding proteins that comprise a domain that does not bind to Glycophorin. Such a domain can be an activity element, as discussed below, and as described in and incorporated by reference from, WO 2008/124086, filed on April 5, 2008, and entitled "Chimeric Activators: Quantitatively Designed Protein Therapeutics and Uses Thereof." Activity Element

According to aspects of the invention, an activity element comprises a protein component that has the desired signaling activity. This component normally binds to a receptor on the surface of the desired target cell. The activity element can be any protein that binds to a cell-surface receptor and stimulates signal transduction. In some embodiments the activity element is a cytokine or a hormone. In some embodiments, the activity element is a portion of a cytokine or a hormone that is sufficient to bind to a receptor on the surface of the target cell and induce an activity in the cell. In some embodiments, the cytokine or hormone is a four-helix-bundle protein. In some embodiments, the activity element is a variant form (e.g., it is mutated) that has an intrinsic binding to its receptor that is weak as compared to the wild-type protein, to the point that binding of the chimeric protein to a cell is driven by the binding of the targeting element to its receptor. For example, the chimeric activator first binds to a cell through the targeting element, and after this initial binding step, the mutated activity element is in a high local concentration relative to its own receptor, so binding takes place in spite of the mutation. In contrast, the chimeric protein may not significantly bind to cells that have only the activity element receptor and not the targeting element receptor, because the inherent binding of the activity element is simply too weak. The presence of one or more weakening mutations may reduce side effects from activation of non-target cells.

It should be appreciated that the choice of a particular mutation can be tuned to the on-rates and dissociation constants of the naturally occurring targeting element and activity element, the relative numbers of receptors of each type on the target cell, and the binding enhancement factor resulting from pre-binding of the chimeric protein to the receptor for the targeting element. In practice, mutations with a range of different strengths may be tested empirically using techniques known to one of ordinary skill in the art. In some embodiments, one or more mutations may be made in the activity element to reduce the on-rate of binding between the activity element and the receptor on the target cell. In some embodiments, one or more mutations may be made in the activity element to enhance the off-rate of the activity element and the receptor on the target cell.

Aspects of the invention relate to creating mutations in the activity element, such that the mutated form of the activity element binds to its receptor with a reduced affinity relative to a wild-type form of the activity element. Accordingly, chimeric activators differ from fusion proteins in which an element with signaling activity may bind to more than one receptor type, and a mutation is introduced into this element to modulate the receptor type that is used (e.g., certain TNF chimeras, see also for example US 2006/0263368, herein incorporated by reference in its entirety). In contrast, chimeric activators of the invention differ from such distinct fusion proteins in that the signaling by a chimeric activator is directed through the receptor for which affinity has been reduced by the mutation in the activity element.

As used herein, a "mutation" refers to a change in the nucleotide sequence encoding the activity element, relative to a wild-type form of the gene, and includes substitution, deletion, and insertion mutations. A change in the nucleotide sequence may or may not lead to a change in the amino acid sequence, the three-dimensional structure of the protein, and/or the activity of the protein, relative to the wild-type form of the protein. In some embodiments a mutation may be a naturally occurring variant of the gene. In some embodiments, the activity element comprises a mutation of at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or more amino acids relative to a wild type/non-mutated activity element. In some embodiments a mutation may be a single amino acid substitution, two or more amino acid substitutions, one or more deletions, one or more insertions, or any combination of two or more thereof, in the protein sequence of the activity element. It will be understood that the selection of a suitable mutation in an activity element for the creation of a chimeric molecule will depend on multiple factors and in some embodiments will need to be determined empirically for different proteins.

It should be appreciated that variant activity elements of the invention may have a reduced binding affinity for their receptor(s) without a loss (or without a significant loss) of signal function (e.g., they substantially or completely retain their ability to promote signal transduction when bound to their receptor(s) even though their affinity for the receptor(s) may be significantly reduced). It also should be appreciated that the reduced binding affinity of the variant activity element preferably does not result in a protein element that will not bind to its natural receptor(s), for example, due to steric hindrance or charge repulsion or other negative interaction between the variant activity element and its natural receptor(s), even when the targeting element binds to a target molecule on the same cell (thereby increasing the local concentration of the variant activity element in the vicinity of its natural receptor(s)).

In some embodiments, appropriate levels of reduced binding affinity can be obtained by introducing one or more mutations in charged or hydrophilic amino acids (or amino acids thought to be pointing outward) that have the effect of shortening the side chain of the amino acid(s). According to aspects of the invention, the charged or hydrophilic side chains are likely to be pointing outward and not into the middle of the protein. Reducing the size of the amino acid side chain(s) removes a contact, but does not create steric hindrance that would completely block binding or signaling. In some embodiments, mutants may be created in an activity element, and those that reduce but do not abolish binding may be selected (e.g., using one or more binding and/or activity assays known to one of skill in the art) and used to construct a chimeric activator. In some embodiments, one or more mutation in an activity element may reduce binding of the activity element to one monomer of a receptor and not affect the binding of the activity element to one or more other monomers of the receptor.

Erythropoietin (EPO)

Aspects of the invention relate to targeting EPO. In some embodiments, the activity element of an engineered protein such as a chimeric activator is EPO. In some embodiments, the activity element is a version of EPO that contains one or more mutations relative to wild type human EPO. In some embodiments, the activity element is a portion of EPO that is sufficient to bind to a receptor or a receptor dimer (e.g. , EPO-R, CD 131) and induce an activity in the cell.

EPO has multiple uses. For example, it is administered to kidney failure patients who fail to produce EPO and therefore can suffer from severe anemia. A further use is for cancer patients receiving chemotherapy, who suffer from anemia due to killing of precursors of red blood cells. Current forms of EPO have significant side effects, such as increased death rates in cancer patients due to venous thrombotic events, and in kidney failure patients due to arterial thrombotic events. EPO acts on receptors found on red blood cell precursors to stimulate their survival and red blood cell formation, but also acts on EPO receptors on platelet precursors, endothelial cells and other cells in the blood system to promote thrombosis.

Clinical trials showing enhanced death rates due to modest dosing of EPO have led to 'black box' warnings on the label of EPO products (e.g. Procrit, Epogen) and a drop in sales from a peak of about $10B in the US to about $7B in 2012.

Aspects of the invention relate to a replacement for the current form of EPO.

Described herein is "targeted EPO," which refers to a form of EPO that is administered in the context of an engineered fusion protein (also referred to herein as a chimeric activator) comprising several elements. In some embodiments, targeted EPO includes: human EPO in which one or more mutations have been introduced that reduce its activity; single-chain Fv (scFv) antibody V regions that bind to Glycophorin, such as a humanized form of the 10F7 antibody; and a linker attaching the EPO element to the scFv element.

The effect of this engineered fusion protein is to target EPO activity to red blood cell precursors and away from platelet precursors, endothelial cells, etc. that promote thrombosis. In this way, targeted EPO avoids side effects seen with regular EPO.

EPO plays multiple biological roles by binding to EPO receptors (EPO-R) on diverse cell types, including erythroid progenitors, macrophages, pro-megakaryocytes, cancer cells, and neurons (Jelkmann et al. Ann. Hematol. (2004) 83: 673-686; Bunn Cold Spring Harb. Perspect. Med. (2013) 3). A therapeutic goal of engineered proteins such as chimeric activators described herein that include targeted EPO is to minimize the side effects of EPO by targeting the protein to red blood cell ("RBC") precursors and away from other cell types. Recombinant EPO has been used for two decades to treat forms of anemia associated with endstage renal failure, AIDS, chemotherapy, or hemoglobinopathies (Jelkmann et al. Ann. Hematol. (2004) 83: 673-686; Bunn Cold Spring Harb. Perspect. Med. (2013) 3). Clinical use of EPO has recently decreased due to concerns over the drug's off-target effects: EPO treatment has been linked to tumor recurrence and platelet formation or activation, which may lead to coronary disease or thrombosis (Jelkmann et al. Ann. Hematol. (2004) 83: 673- 686; Bunn Cold Spring Harb. Perspect. Med. (2013) 3). Targeting EPO to RBC precursors should allow higher doses to fully restore RBC levels without increasing the risk of cardiovascular events or cancer progression. In addition, the targeted form of EPO described herein should have an extended serum half-life and reduced immunogenicity relative to existing EPO drugs.

EPO is a hormone that stimulates RBC production by binding to EPO-Rs on RBC precursors, and can cause a variety of other effects via EPO-Rs on other cell types, such as platelet activation and production, expression of tissue factor on endothelial cells, activation of the renin-angiotensin system, neuroprotection against hypoxia, and acceleration of tumor cell growth (Bunn Cold Spring Harb. Perspect. Med. (2013) 3). Recombinant EPO is used to treat anemia, primarily in chronic kidney disease (CKD) and oncology patients. CKD patients usually lack adequate levels of EPO, which is expressed mainly in the kidneys, and are therefore anemic (Hoffman, et al. Hematology: Basic Principles and Practice (Elsevier Health Sciences, 2013)). Before the introduction of EPO, CKD dialysis patients typically received blood transfusions and suffered complications such as iron overload, with significant morbidity (Shander et al. Vox Sanguinis (2009) 97: 185-197). Between 1992 and 2004, however, 90% of U.S. CKD patients undergoing dialysis were instead managed with erythropoiesis stimulating agents (ESAs) (Pisconi et al. Am. J. Kidney Dis. (2004) 44: 94-

111). Cancer patients can also become anemic as a result of myelosuppressive chemotherapy treatment, which kills dividing cells including RBC precursors (Hoffman et al. Hematology (Elsevier Health Sciences, 2013). Transfusion requirements in an oncology setting have been significantly reduced with ESA therapy (relative risk = 0.64) (Bohlius et al. JNCI (2006) 98: 708-714).

Typically when patients are treated with EPO, RBC levels are purposefully not fully restored, since this is associated with an increased rate of heart attack and stroke (Besarab et al. NEJM (1998) 339: 584-590; Fishban et al. Clin. J. Am. Sco. Nephrol. (2007) 2: 1274- 1282). Following FDA approval of ESAs to minimize blood transfusions in CDK or oncology patients, clinical trials were performed to demonstrate long-term improved patient outcomes with ESA therapy (Besarab et al. (NEJM (1998) 339: 584-590; Hedenus et al. (2003) Br. J. Maematol. 122: 394-403; Leyland- Jones et al. J. Clin. Oncol. (2005) 23: 5960- 5972; Drueke et al. NEJM (2006) 355: 2071-2084; Singh et al. NEJM (2006) 355: 2085- 2098; Wright et al. J. Clin. One. (2007) 25: 1027-1032; Smith et al. J. Clin. One. (2008) 26; 1040-1050; Pfeffer et al NEJM (2009) 261: 2019-2032; Stowell et al. Spine (2009) 34: 2479- 2485; Solomon et al NEJM (2010) 363: 1146-1155; Lewis et al. Clin. J. Am. Soc. Nephrol. (2011) 6: 845-855; Untch et al. Annals of Oncology (2011) 22: 1988-1998). These studies ultimately showed that mortality and morbidity, particularly venous thrombotic events, increased with more aggressive treatment, with the greatest effect among cancer patients (Ross et al. Clin. Ther. (2006) 28: 801-832; Glaspy et al. Br. J. Cancer (2010) 102: 301-315; Seliger et al. Kidney International (2011) 80: 288-294).

Several meta-analyses confirm the finding that treatment of oncology patients with ESAs according to current standards {i.e., not fully restoring hematocrit) causes some mortality, primarily through thrombotic events (Ross et al. Clin. Ther. (2006) 28: 801-832; Henke et al. The Lancet (2003) 362: 1255-1260; Bennett et al. JAMA (2008) 299: 914-924; Bohlius et al. The Lancet (2009) 373: 1532-1542; Tonelli et al. Canadian Med. Assoc. (2009) 180: E62-E71). In one meta-analysis of Phase III trials that reported survival and thrombotic events, 52 trials with 13,611 patients and 38 trials with 8,172 patients, respectively, were analyzed by comparing ESAs with placebo or the standard treatment of anemia among cancer patients (Bennett et al. JAMA (2008) 299: 914-924). ESA-treated patients were more susceptible to thrombotic events (7.5% vs. 4.9%) and mortality (hazard ratio = 1.10), but showed greatly improved quality-of life and required fewer transfusions.

ESA labeling has recently been modified to reduce its therapeutic application and warn clinicians about the associated risks (Hess et al. Am. J. Hematol. (2010) 85: 838-843; Rizzo et al. /. Clin. Oncol. (2010) 28: 4996-5010). The resulting reimbursement changes led to increased use of transfusions to treat anemia. For example, in a 2007 study of oncology patients aged 65 years and older, a 31% increase in transfusions was observed (Hess et al. Am. J. Hematol. (2010) 85: 838-843). These data, in combination with evidence linking transfusions with potentially poor patient outcomes and a destabilizing effect on the U.S. blood inventory, are cause for concern (Adamson Transfusion (2009) 49: 824-826; Vekeman et al. Transfusion (2009) 49: 895-902).

ESAs may stimulate adverse thrombotic events by activating EPO-Rs present on platelet progenitors (megakaryocytes), increasing expression of pro-thrombotic proteins, and/or increasing blood pressure (Bunn Cold Spring Harb. Perspect. Med (2013) 3).

Megakaryocytes express EPO-Rs, and EPO appears to act during their maturation stages (Vainchenker et al. Blood (1979) 54: 940-945; Ishibashi et al. J. Clin. Invest. (1987) 79: 286- 289; Degos et al. Textbook of Malignat Hematology (CRC Press, 2004)). Following EPO treatment, both platelet counts and platelet activation are elevated in mice, rats, and dogs (Berridge, M. V. et al. Blood (1998) 72: 970-977; Tsukada, J. et al. Br. J. Haematol (1990) 76: 260-268; McDonald, T.P. et al. Blood (1992) 80: 352-358; Wolf, R.F. et al. Thromb Haemost (1997) 78: 1505-1509; Wolf, R.F. et al. Thromb Haemost (1997) 77: 1020-1024; Loo, M. et al. Blood (1999) 93: 3286-3293; Kirkeby, A. et al. Thromb Haemost (2008)99(4): 720-8). In humans, platelet activation increases 2-4 fold and platelet counts increase 10-25% upon EPO treatment (these markers can be used as surrogates for overall non-erythroid activity of EPO in clinical trials of targeted EPO molecules described herein) (Eschbach, J.W. N. Engl. K. Med (1987) 316: 73-78; Eschbach, J.W. et al. Ann. Intern. Med (1989) 111: 992- 1000; Fabris, F. et al. Pediatr. Nephrol (1991) 5: 225-228; Cases, A. et al. Kidney Int (1992) 42: 668-672; Taylor, J.E. Nephrol. Dial. Transplant (1992) 7: 235-239; Roger, S.D. et al. Nephrol. Dial. Transplant (1993) 8: 213-217; Beguin, Y. et al. Eur. J. Haematol (1994) 53: 265-270; Malyszko, J.S. et al. Int. J. Clin. Lab. Res (1996) 26: 199-202; Dale, G. L. et al. Lancet (1998) 352: 566-567; Beguin, Y. Haemataologica (1999) 84: 541-547; Stohlawetz, P.J. et al. Blood (2000) 95: 2983-2989; Ando, M. et al. Kidney International (2002) 62: 1757- 1763; Chuang, Y.C. Nephrology Dialysis Transplantation (2003) 18: 947-954; Vaziri, N.D. et al. Nephrology Dialysis Transplantation (2003) 18: 947-954; Heinisch, B.B. et al. Platelets (2012) 23: 352-358). Taken together, these data identify plausible mechanisms for the enhancement of thrombotic events in EPO-treated patients and suggest that an EPO derivative targeted to RBC precursors would avoid such events. Moreover, a secondary analysis of at least one study (Szczech, L.A. et al. Kidney International (2008) 74: 791-798) found that patients unable to reach target hematocrits with a high EPO dose had particularly poor outcomes compared to others in the study, suggesting that the mechanism of mortality was not due to increased numbers of RBCs per se.

Targeted EPO described herein is targeted to RBC precursors. This engineered therapeutic should not stimulate platelet production or activation, expression of other pro- thrombotic proteins, blood pressure increases, or activity of EPO-Rs on cancer cells.

Furthermore, the protein should be less immunogenic and have a longer serum half-life than EPO itself. Overall, targeted EPO should reduce the need for transfusions, reduce the frequency of thrombotic events, and improve a patient's quality of life. Accordingly, aspects of the invention relate to methods comprising administering a therapeutically effective amount of a Glycophorin A-binding protein described herein to a subject in need thereof. In some embodiments, the subject has anemia. In some embodiments, the Glycophorin A- binding protein is targeted EPO.

Previous efforts to engineer EPO have focused on increasing its serum half-life and stability by the addition of carbohydrate chains, disulfide rearrangement, dimerization, fusion to antibody fragments or albumin, or PEGylation (Bitonti et al. PNAS (2004) 101: 9763- 9768; Way et al. PEDS (2005) 18: 111-118; Long et al. Experimental hematology (2006) 34: 697-704; Sathyanarayana et al. Blood (2009) 113: 4955-4962; Bouman-Thio et al J. Clin. Pharm. (2013) 48: 1197-1207). These modifications are not expected to affect targeting of EPO to RBCs. No RBC targeted form of EPO has been tested in an animal models previously.

A problem with targeted protein fusions has been that their design often disregards quantitative information about binding kinetics and structural information about protein ligands and cell surface receptors (Chen et al Mol Pharmaceutics (2011) 8: 457-465;

Atkinson et al. Principles of Clin. Pharm. (2012); Robinson-Mosher at al. Chaos (2013) 23: 025110-13). A typical targeted ligand or activity element ("AE") can be fused to a targeting element ("TE"), such as an antibody, whose binding to a cell should dominate the overall binding reaction. However, most naturally occurring ligands bind quite well to their receptors; binding proceeds at the diffusion-limited on-rate (k_on ~ 10⁶ M^'V¹), and the off-rate is typically slow enough that other processes, such as receptor-mediated endocytosis, occur more quickly and make the off-rate irrelevant (Robinson-Mosher at al. Chaos (2013) 23: 025110-13). Thus, simply adding a TE does not eliminate off-target binding, and often does not accelerate binding to desired target cells, because the on-rate of the ligand and TE are similar. A further difficulty with protein fusions has been achieving simultaneous binding of the AE and TE to their respective receptors on the same cell (Chen et al Mol Pharmaceutics (2011) 8: 457-465; Robinson-Mosher at al. Chaos (2013) 23: 025110-13; Cironi et al. J. Biol. Chem. (2008) 283: 8469-8476; Taylor et al. PEDS (2010) 23: 251-260). Information from crystal structures of ligand-receptor complexes can be integrated into protein fusion designs, such that both the AE and TE are oriented to permit simultaneous binding.

Cironi, et al. addressed the quantitative binding issue by introducing a mutation in the AE of a targeted fusion protein (FIG. 1) (Cironi et al. J. Biol. Chem. (2008) 283: 8469-8476). Initial binding was driven by the TE, and despite the mutation, subsequent binding of the AE occurred due its high local concentration at the cell surface. The mutation contributes an additional 10-20 fold to specificity. Robinson-Mosher, et al (Robinson-Mosher at al. Chaos (2013) 23: 025110-13), recently addressed the spatial issues around simultaneous binding to two receptors.

Taylor, et al. applied these principles to design a targeted EPO molecule that specifically binds RBC precursors and not, for example, platelet precursors, as shown in FIG. 2, which is adapted from Figure 1 of Taylor et al (Taylor et al. PEDS (2010) 23: 251-260). The protein was targeted using a single-chain variable fragment (scFv) derived from the antibody 10F7 (Bigbee et al. Mol. Immunol (1983) 20: 1353-1362.), which binds GYP A (Okumura et al. Blood (2992) 80: 642-650; Southcott et al. Blood (1999) 93: 4425-4435). The EPO element (mutEPO) had the mutation R150A, which reduces its affinity for its receptor (wild-type K_d = 1 nM), but does not affect activity or folding (Elliott et al. Blood (1997) 89P: 493-502; Cheetham et al. Nat. Struct. Mol. Biol. (1998) 5: 861-866). The flexible linker allowed both elements of the protein fusion to bind simultaneously to their receptors. This molecule showed an overall targeting effect of 10-30 fold relative to EPO alone. This improvement in specificity may translate into an improved EC₅₀ in vivo for RBC production compared to induction of pro-thrombotic effects. Furthermore, its molecular weight is > 70 kDa, which should limit renal clearance and extend serum half-life (Atkinson et al. Principles of Clin. Pharma. (2012)). Clearance should also be reduced by the fact that it binds to GYPA on mature RBCs, which should thus act as a reservoir for the drug (Kontos et al. Mol.

Pharmaceutics (2010) 7: 2141-2147) (mature RBCs do not express EPO-R or perform receptor-mediated endocytosis (Agre, P. RedBlood Cell Membranes (CRC Press, 1989)). Finally, the protein should be non-immunogenic; the immunogenicity of RBC-bound proteins is strongly suppressed during RBC turnover, when RBC proteins are internalized by macrophages and presented in a non-immunosuppressive manner (Kontos et al. Mol.

Pharmacuetics (2010) 7: 2141-2147; Kontos et al. Chem. Soc. Rev. (2012) 41: 2686).

The 10F7-mutEPO molecule described in Taylor, et al. (Taylor et al. PEDS (2010) 23: 251-260) is not appropriate for human use. By contrast, described herein are engineered proteins that include humanized versions of the 10F7 antibody V regions, generating an engineered proteinthat is appropriate for human use and that can be used to deliver targeted EPO.

Mutations in Activity Element

Aspects of the invention relate to the affinity of the activity element for its receptor, and the ratio of the affinity of the activity element for its receptor relative to the affinity of the targeting element for its receptor. In some embodiments mutated activity elements will be selected such that the activity element has a lower affinity for its receptor than the affinity of the targeting element for its receptor, thus allowing the targeting element to control which cells are bound by the chimeric molecule. In some embodiments the affinity ratio may be 1/2, 1/5, 1/10, 1/25, 1/50, 1/75, 1/100, 1/500, 1/1000, or an intermediate value, or a smaller value. In some embodiments, the affinity ratio is relative to the affinity of the targeting element to its receptor. In some embodiments, the affinity ratio is relative to the affinity of an activity element that does not comprise the mutation(s). In some embodiments, the affinity ratio is relative to the affinity of an activity element that is not in a chimeric protein comprising the targeting element. Additional aspects of the invention relate to the on-rates, off -rates, and dissociation constants for the binding of the activity element to its receptor and the targeting element to its receptor. In some embodiments mutated activity elements will be selected wherein one or both of the following conditions are met: the on-rate of the activity element is lower than that of the targeting element, or if the off-rates are faster than the internalization rate, then the equilibrium constant (K_on/K_0ff) of the targeting element is higher than that of the activity element.

Another criterion that may be involved in selecting mutated activity elements is an evaluation of the activity level of the activity element alone, and the activity level of the activity element when it is linked to the targeting element. In some embodiments mutated activity elements may be selected that have reduced activity or are inactive alone, relative to a wild type form of the protein, due to their reduced binding efficiency, but are active when linked to a targeting element, because the binding of a targeting element to its receptor at the cell surface increases the local concentration of the activity element in the vicinity of its receptor at the cell surface. It should be appreciated that appropriate activity levels for a given protein and a given mutation can be determined empirically. When the activity element is linked to a targeting element, the increased local concentration of the activity element at the cell surface allows it to be active and bind to its receptor.

In some embodiments, the activity element is a version of EPO that contains one or more mutations relative to wild type human EPO. In some embodiments, the activity element is a version of EPO that contains one or more mutations at a position selected from Argl50, Ala30, His32, Pro87, Trp88, Pro90, Arg53, and Glu55. In some embodiments, the EPO contains a mutation at Arg 150 and any one or more additional mutations at a position selected from Ala30, His32, Pro87, Trp88, Pro90, Arg53, and Glu55. In some embodiments, the activity element is a version of EPO that contains one or more mutations at a position selected from Argl50A, Ala30Asn, His32Thr, Pro87Val, Trp88Asn, Pro90Thr, Arg53Asn, and Glu55Thr. In some embodiments, EPO contains the mutation R150A. In some embodiments, the EPO contains the mutation Argl50Ala and any one or more additional mutations at a position selected from Ala30Asn, His32Thr, Pro87Val, Trp88Asn, Pro90Thr, Arg53Asn, and Glu55Thr.

In some embodiments, one or more mutations may be made in EPO to increase the number of N-linked glycosylation sites in EPO. In some embodiments, the mutations generating additional N-linked glycosylation sites in EPO do not sterically interfere with EPO binding to one or more EPO-R. In some embodiments, the mutations generating additional N-linked glycosylation sites reduce the on-rate of EPO. In some embodiments, the EPO contains at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or at least 10 additional N-linked glycosylation sites. Such an EPO may be referred to as "highly glycosylated." In some embodiments, the EPO contains a mutation at any one or more positions selected from Ala30, His32, Pro87, Trp88, Pro90, Arg53, and/or Glu55. In some embodiments, the EPO contains one or more mutations selected from Ala30Asn, His32Thr, Pro87Val, Trp88Asn, Pro90Thr, Arg53Asn, and/or Glu55Thr. Examples of highly glycosylated engineered proteins/chimeric activators are provided by SEQ ID NOs: 64 and 65.

In some embodiments, one or more additional mutations may be made in EPO to reduce the interaction between EPO and at least one EPO receptor (EPO-R). In some embodiments, the EPO contains one or more additional mutations in one or more surfaces of the protein that interact with a receptor(s). In some embodiments, the EPO contains one or more additional mutations in one or more surfaces of the protein that interact with a receptor(s) such that the binding affinity of the mutated EPO to the EPO-R is reduced relative to EPO that does not contain the mutation(s). In some embodiments, one or more mutations are made in helix 3, helix 1, or nearby residues of EPO. In some embodiments, the EPO contains a mutations at one or more positions selected from residues Leu5, Argl4, Leu93, Asp96, Lys97, SerlOO, Argl03, Serl04, Thrl07, and Leul08. In some embodiments, the EPO contains one or more mutations selected from Leu5Val, Leu5Ala, Argl4Lys,

Argl4Met, Argl4Ile, Arg 14Ala, Argl4Ser, Leu93Ala, Asp96Arg, Asp96Lys, Lys97Ala, Lys97Ser, SerlOOArg, Argl03Ser, Argl03Met, Argl03Ile, Argl03Lys, Argl03Glu,

Serl04Ala, Serl04Gly, Thrl07Ser, Thrl07Ala, Leul08Val, Leul08Ala, and Leul08Ser. In some embodiments, the one or more mutations in EPO reduce binding to the EPO-R relative to the binding of EPO in the absence of the targeting element (scFv). In some embodiments, the one or more mutations reduce binding of EPO to a homodimer EPO-R, in the absence of the targeting element, by at least 1.5- fold, 2-fold, 3-fold, 4-fold or at least 5-fold relative to binding of wild-type EPO to the homodimer EPO-R. In some embodiments, the one or more mutations in EPO do not have an effect on binding of the EPO to the heterodimer EPO- R/CD131 in the absence of the targeting element. In some embodiments, the one or more mutations in EPO have less than a 5-fold, 4-fold, 3-fold, 2-fold, or less than a 1.5-fold effect on binding of the EPO to the heterodimer EPO-R/CD 131.

In some embodiments an engineered protein includes a mutant EPO (e.g., R150A) attached to a GYPA targeting element (e.g., including an antibody polypeptide having a sequence of any one of SEQ ID Nos.: 2-5, 7-9, or 52-60). In some embodiments, the mutant EPO is C-terminal relative to the GYPA targeting element. In some embodiments, the N- terminus of the mutant EPO is attached directly to the C-terminus of the GYPA binding antibody polypeptide. In some embodiments, the N-terminus of the mutant EPO is attached to the C-terminus of the GYPA binding antibody polypeptide via a linker peptide. It should be appreciated that antibodies associated with the invention can have any combination of CDR sequences, for example any combination of the CDR sequences listed in Table 1. For example, any of the sequences listed as heavy chain CDR1 sequences (SEQ ID NOs: 11-16) can be combined with any of the sequences listed as heavy chain CDR2 sequences (SEQ ID NOs: 18-22) and/or with any of the sequences listed as heavy chain CDR3 sequences (SEQ ID NOs: 24-28). Any of the sequences listed as light chain CDR1 sequences (SEQ ID NOs: 30-33) can be combined with any of the sequences listed as light chain CDR2 sequences (SEQ ID NOs: 35 or 36) and/or with any of the sequences listed as light chain CDR3 sequences (SEQ ID NOs: 37 or 38).

Aspects of the invention relate to the use of EPO moieties that resemble human EPO as much as possible in primary sequence and three-dimensional structure. In some embodiments, the mutations in the EPO moiety are kept to a minimum. For example, in some embodiments, the EPO moieties are at least 90% identical to human EPO, and more preferably at least 95% identical to human EPO. In some embodiments, the amino acid sequence of the EPO moiety is at least 90%, 91%, 92,%, 93%, 94%, 95%, 96%, 97%, 98%, or at least 99% identical to the amino acid sequence of human EPO. As human EPO has 166 amino acids in its mature form without C-terminal processing, an EPO moiety that is 90% and 95% identical to human EPO has at most 16 and 8 amino acids, respectively, that differ from human EPO.

Linkers

In some embodiments, targeting and activity elements are attached by a linker. The linker can be any amino acid sequence that allows the simultaneous binding of the activity element and the targeting element to their receptors on a given cell surface. Typically the linker is a non-folding, protease-resistant polypeptide segment that permits the targeting element and the activity element to bind to their cell-surface receptors, typically on the same cell surface, at the same time. Linkers may be of varied length {e.g., about 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100 amino acids long or intermediate or longer lengths) and/or amino acid composition. Linkers may be flexible or rigid. In some embodiments, the targeting element may also comprise a linker, such as a linker between the variable region of a heavy chain and the variable region of a light chain of a scFv. In some embodiments, the length of the linker between the targeting and activity elements will be optimized for maximizing binding efficiency of the activity element to the second binding site when the targeting element is bound to the first binding site. Optimal linker length for a linker between the targeting and activity elements a given chimeric activator can be determined at least in part using models disclosed herein. In optimizing linker length, multiple factors will be considered including, but not limited to: the

concentration of the first and second binding sites, the sizes of the activity element and targeting element, the minimum distance between the binding sites of the targeting element and the activity element, and the affinities of the activity element and the targeting element for the first and second binding sites. In some embodiments a linker length will be chosen that is at least equal to the minimum distance between the first and second binding sites on the cell surface. In other embodiments a linker length will be chosen that is at least twice the minimum distance between the first and second binding sites on the cell surface. In some embodiments a linker length will be chosen that is at most five-six times the minimum distance between the first and second binding sites on the cell surface. It will be understood that optimal linker length will vary depending on the specific proteins selected as activity elements and targeting elements. In some embodiments a protein which is blocked at the N- or C-terminus may be selected as an activity element or a targeting element. In these embodiments a linker may be added elsewhere on the protein.

In some embodiments, the length of the linker between the targeting and activity elements is chosen based in part on the known or surmised heights from the cell surface of the targeting element and activity element when each is bound to its receptor. For example, when EGF is bound to EGFR, its N- and C-termini are about 90 Angstroms from the cell surface, and when interferon alpha is bound to its receptor, its N- and C-termini are about 75 Angstroms from the cell surface, so that a linker of at least 15 Angstroms must be used.

The invention also provides methods for calculating an optimal length of a linker between the targeting and activity elements for a chimeric protein with two elements that simultaneously bind to a cell surface, based on the density of the receptors for each of the two elements on a cell surface and the difference in height from the cell surface of the termini of the two elements when bound to their receptors. The method optionally involves knowledge of the on-rates and off-rates of each element for its receptor. The method optionally involves calculating, as linker length is increased, the trade-off on the one hand between the increasing volume that can be occupied by one element when the other is bound to its receptor on a cell surface (which negatively impacts on a second binding event), and the increasing portion of cell surface area that may be sampled by one element when the other is bound to its receptor on that cell surface (which positively impacts on a second binding event). For many situations, the density of a targeting element receptor is more than 10,000 per cell and the difference in height from the cell membrane of the bound elements is less than about 50 Angstroms, and in such situations a linker of about 25 to 70 amino acids may be optimal, with a linker length of about 30 to 40 amino acids often being more optimal. In some embodiments, the linker is flexible. In some embodiments, the linker is a glycine serine linker. In some embodiments, the linker comprises glycine, serine, and glutamic acid residues. An example of a linker is provided by amino acid residues 243-281 of SEQ ID NO: 39.

Polynucleotides

Nucleic acids of the invention include isolated or recombinant nucleic acids comprising coding sequences for chimeric activators or components thereof described herein. The nucleic acids may be on vectors. The nucleic acids may be in host cells. It should be appreciated that the nucleic acids also may include regulatory sequences for transcription (e.g., promoters, activators, terminators, etc.) and translation (e.g., ribosome binding sequences, IRES elements, terminators, etc.) of the chimeric proteins in addition to sequences required for vector replication and/or selection and/or packaging in host cells (e.g., prokaryotic or eukaryotic host cells, including, but not limited to bacterial, mammalian, yeast, insect, and/or other host cells). Accordingly, aspects of the invention relate to recombinant vectors which include one or more nucleic acids of the invention, as well as host cells containing the vectors or which are otherwise engineered to contain or express nucleic acids or polypeptides of the invention, and methods of making such vectors and host cells and their use in production of polypeptides of the invention by recombinant or synthetic techniques. In some embodiments, targeting and activity elements may be isolated (e.g. expressed and purified) independently and linked via a synthetic linker.

Aspects of the invention relate to nucleic acid sequences encoding a Glycophorin A- binding protein, comprising a sequence selected from the group consisting of SEQ ID NOs: 41-49, and 73 .

In some embodiments, the polynucleotides of the invention are joined to a vector (e.g., a cloning or expression vector). The vector may be, for example, a phage, plasmid, or viral vector. Viral vectors may be replication competent or replication defective. In the latter case, viral propagation generally will occur only in complementing host cells. The polynucleotides may be joined to a vector containing a selectable marker for propagation in a host. Introduction of the vector construct into the host cell can be effected by techniques known in the art which include, but are not limited to, calcium phosphate transfection, DEAE-dextran mediated transfection, cationic lipid-mediated transfection, electroporation, transduction, infection or other methods. Such methods are described in many standard laboratory manuals, such as Davis et al. (1986) Basic Methods In Molecular Biology.

The polypeptides of the invention or fragments thereof may be provided in

substantially pure form, that is to say free, to a substantial extent, from other proteins. Thus, a polypeptide may be provided in a composition in which it is the predominant component present (i.e., it is present at a level of at least 50%; preferably at least 75%, at least 90%, or at least 95%; when determined on a weight/weight basis excluding solvents, carriers, or coupling agents).

Polypeptides of the invention can be recovered and purified from recombinant cell cultures or organisms by well-known methods including, ammonium sulphate or ethanol precipitation, acid extraction, anion or cation exchange chromatography, phosphocellulose chromatography, affinity chromatography, hydrophobic interaction chromatography, hydroxylapatite chromatography, molecular sieving chromatography, centrifugation methods, electrophoresis methods and lectin chromatography. In one embodiment, chimeric activator proteins are expressed in and secreted from the yeast Pichia pastoris. In one embodiment, a combination of these methods is used. In another embodiment, high performance liquid chromatography is used. In a further embodiment, an antibody which specifically binds to a polypeptide of the invention can be used to deplete a sample comprising a polypeptide of the invention of the polypeptide or to purify the polypeptide. Techniques well-known in the art, may be used for refolding to regenerate native or active conformations of the polypeptides of the invention when the polypeptides have been denatured during isolation and or purification, should such be desired.

Polypeptides of the invention (and their corresponding DNA and RNA coding sequences) may comprise sequences, for example, epitope or affinity tags, that may aid in isolation or purification of the protein. However, they need not do so. For example, proteins of the invention may be purified through use of reagents that bind myc and/or His tags, but the invention also contemplates embodiments in which the proteins lack the myc and/or His tags, or which utilize different epitope or affinity tags.

Chimeric activator nucleic acids and/or polypeptides may be screened, such as in high-throughput screening of a library of chimeric activators to select for a chimeric activator of the desired characteristics, or may be screened individually. For example, measurement of signal transduction in mammalian cells using Western blots that detect phosphorylation at specific sites may be used to screen for chimeric activators of a desired signal transduction activity.

Administration

Humanized anti-GYPA antibodies described herein have multiple applications.

Several non-limiting examples include targeting of EPO to red blood cell precursors, targeting of proteins with a short serum half-life to mature red blood cells to extend serum half-life, and targeting of immunogenic proteins to red blood cells to reduce their

immunogenicity .

In some embodiments, humanized anti-GYPA antibodies are components of engineered proteins such as chimeric activators in which the targeting element targets the activity element to a specific cell or cell type such as an RBC precursor. GYPA is present on mature RBCs at ~10⁵-10⁶ molecules per cell, while its expression is 10-100 fold less on erythroid precursors (Merry et al. Biochem J. (1986) 233: 93-98; Loken et al. Blood (1987) 69: 255-263). It is expected that GYPA on mature RBCs will act as a sink for the engineered protein/chimeric activator protein (Taylor et al. PEDS (2010) 23: 251-260). Binding to RBCs should significantly extend the serum half-life of the fusion protein (Kontos et al. Mol.

Pharma. (2010) 7: 2141-2147; Kontos et al. Chem. Soc. Revi. (2012) 41: 2686; Konotos et al. PNAS (2013) 110: E60-8; Holt et al. PEDS (2008) 21: 283-288).

Chimeric activator nucleic acids and/or polypeptides may be used in gene therapy or in pharmaceutical compositions. Chimeric activator nucleic acids and/or polypeptides may be used in treatment of a subject or patient, and may be used in combination with other therapies. Use of chimeric activator nucleic acids and/or polypeptides may be indicated by a diagnostic or theranostic, for example, a biopsy, for the appropriate presence of receptors for the targeting and/or activity element.

Aspects of the invention relate to the use of Glycophorin-binding proteins/chimeric activator proteins in the treatment of disorders. In some embodiments, Glycophorin-binding proteins/chimeric activator proteins are used to treat subjects who have or are at risk of having anemia. In some embodiments, Glycophorin-binding proteins/chimeric activator proteins are used to treat subjects who have or are at risk of having kidney failure. As used herein, the term "subject" refers to a human or non-human mammal or animal. Non-human mammals include livestock animals, companion animals, laboratory animals, and non-human primates. Non-human subjects also specifically include, without limitation, chickens, horses, cows, pigs, goats, dogs, cats, guinea pigs, hamsters, mink, and rabbits. In some embodiments of the invention, a subject is a patient. As used herein, a "patient" refers to a subject who is under the care of a physician or other health care worker, including someone who has consulted with, received advice from or received a prescription or other recommendation from a physician or other health care worker. A patient is typically a subject having or at risk of having anemia. The term "treatment" or "treating" is intended to refer to prophylaxis, amelioration, prevention and/or cure of a condition (e.g., anemia, kidney failure). Treatment after a condition (e.g., anemia, kidney failure) that has started aims to reduce, ameliorate or altogether eliminate the condition, and/or its associated symptoms, or prevent it from becoming worse. Treatment of subjects before a condition (e.g., anemia, kidney failure) has started (i.e., prophylactic treatment) aims to reduce the risk of developing the condition and/or lessen its severity if the condition does develop. As used herein, the term "prevent" refers to the prophylactic treatment of a subject who is at risk of developing a condition (e.g., anemia, kidney failure) resulting in a decrease in the probability that the subject will develop the disorder, and to the inhibition of further development of an already established disorder.

Compositions of the invention may be administered in effective amounts. An effective amount is a dosage of the composition of the invention sufficient to provide a medically desirable result. An effective amount means that amount necessary to delay the onset of, inhibit the progression of or halt altogether the onset or progression of the particular condition (e.g., anemia, kidney failure) being treated. An effective amount may be an amount that reduces one or more signs or symptoms of the condition (e.g., anemia, kidney failure). When administered to a subject, effective amounts will depend, of course, on the particular condition being treated (e.g., the anemia, kidney failure), the severity of the condition, individual subject parameters including age, physical condition, size and weight, concurrent treatment, frequency of treatment, and the mode of administration. These factors are well known to those of ordinary skill in the art and can be addressed with no more than routine experimentation.

Actual dosage levels of active ingredients in the compositions of the invention can be varied to obtain an amount of the composition of the invention that is effective to achieve the desired therapeutic response for a particular subject, compositions, and mode of

administration. The selected dosage level depends upon the activity of the particular composition, the route of administration, the severity of the condition being treated, the condition, and prior medical history of the subject being treated. However, it is within the skill of the art to start doses of the composition at levels lower than required to achieve the desired therapeutic effort and to gradually increase the dosage until the desired effect is achieved. In some embodiments, lower dosages would be required for combinations of multiple compositions than for single compositions.

It should be appreciated that an optimized dosage may be based, at least in part, on the residual non-specific toxicity of the activity element in an engineered protein/chimeric activator, the concentration of target molecules on the target cells, the binding affinity of the targeting element of the chimeric activator for the target molecules, and the length and flexibility of the linker. The amount of compound added should be sufficient to obtain a concentration of chimeric activator on the target cells that is sufficient to promote the desired response in the target cells. However, the amount should not be higher than necessary to avoid exposing non-target cells with levels of variant activity element that would be sufficient to activate the non-target cells. It should be appreciated that the optimal amount may be different for different chimeric activators (e.g., with different relative binding of targeting and activity elements, etc.). Suitable dosages and dosage regimens for a subject will be evident or may be determined by one of skill in the art, such as a medical practitioner. In some embodiments, the subject is administered one or more doses of a chimeric activator between about 0.01-100 mcg/kg, 0.1- 10 mcg/kg, or 0.3-3 mcg/kg. In some embodiments, the subject is administered one or more doses of about 1 mcg/kg. In some embodiments, the one or more doses of the chimeric activator are administered once per two weeks or once per four weeks.

The compositions of the invention, can be administered to a subject by any suitable route. For example, the compositions can be administered orally, including sublingually, rectally, parenterally, intracisternally, intravaginally, intraperitoneally, topically and transdermally (as by powders, ointments, or drops), bucally, or nasally. The term

"parenteral" administration as used herein refers to modes of administration other than through the gastrointestinal tract, which include intravenous, intramuscular, intraperitoneal, intrasternal, intramammary, intraocular, retrobulbar, intrapulmonary, intrathecal,

subcutaneous and intra-articular injection and infusion. Surgical implantation also is contemplated, including, for example, embedding a composition of the invention in the body such as, for example, in the brain, in the abdominal cavity, under the splenic capsule, or in the cornea.

Dosage forms for topical administration of a composition of this invention include powders, sprays, ointments, and inhalants as described herein. The composition is mixed under sterile conditions with a pharmaceutically acceptable carrier and any needed preservatives, buffers, or propellants that may be required.

Pharmaceutical compositions of the invention for parenteral injection comprise pharmaceutically acceptable sterile aqueous or nonaqueous solutions, dispersions, suspensions, or emulsions, as well as sterile powders for reconstitution into sterile injectable solutions or dispersions just prior to use. Examples of suitable aqueous and nonaqueous carriers, diluents, solvents, or vehicles include water, ethanol, polyols (such as glycerol, propylene glycol, polyethylene glycol, and the like), and suitable mixtures thereof, vegetable oils (such as olive oil), and injectable organic esters such as ethyl oleate. Proper fluidity can be maintained, for example, by the use of coating materials such as lecithin, by the maintenance of the required particle size in the case of dispersions, and by the use of surfactants.

These compositions also can contain adjuvants such as preservatives, wetting agents, emulsifying agents, and dispersing agents. Prevention of the action of microorganisms can be ensured by the inclusion of various antibacterial and antifungal agents, for example, paraben, chlorobutanol, phenol sorbic acid, and the like. It also may be desirable to include isotonic agents such as sugars, sodium chloride, and the like. Prolonged absorption of the injectable pharmaceutical form can be brought about by the inclusion of agents that delay absorption, such as aluminum monostearate or gelatin.

In some cases, in order to prolong the effect of the composition, it is desirable to slow the absorption of the composition from subcutaneous or intramuscular injection. This result can be accomplished by the use of a liquid suspension of crystalline or amorphous materials with poor water solubility. The rate of absorption of the composition then depends upon its rate of dissolution, which, in turn, may depend upon crystal size and crystalline form.

Alternatively, delayed absorption of a parenterally administered composition from is accomplished by dissolving or suspending the composition in an oil vehicle.

Injectable depot forms are made by forming microencapsule matrices of the composition in biodegradable polymers such a polylactide-polyglycolide. Depending upon the ratio of composition to polymer, and the nature of the particular polymer employed, the rate of composition release can be controlled. Examples of other biodegradable polymers include poly(orthoesters) and poly(anhydrides). Depot injectable formulations also are prepared by entrapping the drug in liposomes or microemulsions that are compatible with body tissue.

The injectable formulations can be sterilized, for example, by filtration through a bacterial- or viral-retaining filter, or by incorporating sterilizing agents in the form of sterile solid compositions, which can be dissolved or dispersed in sterile water or other sterile injectable medium just prior to use.

The invention provides methods for oral administration of a pharmaceutical composition of the invention. Oral solid dosage forms are described generally in

Remington's Pharmaceutical Sciences, 18th Ed., 1990 (Mack Publishing Co. Easton Pa. 18042) at Chapter 89. Solid dosage forms for oral administration include capsules, tablets, pills, powders, troches or lozenges, cachets, pellets, and granules. Also, liposomal or proteinoid encapsulation can be used to formulate the present compositions (as, for example, proteinoid microspheres reported in U.S. Pat. No. 4,925,673). As is known in the art, liposomes generally are derived from phospholipids or other lipid substances. Liposomes are formed by mono- or multi-lamellar hydrated liquid crystals that are dispersed in an aqueous medium. Any nontoxic, physiologically acceptable, and metabolizable lipid capable of forming liposomes can be used. The present compositions in liposome form can contain, in addition to a compound of the present invention, stabilizers, preservatives, excipients, and the like. The preferred lipids are the phospholipids and the phosphatidyl cholines (lecithins), both natural and synthetic. Methods to form liposomes are known in the art. See, for example, Prescott, Ed., Methods in Cell Biology, Volume XIV, Academic Press, New York, N.Y. (1976), p 33, et seq. Liposomal encapsulation may include liposomes that are derivatized with various polymers (e.g., U.S. Pat. No. 5,013,556). In general, the formulation includes a composition of the invention and inert ingredients which protect against degradation in the stomach and which permit release of the biologically active material in the intestine.

In such solid dosage forms, the composition is mixed with, or chemically modified to include, a least one inert, pharmaceutically acceptable excipient or carrier. The excipient or carrier preferably permits (a) inhibition of proteolysis, and (b) uptake into the blood stream from the stomach or intestine. In one embodiment, the excipient or carrier increases uptake of the composition of the invention, overall stability of the composition, and/or circulation time of the composition in the body. Excipients and carriers include, for example, sodium citrate, or dicalcium phosphate, and/or (a) fillers or extenders such as starches, lactose, sucrose, glucose, cellulose, modified dextrans, mannitol, and silicic acid, as well as inorganic salts such as calcium triphosphate, magnesium carbonate and sodium chloride, and commercially available diluents such as FAST-FLO^®, EMDEX^®, STA-RX 1500^®,

EMCOMPRESS^® and AVICEL^®, (b) binders such as, for example, methylcellulose ethylcellulose, hydroxypropyhnethyl cellulose, carboxymethylcellulose, gums (e.g., alginates, acacia), gelatin, polyvinylpyrrolidone, and sucrose, (c) humectants, such as glycerol, (d) disintegrating agents, such as agar-agar, calcium carbonate, potato or tapioca starch, alginic acid, certain silicates, sodium carbonate, starch including the commercial disintegrant based on starch, EXPLOTAB^®, sodium starch glycolate, AMBERLITE^®, sodium

carboxymethylcellulose, ultramylopectin, gelatin, orange peel, carboxymethyl cellulose, natural sponge, bentonite, insoluble cationic exchange resins, and powdered gums such as agar, karaya or tragacanth; (e) solution retarding agents such a paraffin, (f) absorption accelerators, such as quaternary ammonium compounds and fatty acids including oleic acid, linoleic acid, and linolenic acid (g) wetting agents, such as, for example, cetyl alcohol and glycerol monosterate, anionic detergent surfactants including sodium lauryl sulfate, dioctyl sodium sulfosuccinate, and dioctyl sodium sulfonate, cationic detergents, such as

benzalkonium chloride or benzethonium chloride, nonionic detergents including

lauromacrogol 400, polyoxyl 40 stearate, polyoxyethylene hydrogenated castor oil 10, 50 and 60, glycerol monostearate, polysorbate 40, 60, 65, and 80, sucrose fatty acid ester, methyl cellulose and carboxymethyl cellulose; (h) absorbents, such as kaolin and bentonite clay, (i) lubricants, such as talc, calcium stearate, magnesium stearate, solid polyethylene glycols, sodium lauryl sulfate, polytetrafluoroethylene (PTFE), liquid paraffin, vegetable oils, waxes, CARBOWAX^® 4000, CARBOWAX^® 6000, magnesium lauryl sulfate, and mixtures thereof; (]) glidants that improve the flow properties of the drug during formulation and aid rearrangement during compression that include starch, talc, pyrogenic silica, and hydrated silicoaluminate. In the case of capsules, tablets, and pills, the dosage form also can comprise buffering agents.

Solid compositions of a similar type also can be employed as fillers in soft and hard- filled gelatin capsules, using such excipients as lactose or milk sugar, as well as high molecular weight polyethylene glycols and the like.

The solid dosage forms of tablets, dragees, capsules, pills, and granules can be prepared with coatings and shells, such as enteric coatings and other coatings well known in the pharmaceutical formulating art. They optionally can contain opacifying agents and also can be of a composition that they release the active ingredients(s) only, or preferentially, in a part of the intestinal tract, optionally, in a delayed manner. Exemplary materials include polymers having pH sensitive solubility, such as the materials available as EUDRAGIT^® Examples of embedding compositions that can be used include polymeric substances and waxes. The composition of the invention also can be in microencapsulated form, if appropriate, with one or more of the above-mentioned excipients.

Liquid dosage forms for oral administration include pharmaceutically acceptable emulsions, solutions, suspensions, syrups, and elixirs. In addition to the composition of the invention, the liquid dosage forms can contain inert diluents commonly used in the art, such as, for example, water or other solvents, solubilizing agents and emulsifiers, such as ethyl alcohol, isopropyl alcohol ethyl carbonate ethyl acetate, benzyl alcohol, benzyl benzoate, propylene glycol, 1,3-butylene glycol, dimethyl formamide, oils (in particular, cottonseed, groundnut, corn, germ, olive, castor, and sesame oils), glycerol, tetrahydrofurfuryl alcohol, polyethylene glycols, fatty acid esters of sorbitan, and mixtures thereof.

Besides inert diluents, the oral compositions also can include adjuvants, such as wetting agents, emulsifying and suspending agents, sweetening, coloring, flavoring, and perfuming agents. Oral compositions can be formulated and further contain an edible product, such as a beverage.

Suspensions, in addition to the composition of the invention, can contain suspending agents such as, for example ethoxylated isostearyl alcohols, polyoxyethylene sorbitol and sorbitan esters, microcrystalline cellulose, aluminum metahydroxide, bentonite, agar-agar, tragacanth, and mixtures thereof.

Also contemplated herein is pulmonary delivery of the composition of the invention. The composition is delivered to the lungs of a mammal while inhaling, thereby promoting the traversal of the lung epithelial lining to the blood stream. See, Adjei et al., Pharmaceutical Research 7:565-569 (1990); Adjei et al., International Journal of Pharmaceutics 63: 135-144 (1990) (leuprolide acetate); Braquet et al., Journal of Cardiovascular Pharmacology 13 (suppl.5): s.143-146 (1989)(endothelin-l); Hubbard et al., Annals of Internal Medicine 3:206- 212 (1989)(al-antitrypsin); Smith et al., J. Clin. Invest. 84: 1145-1146 (1989) (a 1 -proteinase); Oswein et al., "Aerosolization of Proteins," Proceedings of Symposium on Respiratory Drug Delivery II, Keystone, Colorado, March, 1990 (recombinant human growth hormone); Debs et al., The Journal of Immunology 140:3482-3488 (1988) (interferon-γ and tumor necrosis factor a) and Platz et al., U.S. Pat. No. 5,284,656 (granulocyte colony stimulating factor). Contemplated for use in the practice of this invention are a wide range of mechanical devices designed for pulmonary delivery of therapeutic products, including, but not limited to, nebulizers, metered dose inhalers, and powder inhalers, all of which are familiar to those skilled in the art. Some specific examples of commercially available devices suitable for the practice of the invention are the ULTRA VENT^® nebulizer, manufactured by Mallinckrodt, Inc., St. Louis, MO; the ACORN II^® nebulizer, manufactured by Marquest Medical Products, Englewood, CO.; the VENTOL^® metered dose inhaler, manufactured by Glaxo Inc., Research Triangle Park, N.C.; and the SPINHALER^® powder inhaler, manufactured by Fisons Corp., Bedford, MA.

All such devices require the use of formulations suitable for the dispensing of a composition of the invention. Typically, each formulation is specific to the type of device employed and can involve the use of an appropriate propellant material, in addition to diluents, adjuvants, and/or carriers useful in therapy.

The composition may be prepared in particulate form, preferably with an average particle size of less than 10 μιη, and most preferably 0.5 to 5 μιη, for most effective delivery to the distal lung.

Carriers include carbohydrates such as trehalose, mannitol, xylitol, sucrose, lactose, and sorbitol. Other ingredients for use in formulations may include lipids, such as DPPC, DOPE, DSPC and DOPC, natural or synthetic surfactants, polyethylene glycol (even apart from its use in derivatizing the inhibitor itself), dextrans, such as cyclodextran, bile salts, and other related enhancers, cellulose and cellulose derivatives, and amino acids.

In addition, the use of liposomes, microcapsules or microspheres, inclusion complexes, or other types of carriers is contemplated.

Formulations suitable for use with a nebulizer, either jet or ultrasonic, typically comprise a composition of the invention dissolved in water at a concentration of about 0.1 to 25 mg of biologically active protein per mL of solution. The formulation also can include a buffer and a simple sugar (e.g., for protein stabilization and regulation of osmotic pressure). The nebulizer formulation also can contain a surfactant to reduce or prevent surface-induced aggregation of the inhibitor composition caused by atomization of the solution in forming the aerosol.

Formulations for use with a metered-dose inhaler device generally comprise a finely divided powder containing the composition of the invention suspended in a propellant with the aid of a surfactant. The propellant can be any conventional material employed for this purpose, such as a chlorofluorocarbon, a hydrochlorofluorocarbon, a hydrofluorocarbon, or a hydrocarbon, including trichlorofluoromethane, dichlorodifluoromethane,

dichlorotetrafluoroethanol, and 1,1,1,2-tetrafluoroethane, or combinations thereof. Suitable surfactants include sorbitan trioleate and soya lecithin. Oleic acid also can be useful as a surfactant.

Formulations for dispensing from a powder inhaler device comprise a finely divided dry powder containing the composition of the invention and also can include a bulking agent, such as lactose, sorbitol, sucrose, mannitol, trehalose, or xylitol, in amounts that facilitate dispersal of the powder from the device, e.g., 50 to 90% by weight of the formulation.

Nasal delivery of the composition of the invention also is contemplated. Nasal delivery allows the passage of the composition to the blood stream directly after

administering the therapeutic product to the nose, without the necessity for deposition of the product in the lung. Formulations for nasal delivery include those with dextran or cyclodextran. Delivery via transport across other mucous membranes also is contemplated.

Compositions for rectal or vaginal administration are preferably suppositories that can be prepared by mixing the composition of the invention with suitable nonirritating excipients or carriers, such as cocoa butter, polyethylene glycol, or suppository wax, which are solid at room temperature, but liquid at body temperature, and therefore melt in the rectum or vaginal cavity and release the active compound.

The following examples are intended to be illustrative of certain embodiments and are non-limiting. The entire contents of all of the references (including literature references, issued patents, published patent applications, and co pending patent applications) cited throughout this application are hereby expressly incorporated by reference.

EXAMPLES

Example 1; Design of humanized 10F7 heavy and light-chain V regions

A "human-grade" molecule is generated by humanization of the 10F7 V regions. The sequences of murine 10F7 V regions and candidate humanized derivatives were designed with two primary goals: to reduce potential immunogenicity that would result from B cell epitopes and T cell epitopes that would not be part of the human 'self repertoire of sequences; and to enhance the stability of the folded sequence.

The starting sequences were: 10F7 VH (SEQ ID NO: 1)

QVKLQQSGAELVKPGASVKLSCKASGYTFNSYFMHWMKQRPVQGLEWIGMIRPNGGTTDYNE KFKNKATLTVDKSSNTAYMQLNSLTSGDSAVYYCARWEGSYYALDYWGQGTTVTVSS 10F7 VL (SEQ ID NO: 6)

DIELTQSPAIMSATLGEKVTMTCRASSNVKYMYWYQQKSGASPKLWIYYTSNLASGVPGRFS GSGSGTSYSLTI SSVEAEDAATYYCQQFTSSPYTFGGGTKLEIK

Four heavy chain and four light chains were designed that were humanized to various extents. Based on considerations of similarity to human antibody V regions and domain stability, the following protein sequences were designed: Humanized 10F7 VH-A (SEQ ID NO: 2)

EVQLVESGGGLVQPGGSLRLSCKASGYTFNSYFMHWVRQAPGKGLVWVSMIRPNGGTTDYAD SVKGRFTI SVDNSKNTLYLQMNSLRAEDTAVYYCARWEGSYYALDVWGQGTTVTVSS

Humanized 10F7 VH-B (SEQ ID NO: 3)

QVQLVQSGAELVQPGGSVRLSCKASGYTFNSYFMHWMKQAPGKGLEWVSMIRPNGGTTDYNE KFKGRATLSVDKSKNTAYMQLNSLRAEDTAVYYCARWEGSYYALDYWGQGTTVTVSS

Humanized 10F7 VH-C (SEQ ID NO: 4)

QVQLVQSGAEVKKPGASVKVSCKASGYTFNSYFMHWVRQAPGQGLEWMGMIRPNGGTTDYAQ KFQGRVTSTVDTS I STAYMELSRLRSDDTVVYYCARWEGSYYALDYWGQGTTVTVSS

Humanized 10F7 VH-D (SEQ ID NO: 5)

QVKLQQSGAEVVKPGASVKLSCKASGYTFNSYFMHWMKQAPGQGLEWIGMIRPNGGTTDYNE KFQGRVTLTVDKS I STAYMELSRLRSGDTVVYYCARWEGSYYALDYWGQGTTVTVSS

Humanized 10F7 VL-A (SEQ ID NO: 7)

DIQMTQSPSSLSASVGDRVTITCRASSNVKYLAWYQQKPGKAPKLLIYYTSNLQS GVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIK Humanized 10F7 VL-B (SEQ ID NO: 8)

DIQMTQSPSSLSASVGDRVTITCRASSNVKYLYWYQQKPGKAPKLLIYYTSNLAS GVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIK

Humanized 10F7 VL-C (SEQ ID NO: 9)

DIQLTQSPSSLSASVGDRVTITCRASSNVKYMYWYQQKPGKAPKLWIYYTSNLAS GVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIK The sequences 10F7 VH-A and VH-B were based on the human VH3 family of heavy chain V regions, specifically VH3-74. The sequences 10F7 VH-C and VH-D were based on the human VHl family of heavy chain V regions, specifically VHl-2. The 10F7 VH aligned more closely with VHl than VH3, but VH3 domains are generally observed to fold particularly well, so both types were designed.

The sequences 10F7 VL-A, B and C were all based on human Vkappal, specifically Vkappal-16. The human Vkappal family is thought to be the best-folding family of human light chains.

Coding sequences were designed as follows. The coding sequences for the murine VH and VL regions were described previously in, and are incorporated by reference from, US Patent Publication No.2011/0274658, which were used as a starting point.

Coding sequence for 10F7 VH domain (SEQ ID NO: 41)

CAAGTTAAGTTGCAACAATCTGGTGCTGAATTGGTTAAGCCAGGTGCTTCTGTTAAGTTGTC TTGTAAGGCTTCTGGTTACACCTTCAACTCTTACTTTATGCATTGGATGAAGCAAAGACCAG TTCAAGGTTTGGAATGGATTGGTATGATTAGACCAAACGGTGGTACTACCGATTACAACGAG AAGTTTAAGAACAAGGCTACTTTGACTGTTGATAAGTCCTCTAACACTGCTTACATGCAATT GAACTCTTTGACTTCTGGTGATTCTGCTGTTT ACT ACTGTGCTAGATGGGAAGGTTCT TACT ACGCTTTGGATTACTGGGGTCAAGGTACCACTGTTACTGTTTCT

Coding sequence for 10F7 VL domain (SEQ ID NO: 42)

GATATTGAGTTGACTCAATCTCCAGCTATTATGTCTGCTACCTTGGGTGAGAAGGTTACTAT GACTTGTAGAGCTTCATCTAACGTTAAGTACATGTACTGGTACCAACAGAAGTCTGGTGCTT CTCCAAAGTTGTGGATTTACTACACTTCTAACTTGGCTTCTGGTGTTCCAGGTAGATTTTCT GGTTCAGGTTCTGGTACTTCCTACTCTTTGACTATTTCCTCTGTTGAAGCTGAAGATGCTGC TACTTACTACTGTCAACAATTCACTTCTTCCCCATACACTTTTGGAGGAGGTACTAAGTTGG AAATCAAG

Based on these sequences, coding sequences for the modified protein sequences were generated by editing in Microsoft Word a text file of the murine coding sequences (SEQ ID 50 or 51), based on manual inspection of an alignment of the starting and final protein sequences and a codon usage table for human genome-encoded proteins. After editing, the coding sequences were translated using DNASTAR/Lasergene SeqBuilder version 11.2.1, and then compared by a 2-sequence BLAST alignment to the target sequence; this allowed for identification and correction of typographical and other errors. In this way the following coding sequences were generated. Coding sequence for humanized 10F7 VH-A domain (SEQ ID NO: 43) gAA GTT cAG TTG gtc gAA TCT GGT Gga Ggt TTG GTT cAG CCA GGT

GgT TCT tTg AgG TTG TCT TGT AAG GCT TCT GGT TAC ACC TTC AAC

TCT TAC TTT ATG CAT TGG gTG agG CAA gcA CCA GgT aAA GGT TTG

GtA TGG gTT tcT ATG ATT AGA CCA AAC GGT GGT ACT ACC GAT TAC gcC GAc tct gTT AAG ggC AgG ttT ACT aTc tCT GTT GAT AAc TCC aag AAC ACT etc TAC etc CAA aTG AAC TCT TTG Agg gCT Gag GAT aCT GCT GTT TAC TAC TGT GCT AGA TGG GAA GGT TCT TAC TAC GCT

TTG GAT gtC TGG GGT CAA GGT ACC ACT GTT ACT GTT TCT TCC

Coding sequence for humanized 10F7 VH-B domain (SEQ ID NO: 44)

CAA GTT cAG TTG gtc CAA TCT GGT GCa GAA TTG GTT cAG CCA GGT

GgT TCT GTg AgG TTG TCT TGT AAG GCT TCT GGT TAC ACC TTC AAC

TCT TAC TTT ATG CAT TGG ATG AAG CAA gcA CCA GgT aAA GGT TTG GAA TGG gTT tcT ATG ATT AGA CCA AAC GGT GGT ACT ACC GAT TAC

AAC GAG AAG TTT AAG ggC AgG GCT ACT TTg tCT GTT GAT AAG TCC aag AAC ACT GCc TAC ATG CAA CTG AAC TCT TTG Agg gCT Gag GAT aCT GCT GTT TAC TAC TGT GCT AGA TGG GAA GGT TCT TAC TAC GCT

TTG GAT TAC TGG GGT CAA GGT ACC ACT GTT ACT GTT TCT TCC

Coding sequence for humanized 10F7 VH-C domain (SEQ ID NO: 45)

CAA GTT cAG TTG gtc CAA TCT GGT GCT GAA gTG aag AAG CCA GGT

GCT TCT GTT AAG gTG TCT TGT AAG GCT TCT GGT TAC ACC TTC AAC

TCT TAC TTT ATG CAT TGG gTG AgG CAA gcA CCA GgT CAA GGT TTG GAA TGG ATg GGT ATG ATT AGA CCA AAC GGT GGT ACT ACC GAT TAC get cAG AAG TTT cAG ggC AgG GtT ACT Tct ACT GTT GAT act TCC ate tcC ACT GCT TAC ATG gAA TTG tcC agg TTG Agg TCT GaT GAT aCT GtT GTT TAC TAC TGT GCT AGA TGG GAA GGT TCT TAC TAC GCT

TTG GAT TAC TGG GGT CAA GGT ACC ACT GTT ACT GTT TCT TCC

Coding sequence for humanized 10F7 VH-D domain (SEQ ID NO: 46)

CAA GTT AAG TTG CAG CAA TCT GGT GCT GAA gTG GTg AAG CCA GGT GCT TCT GTT AAG TTG TCT TGT AAG GCT TCT GGT TAC ACC TTC AAC TCT TAC TTT ATG CAT TGG ATG AAG CAA gcA CCA GgT CAA GGT TTG GAA TGG ATC GGT ATG ATT AGA CCA AAC GGT GGT ACT ACC GAT TAC AAt GAG AAG TTT cAG ggC AgG GtT ACT TTG ACT GTT GAT aAA TCC ate tcC ACT GCT TAC ATG gAA TTG tcC agg TTG Agg TCT GgT GaT aCT GtT GTT TAC TAC TGT GCT AGA TGG GAA GGT TCT TAC TAC GCT TTG GAT TAC TGG GGT CAA GGT ACC ACT GTT ACT GTT TCT TCC

Coding sequence for humanized 10F7 VL-A domain (SEQ ID NO: 47) GAT ATT cAG aTG ACT CAA TCT CCA tCT AgT tTG TCT GCT Agt gTG GGT GAc AgG GTT ACT ATc ACT TGT AGA GCT TCA TCT

AAC GTT AAG TAC tTG gcC TGG TAC CAA CAG AAG cCT GGT aag gCT CCA AAG TTG etc ATT TAC TAC ACT TCT AAC TTG cag

TCT GGT GTT CCA aGT AGA TTT TCT GGT TCA GGT TCT GGT ACT gaC TAC aCT TTG ACT ATT TCC TCT cTT cAA cCT GAA GAT

Gtc GCT ACT TAC TAC TGT CAA CAA TTC ACT TCT TCC CCA TAC ACT TTT GGA caA GGT ACT AAG TTG GAA ATC AAG Coding sequence for humanized 10F7 VL-B domain (SEQ ID NO: 48)

GAT ATT cAG aTG ACT CAA TCT CCA tCT AgT tTG TCT GCT Agt gTG

GGT GAc AgG GTT ACT ATc ACT TGT AGA GCT TCA TCT

AAC GTT AAG TAC tTG TAC TGG TAC CAA CAG AAG cCT GGT aag gCT

CCA AAG TTG etc ATT TAC TAC ACT TCT AAC TTG GCT

TCT GGT GTT CCA aGT AGA TTT TCT GGT TCA GGT TCT GGT ACT gaC

TAC aCT TTG ACT ATT TCC TCT cTT cAA cCT GAA GAT

Gtc GCT ACT TAC TAC TGT CAA CAA TTC ACT TCT TCC CCA TAC ACT

TTT GGA caA GGT ACT AAG TTG GAA ATC AAG Coding sequence for humanized 10F7 VL-C domain (SEQ ID NO: 49)

GAT ATT cAG TTG ACT CAA TCT CCA tCT AgT tTG TCT GCT Agt gTG

GGT GAc AgG GTT ACT ATc ACT TGT AGA GCT TCA TCT

AAC GTT AAG TAC ATG TAC TGG TAC CAA CAG AAG cCT GGT aag gCT

CCA AAG TTG TGG ATT TAC TAC ACT TCT AAC TTG GCT TCT GGT GTT CCA aGT AGA TTT TCT GGT TCA GGT TCT GGT ACT gaC

TAC aCT TTG ACT ATT TCC TCT cTT cAA cCT GAA GAT

Gtc GCT ACT TAC TAC TGT CAA CAA TTC ACT TCT TCC CCA TAC ACT

TTT GGA caA GGT ACT AAG TTG GAA ATC AAG Example 2. Construction of coding sequences to express humanized 10F7 scFv domains

Humanized 10F7 V regions can be used as scFv fusions to other proteins such as EPO or self -proteins to which sufferers of an autoimmune disease have developed an immune response. Therefore specific DNA sequences were designed to facilitate testing in this format.

Testing of candidate humanized antibody V region sequences generally involves design of the individual heavy and light chains, and then construction of all possible pairwise combinations of designed and parental heavy and light chains, to identify the most humanized sequences that still show good expression and function.

The following sequences were ordered as G-Blocks™ from IDT to facilitate Gibson assembly® cloning of humanized VH and VL elements, so that an scFv would be expressed from the mammalian expression vector pSecTag2. SEQ ID NOs: 50 and 51 illustrate the approach. Analogous sequences for VH-A, VH-B, VH-C, VH-D, VL-A, VL-B, and VL-C were constructed. Murine 10F7 VH with adapters for Gibson assembly®: SEQ ID NO: 50

CTG CTG CTC TGG GTT CCA GGT TCC ACT GGT

CAA GT T AAG T TG CAA CAA TCT GGT GCT GAA T TG GT T AAG CCA GGT

GCT TCT GT T AAG T TG TCT TGT AAG GCT TCT GGT TAC ACC T TC AAC

TCT TAC T T T ATG CAT TGG ATG AAG CAA AGA CCA GT T CAA GGT T TG

GAA TGG AT T GGT ATG AT T AGA CCA AAC GGT GGT ACT ACC GAT TAC

AAC GAG AAG T T T AAG AAC AAG GCT ACT T TG ACT GT T GAT AAG TCC

TCT AAC ACT GCT TAC ATG CAA T TG AAC TCT T TG ACT TCT GGT GAT

TCT GCT GT T TAC TAC TGT GCT AGA TGG GAA GGT TCT TAC TAC GCT

T TG GAT TAC TGG GGT CAA GGT ACC ACT GT T ACT GT T TCT TCC GGT

GGA GGT GGA TCT GGT GGT GGA GGA TCT TCA GGA GGT

Murine 10F7 VL with adapters for Gibson assembly®: SEQ ID NO: 51

GGA TCT GGT GGT GGA GGA TCT TCA GGA GGT GGT GGA TCT TCC

GAT AT T GAG T TG ACT CAA TCT CCA GCT AT T ATG TCT GCT ACC T TG

GGT GAG AAG GT T ACT ATG ACT TGT AGA GCT TCA TCT

AAC GT T AAG TAC ATG TAC TGG TAC CAA CAG AAG TCT GGT GCT TCT

CCA AAG T TG TGG AT T TAC TAC ACT TCT AAC T TG GCT

TCT GGT GT T CCA GGT AGA T T T TCT GGT TCA GGT TCT GGT ACT TCC

TAC TCT T TG ACT AT T TCC TCT GT T GAA GCT GAA GAT

GCT GCT ACT TAC TAC TGT CAA CAA T TC ACT TCT TCC CCA TAC ACT

T T T GGA GGA GGT ACT AAG T TG GAA ATC AAG GGG ccc GAA CAA AAA

CTC ATC TCA GAA GAG

Expression vectors are constructed as follows. In total, 20 expression vectors are constructed, consisting of all possible pairwise combinations of the candidate humanized VH and VL sequences, as well as the murine VH and VL sequences. Using the G-blocks for each VH and VL segment and a PCR-amplified product derived from the vector used to express the 10F7 -Erythropoietin construct described in Taylor et al. (2010) Protein Engineering Design & Selection 23(4):251-60. Gibson assembly® reactions are performed, followed by treatment of the reaction mixture with Dpnl, transformation of E. coli and identification of clones with the correct plasmid as determined by restriction digestion and DNA sequencing of the V regions.

Example 3: Expression of humanized scFvs from mammalian cells The plasmids constructed in Example 2 were tested for their ability to express 10F7 scFvs upon transient transfection in mammalian cells as follows. Qiagen Maxiprep plasmid preparations of each expression plasmid were generated. Freestyle™ 293-F HEK cells were transfected with an equal amount of each plasmid in Freestyle™ 293 Expression Medium (Life Technologies; Carlsbad, CA). Alternatively, Freestyle™ Chinese Hamster Ovary (CHO) cells were transfected with an equal amount of each plasmid. Cells were then incubated for about 3-5 days, after which the cells were removed by centrifugation. The supernatants were collected, filtered, and tested for expression of each scFv. Expression was measured by examination of total supernatant on a stained SDS protein gel, and a band was identified at a molecular weight of about 25 kilodaltons corresponding to the scFv. This was confirmed by Western blotting, using an antibody directed against the His₆ tag. Based on visual inspection of the Western blot, the following results were obtained, as summarized in Table 2. Table 2: Relative expression levels of 10F7 scFv humanized variants in 293 cells

Based on these results, V regions corresponding to SEQ ID NO: 2, 4 and 8 were considered to express poorly and were not characterized further.

Example 4: Testing affinity of scFvs for human GYPA binding

The scFvs that expressed well were tested for affinity to GYPA as follows. Because GYPA is a transmembrane protein, it was most convenient to assay for binding activity on red blood cells. To qualitatively measure scFvs for binding, transfected cell supernatant was incubated with washed human red blood cells, and then incubated with a fluorescently labeled antibody directed against the His₆ tag on the scFv. Cells were then analyzed by FACS. Based on inspection of the FACS plots, and rough normalization of the binding signal to the expression level as measured in Example 3, the following results were obtained, as summarized in Table 3.

Table 3: Affinity of scFvs for GYPA

Taken together, the results of Examples 3 and 4 indicated that SEQ ID NOs: 3, 5 and 9 showed reduced expression but that the molecules that were secreted could bind with roughly wild-type affinity. SEQ ID NO. 7 was expressed at levels comparable to expression of SEQ ID NO: 1 but did not allow binding to GYPA. Based on these observations, SEQ ID NO: 3, 5 and 9 were further modified, as described in Example 8.

Example 5. Construction of a DNA encoding a Targeted EPO using V regions described herein

A non-limiting pair of the humanized V regions described herein are joined with a 15 to 20 amino acid linker, preferably in the configuration VH-L-VL, and then further joined to a mutated erythropoietin element, such as human erythropoietin alpha with the mutation Argl50Ala. For example, the protein sequences of VH-A and VL-A are joined with a long linker to erythropoietin with Argl50Ala in the configuration VH-L-VL - L(long) -

EPO(mut), as shown in the sequence below (the mutation Argl50Ala is shown in bold and underlined). This specific sequence is shown for illustration purposes only and is not to be considered limiting; any of the novel VH and VL regions described here may be combined with each other and with a mutated erythropoietin to create a targeted EPO protein. VH-A-VL-A-linker- EPO_RI_SOA (SEQ ID NO:39):

EVQLVESGGGLVQPGGSLRLSCKASGYTFNSYFMHWVRQAPGKGLVWVSMIRPNGGTTDYAD SVKGRFTI SVDNSKNTLYLQMNSLRAEDTAVYYCARWEGSYYALDVWGQGTTVTVSS GGGGSGGGGSSGGGGSS DIQMTQSPSSLSASVGDRVTITCRASSNVKYLAWYQQKPGKAPKLLIYYTSNLQS GVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIK GGSSGGGGSGGGSSGGGSSSGGGGSGGGGSSGGGSGGGS

A DNA segment encoding this protein sequence is constructed by standard techniques, expressed in Chinese Hamster Ovary cells by standard techniques, tested in cell- based assays as described by Taylor et al., evaluated in preclinical animals for safety and ultimately administered to humans. It is predicted that humans treated with the fusion protein do not develop neutralizing antibodies that cross-react with endogenous erythropoietin and do not develop pure red cell aplasia.

Nucleic acid sequences of additional examples of engineered proteins/chimeric activators are provided below.

10F7_W99G-linker-EPO_Ri50A (SEQ ID NO: 66) (The W99G mutation is bold, italicized, and underlined)

CAAGTTAAGTTGCAACAATCTGGTGCTGAATTGGTTAAGCCAGGTGCTTCTGTTAAGTTGTCTTGTAA GGCTTCTGGTTACACCTTCAACTCTTACTTTATGCATTGGATGAAGCAAAGACCAGTTCAAGGTTTGG AATGGATTGGTATGATTAGACCAAACGGTGGTACTACCGATTACAACGAGAAGTTTAAGAACAAGGCT ACTTTGACTGTTGATAAGTCCTCTAACACTGCTTACATGCAATTGAACTCTTTGACTTCTGGTGATTC TGCTGTTTACTACTGTGCTAGAGGGGAAGGTTCTTACTACGCTTTGGATTACTGGGGTCAAGGTACCA CTGTTACTGTTTCTTCCGGTGGAGGTGGATCTGGTGGTGGAGGATCTTCAGGAGGTGGTGGATCTTCC GATATTGAGTTGACTCAATCTCCAGCTATTATGTCTGCTACCTTGGGTGAGAAGGTTACTATGACTTG TAGAGCTTCATCTAACGTTAAGTACATGTACTGGTACCAACAGAAGTCTGGTGCTTCTCCAAAGTTGT GGATTTACTACACTTCTAACTTGGCTTCTGGTGTTCCAGGTAGATTTTCTGGTTCAGGTTCTGGTACT TCCTACTCTTTGACTATTTCCTCTGTTGAAGCTGAAGATGCTGCTACTTACTACTGTCAACAATTCAC TTCTTCCCCATACACTTTTGGAGGAGGTACTAAGTTGGAAATCAAGAGAGCTGCCGCAGGTGGAGGTG GTTCCGGAGGAGGATCTTCCGGTGGTGGATCTTCTTCTGGAGGTGGAGGATCCGGTGGTGGAGGATCA TCTGGTGGAGGATCTGGTGGTGGTTCCGCTCC ACCTAGATTGATTTGTGATTCCAGAGTTTTGGAAAG ATACTTGTTGGAAGCTAAGGAGGCTGAAAAGATTACTACTGGTTGTGCTGAACATTGTTCTTTGAACG AGAAGA TTACTGTTCCAGATACTAAGGTTAACTTTTACGCTTGGAAGAGAA TGGAAGTTGGTCAGCAA GCTGTTGAAGTTTGGCAAGGTTTGGCTTTGTTGTCTGAAGCTGTTTTGAGAGGTCAAGCTTTGTTGGT TAAGTCTTCTCAACCA TGGGAACCATTGCAATTGCA TGTTGA TAAGGCTGTTTCTGGTTTGAGA TCTT TGACTACCTTGTTGAGAGCTTTGGGTGCTCAAAAGGAAGCTA TTTCTCCTCCAGATGCTGCTTCTGCC GCTCCA TTGAGAACTA TTACTGCTGA TACTTTTAGAAAGTTGTTTAGAGTTTACTCTAACTTCTTGGC CGGTAA G T TGAA GTTGTACACTGG TGAA GCTTGTA GAA CTGG TGA TCGG

10F7-linker-EPO_K45D (SEQ ID NO: 67) (The K45D mutation is bold, italicized, and underlined)

CAAGTTAAGTTGCAACAATCTGGTGCTGAATTGGTTAAGCCAGGTGCTTCTGTTAAGTTGTCTTGTAA GGCTTCTGGTTACACCTTCAACTCTTACTTTATGCATTGGATGAAGCAAAGACCAGTTCAAGGTTTGG AATGGATTGGTATGATTAGACCAAACGGTGGTACTACCGATTACAACGAGAAGTTTAAGAACAAGGCT ACTTTGACTGTTGATAAGTCCTCTAACACTGCTTACATGCAATTGAACTCTTTGACTTCTGGTGATTC TGCTGTTTACTACTGTGCTAGAGTGGAAGGTTCTTACTACGCTTTGGATTACTGGGGTCAAGGTACCA CTGTTACTGTTTCTTCCGGTGGAGGTGGATCTGGTGGTGGAGGATCTTCAGGAGGTGGTGGATCTTCC GATATTGAGTTGACTCAATCTCCAGCTATTATGTCTGCTACCTTGGGTGAGAAGGTTACTATGACTTG TAGAGCTTCATCTAACGTTAAGTACATGTACTGGTACCAACAGAAGTCTGGTGCTTCTCCAAAGTTGT GGATTTACTACACTTCTAACTTGGCTTCTGGTGTTCCAGGTAGATTTTCTGGTTCAGGTTCTGGTACT TCCTACTCTTTGACTATTTCCTCTGTTGAAGCTGAAGATGCTGCTACTTACTACTGTCAACAATTCAC TTCTTCCCCATACACTTTTGGAGGAGGTACTAAGTTGGAAATCAAGAGAGCTGCCGCAGGTGGAGGTG GTTCCGGAGGAGGATCTTCCGGTGGTGGATCTTCTTCTGGAGGTGGAGGATCCGGTGGTGGAGGATCA TCTGGTGGAGGATCTGGTGGTGGTTCC GCTCCACCTAGATTGATTTGTGA TTCCAGAGTTTTGGAAAG ATACTTGTTGGAAGCTAAGGAGGCTGAAAAGA TTACTACTGGTTGTGCTGAACA TTGTTCTTTGAACG AGAAGA TTACTGTTCCAGATACTGATGTTAACTTTTACGCTTGGAAGAGAATGGAAGTTGGTCAGCAA GCTGTTGAAGTTTGGCAAGGTTTGGCTTTGTTGTCTGAAGCTGTTTTGAGAGGTCAAGCTTTGTTGGT TAAGTCTTCTCAACCA TGGGAACCATTGCAATTGCA TGTTGA TAAGGCTGTTTCTGGTTTGAGA TCTT TGACTACCTTGTTGAGAGCTTTGGGTGCTCAAAAGGAAGCTA TTTCTCCTCCAGATGCTGCTTCTGCC GCTCCA TTGAGAACTA TTACTGCTGA TACTTTTAGAAAGTTGTTTAGAGTTTACTCTAACTTCTTGAG AGGTAAGTTGAAGTTGTACACTGGTGAAGCTTGTAGAACTGGTGATCGG

Example 6: Treatment of a mammal expressing human glycophorin A with fusion proteins that include a glycophorin-binding element and an erythropoietin element

The engineered proteins/chimeric activators described herein contain a "targeting element" (e.g., an antibody fragment) that binds to a cell-specific surface marker (FIG. 6A, top panel) and is tethered to a mutated "activity element" (e.g. a hormone or cytokine) by a flexible peptide linker that permits simultaneous binding of both elements to the same cell surface. The targeting element anchors the mutated activity element to the desired cell surface (FIG. 6A, middle panel), thereby creating a high local concentration and driving receptor binding in spite of the mutation (FIG. 6A, bottom panel). Off-target signaling should be minimal (FIG. 6B) and decrease in proportion to the mutation strength.

The chimeric activator strategy was tested in vivo using EPO as the drug to be targeted to a cell. As described herein, EPO is a pleiotropic hormone that signals in diverse cell types. The design features of the chimeric activator 10F7-EPO_RI5_OA were chosen to direct human EPO activity to RBC precursors using a single-chain variable fragment (scFv) from the antibody 10F7 (FIG. 6C). 72 ' 120 10F7 binds the common variants of human glycophorin A (huGYPA) (FIG. 6C), which is restricted to the RBC lineage and expressed at about 800,000 copies on mature RBCs. 82 ' 83 Several characteristics of huGYPA make it a desirable receptor to target: (1) it has a small extracellular domain such that an scFv-EPO fusion protein could likely simultaneously bind to both huGYPA and EPO-R (FIG. 6C), 72 (2) the sequence of the 10F7 V regions are available (GI: 15149451), and (3) loss of huGYPA is pheno typically silent,¹²¹ so binding to huGYPA is unlikely to cause side-effects. 10F7 was tethered via a flexible thirty-five amino acid glycine/serine linker to a form of EPO mutated

71 75

at position 150 from arginine to alanine (EPO_RI_SOA) ' : this mutation changes a surface residue that makes a strong contact with EPO-R, minimizes binding, and does not affect

75

protein folding. The resulting 10F7-EPO_RI_SOA showed in vitro activity that was enhanced relative to EPO_RI_SOA alone by 10- to 27-fold on erythroleukemic cell lines that express EPO-R and huGYPA (FIGs. 11A-11D).⁷¹

It was inferred that huGYPA and EPO-R were co-expressed on RBC precursors in vivo (FIG. 6D). Late RBC precursors (BFU-E descendants) express huGYPA approximately five days before maturing into erythrocytes, 73 while EPO-R binding events on cultured late- stage precursor cells (CFU-E stage) could be detected as late as twenty-four hours before the cells matured into reticulocytes or RBCs. 122 These results suggested that EPO-R and huGYPA are co-expressed for a period during late erythropoiesis, although this has not been directly addressed in vivo until the experiments described herein.

The chimeric activator 10F7-EPO_RI_SOA and control variants were tested in huGYPA transgenic mice 123 because 10F7 does not cross-react with murine GYPA (27) (human EPO activates murine EPO-Rs ). This animal model reflects the normal expression pattern of huGYPA.¹²³ Transgenic RBCs express less huGYPA than human RBCs (FIG. 12A): 1.6 μΜ huGYPA is exposed to plasma in transgenic mice, versus 2.8 μΜ in humans (FIG. 12B). Treatment of huGYPA transgenic mice with 10F7-EPO_RI5_OA should stimulate erythropoiesis, and this production should depend on huGYPA expression and a functional 10F7 element.

In vitro characterization of chimeric activator variants The experimental strategy was designed to identify the chimeric activator features required for the desired in vivo behavior: RBC expansion in the absence of platelet production. 10F7-EPO_RI5_OA was compared with variants in which EPO was not mutated (10F7-EPO), 10F7 was mutated to mitigate huGYPA binding (10F7_W99G-EPO_RI₅OA), and EPO

125

was mutated to eliminate EPO-R affinity (10F7-EPO_K4SD)- Darbepoetin alfa (Aranesp™) was used as a non-targeted form of EPO. FIG. 7A shows schematics of the tested engineered proteins and verification of their size and N-linked glycosylation.

The effect of the R150A mutation on the interaction between EPO and EPO-R was quantified (FIG. 7B). EPO and 10F7-EPO had similar binding kinetics (K_D = 5.4 nM and 7.2 nM, respectively) in accordance with published values,¹²⁶ indicating that the 10F7 targeting element does not interfere with EPO binding to EPO-R. The results also confirmed that

R150A weakened the interaction between EPO and EPO-R: while on-rates of 10F7-EPO and 10F7-EPO_RI5_OA were similar (3.5xl0⁴ M^"1sec^_1 versus 3.6xl0⁴ M^"1sec^"1, respectively), their off -rates differed by 12-fold (2.5xl0^~4 sec^"1 versus 3.1xl0^~3 sec^"1, respectively).

Mutations in 10F7 and EPO showed predicted effects in cell-based assays. In vitro activity was measured via proliferation of erythroleukemia cells that express huGYPA and EPO-R (FIGs. 11A-11D).⁷¹ Fusion of 10F7 to wild-type EPO enhanced in vitro activity by 1.7 fold relative to EPO alone (FIG. 11A). By contrast, fusion 10F7 to EPO_RI_SOA improved in vitro activity by 14-fold relative to EPO_RI_SOA alone (FIG. 11B), in agreement with previous in

71 127 128

vitro testing. ' ' 10F7W99G-EPO_RISOA and EPO_RISOA had similar in vitro activities (FIG. 11C), implying that the mutation W99G weakens the affinity of 10F7 for huGYPA. Finally, neither 10F7-EPO_K4SD nor EPO_K4SD alone had detectable in vitro activity (FIG. 11D), in agreement with previous work indicating that K45D is a null mutation. 75

In a separate in vitro assay, the EPO mutation R150Awas found to prevent enhanced proliferation of EPO-R-positive tumor cells (FIG. 7C). EPO can stimulate the growth of tumor cells that express EPO-R and thereby enhance tumor growth in patients with EPO-R- positive tumors. 23 Proliferation of EPO-R-positive MCF-7 cells was compared with EPO-R- negative HeLa cells. 127 ' 128 Exposure to 50 nM of 10F7-EPO for three days caused a 2.4-fold increase in MCF-7 proliferation, as compared to 10F7-EPO_Riso_A (P-value < 0.04); no effect was observed in HeLa cells. These results indicated that the mutation R150A mitigated undesired EPO activity on non-erythroid cells.

Pharmacodynamics of chimeric activator variants

Animal testing indicated that 10F7-EPO_RI5_OA targets EPO activity to RBCs, and that the structural features of chimeric activators are essential for the desired in vivo behavior (FIGs. 8A-Q and FIGs. 9A-9P). The chimeric activatorlOF7-EPORi50A was compared to darbepoetin, control proteins, and saline. Pharmacodynamics will be a function of receptor on- and off-rates, plasma half-life, and sequestration onto RBCs. Accordingly, to fairly assess the corresponding impact on platelets, proteins were compared at doses that achieved similar effects on RBC expansion: 50 pmol of darbepoetin and 125 pmol for 10F7-EPO, 10F7-EPORI5OA, and 10F7W99G-EPORISOA- Reticulocytes (RBC precursors as a percentage of total RBCs), hematocrits (volume percentage of total RBCs), reticulated platelets (platelet precursors as a percentage of total platelets), and platelets (total platelet count per whole blood volume) were assessed. Reticulocytes and reticulated platelets are less than twenty- four hours old and thus measure new cell production. 129 ' 130 ' 131 Animals received a single intra-peritoneal injection, and responses were measured after 4, 7, and 11 days post-dosing, with four days being roughly when a robust reticulocyte response can first be observed.

In huGYPA transgenic mice (FIG. 8A), 10F7-EPO_RI_SOA stimulated expansion of reticulocytes but not reticulated platelets (FIGs. 8A-8Qand FIGs. 13A-13K). Average baseline reticulocyte and reticulated platelet counts were 5.0% and 19%, respectively. At the highest dose, darbepoetin, 10F7-EPO, and 10F7-EPO_RI_SOA raised reticulocytes by 12-14% by day 4 (FIGs. 8B-8D). Darbepoetin and 10F7-EPO also strongly impacted reticulated platelets: by 12% and 9.1%, respectively (FIGs. 8G and 8H). Only 10F7-EPO_RI_5OA had a specific effect on reticulocytes relative to reticulated platelets (FIG. 8D compared to FIG. 81), causing a marginal increase in reticulated platelets of 3.1% by day 4, comparable to the effect of 10F7_W99G-EPO_RI5OA or saline (FIG. 8J compared to FIG. 8K: Δ4.9% and Δ2.8%, respectively). These trends were similar for all tested doses.

The synthesis of reticulated platelets by darbepoetin and 10F7-EPO_RI_SOA was not due to treatment with saturating doses. By day 4, treatment with a low dose of darbepoetin caused reticulocytes to increase by 5.2% (FIG. 8B), while a high dose of 10F7-EPO_RI_SOA increased reticulocytes by 12.2% (FIG. 8D). However, at these same doses, darbepoetin increased reticulated platelets by 7.6% (FIG. 8G), while 10F7-EPO_RI_SOA increased reticulated platelets by only 2.9% (FIG. 81). Thus, compared to darbepoetin, 10F7-EPO_RI5_OA causes a greater stimulation of reticulocytes but less stimulation of reticulated platelets.

Pharmacodynamics of 10F7-EPO_RI5_OA depended on huGYPA expression. At all doses, 10F7-EPO_RI5_OA produced a lasting reticulocyte response in transgenic mice (FIG. 8L), but had little effect in non-transgenic mice (FIG. 8M); no effect on reticulated platelets was observed in either group (FIGs. 8N and 80). Furthermore, the 10F7 element does not signal on its own: 10F7-EPO_K45_D, in which EPO is completely nonfunctional, had no effect on reticulocytes or reticulated platelets (FIGs. 8P and 8Q).

In huGYPA transgenic mice, 10F7-EPO_RI5_OA stimulated total RBCs but not total platelets (FIGs. 9A-9P, 14A-14P, and 15A-15P). Average baseline reticulocyte, hematocrit, and platelet counts were 5.0%, 51%, and 1.2χ10⁶/μΕ, respectively. Reticulocytes increased by 13-15% of total RBCs at the highest doses of darbepoetin, 10F7-EPO, and 10F7-EPO_RI_5OA (FIGs. 9B-9D), and hematocrit changes mirrored reticulocytes (FIGs. 9G-9I). These effects were dose-dependent. Platelets had a different response pattern: darbepoetin and 10F7-EPO caused platelet counts to significantly increase from baseline at all tested doses (FIGs. 9L and 9M), while 10F7-EPO_RI_SOA had little or no impact on platelet counts at any dose (FIG. 9N). This differential effect on RBCs versus platelets depends on the R150A mutation in EPO, as it was not observed with 10F7-EPO (FIG. 9M). In all experiments, 10F7_W99_G-EPO_RI5_OA (FIGs. 9E, 9 J, and 9K) and saline (FIGs. 9F, 9K, and 9P) had minimal effects.

In sum, only 10F7-EPO_RI_SOA caused a specific increase in reticulocytes and RBCs without also increasing reticulated and mature platelets. This specificity required a weakened EPO element, a functional 10F7 targeting element, and expression of the targeting receptor huGYPA. These results illustrate how cell-specific signaling can be achieved with a targeted fusion protein, but only if the binding properties of its elements have been modulated.

Pharmacokinetics of chimeric activator variants

Binding of 10F7-EPO_RI_SOA to huGYPA reduces its maximal plasma concentration (Cmax) and increases its terminal plasma half-life. EPO pharmacokinetics can be influenced by receptor binding, glycosylation, and molecular weight, which respectively affect clearance through receptor- mediated endocytosis by EPO-Rs, liver asialoglycoprotein receptors, and kidney filtration. 2 ' 126 ' 132 Moreover, binding to huGYPA on mature RBCs is expected to

77

create a sink effect, by which most of 10F7-EPO_RI_SOA should be maintained in equilibrium with a free plasma state. FIG. 10A illustrates a bio-distribution compartment model for 10F7-EPO_RI5_OA- Clearance should mainly occur through binding EPO-Rs on late RBC precursors. Kidney clearance should be minimal due to the molecule's large size. Binding to non-erythroid EPO- Rs should be reduced due to the R150A EPO mutation, and binding to asialoglycoprotein receptors should remove only a subpopulation of drug molecules.¹²⁶ Finally, clearance of

RBC-bound drug via splenic apoptosis should be slow. 133

The Cmax of 10F7-EPO_RI5OA is strongly influenced by binding to huGYPA. To measure the protein's terminal half-life in plasma or bound to RBCs, transgenic or non- transgenic mice were injected with 100 μg (1.39 nmol) of 10F7-EPO_RI_SOA, and plasma or whole blood was collected in a five-day time-course (FIG. 10B). In wild-type mice, the initial plasma concentration of 10F7-EPO_RI_SOA was 82 g/ml, which corresponds to the injected dose in a 1.1 ml plasma volume (FIG. 12B). By contrast, the initial plasma concentration in transgenic mice was 16 g/ml, suggesting that about 80% of the injected protein immediately bound to huGYPA on RBCs. These results agree with the predicted effect of 10F7-EPO_RI₅OA binding to huGYPA (FIG. 12B).

The terminal plasma half-life of was extended by binding to huGYPA on mature RBCs. In transgenic mice, 10F7-EPO_RI_SOA had terminal plasma and RBC-bound half-lives of 27.7 h and 24.8 h, respectively (FIGs. IOC and 10D). The ratio of 10F7-EPO_RI₅₀A in plasma to that bound to RBCs was roughly constant at all time-points (FIGs. 16 and 17), consistent with a rapid equilibration between the bound and unbound states. By comparison, in non- transgenic mice, 10F7-EPORisoA had a terminal plasma half-life of 15.1 h (FIGs. IOC and 17), and RBC-binding was not meaningfully detected. These results indicate that binding to huGYPA extends the plasma half-life and reduces the C_max of 10F7-EPO_RISOA- METHODS

Cell culture

Human erythroleukemia TF-1 cells (American Type Culture Collection; ATCC;

Manassas, VA) were cultured in RPMI-1640 Medium with 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 U/mL streptomycin, and 2 ng/mL recombinant human granulocyte - macrophage colony- stimulating factor (GM-CSF, Life Technologies; Calsbad, CA) unless otherwise specified. FreeStyle™ Chinese Hamster Ovary (CHO-S) cells were cultured in FreeStyle™ CHO Expression Medium (Life Technologies; Calsbad, CA). CHO DG44 cells were cultured in complete DG44 Medium (Life Technologies; Calsbad, CA). Human breast cancer MCF-7 cells and cervical cancer HeLa cells were cultured in Eagle's Minimal Essential Medium (ATCC: Manassas, VA), 10% FBS, and 1% insulin. TF-1, MCF-7, and HeLa cells were cultured at 37°C in 5% C0₂. CHO-S and CHO DG44 were cultured at 37°C in 8% C0₂ with shaking at 130 RPM.

Construction of chimeric activator variants

Examples of chimeric activator coding sequences are provided herein, for example SEQ ID NOs: 39, 66, and 67. The general structure is as follows: the 10F7 scFv (GI:

15149451) linked by a 35 amino acid glycine/serine linker to human EPO (GI: 5822016); unfused EPO variants were also made. Constructs were sub-cloned into pSecTag2A or pOptiVEC vectors (Life Technologies; Carlsbad, CA) for transient or stable expression, respectively. These vectors contain a CMV promoter, murine Ig k-chain leader sequence, and a C-terminal c-myc epitope and His6 tag for purification.

The mutation EPO_K4SD was designed based on published site-directed mutagenesis results,⁷⁵ and inspection of the EPO:EPO-R crystal structure. EPO mutation 10F7w₉₉G alters CDR3 of the heavy chain and was designed as follows: a model of 10F7 V regions was constructed using the Rosetta Online Server, and a BLAST alignment of 10F7 heavy and light chains was performed to identify common CDR substitutions in closely related antibodies. 10F7w₉₉G was tested based on inspection of the structure model and alignments: the construct 10F7_W99_G-EPO_RI_SOA expressed well and exhibited the desired lack of activity in a TF-1 cell proliferation assay (FIG. 11D).

Expression and purification of chimeric activator variants

Transient and stable expression of proteins was carried out using FreeStyle™ CHO-S cells (Life Technologies; Carlsbad, CA) or CHO DG44 cells (Life Technologies; Carlsbad, CA), respectively, according to standard procedures. Proteins were purified by a two-step process.

For transient expression, Freestyle™ CHO-S cells (Life Technologies; Carlsbad, CA) were transfected with pSecTag2A plasmid according to supplier's protocol. After five days of culture, cells were pelleted at 500 g and supernatant protein expression was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using Coomassie Fluor™ Orange stain (Life Technologies; Carlsbad, CA). For stable expression, CHO DG44 cells were transfected with pOptiVec plasmid according to supplier's protocol. After two days of culture, cells were moved to complete DHFR-negative CD OptiCHO™ medium (Life Technologies; Carlsbad, CA) for selection. Selection continued until cells recovered to >90% viability. Expression was assayed as described above. To enhance expression, each stable pool underwent one round of methotrexate amplification and expression was again assayed as described above.

Proteins derived from transient transfection or stable cell lines were purified as follows. Supernatant was concentrated to 7 mL using Macrosep Advanced Centrifugal devices (Pall Corporation; Westborough, MA) or tangential flow filtration (Labscale™ TFF Lab System, EMD Millipore; Billerica, MA). Concentrated protein was bound to 700 of ProBond™ Nickel Chelating Resin (Life Technologies; Carlsbad, CA) for 1 h at 4°C while rotating in a 15 mL Falcon tube, washed thrice with native purification buffer (50 mM

NaH₂P0₄, 0.5 M NaCl, pH 8.0) plus 20 mM imidazole, and eluted with native purification buffer plus increasing imidazole concentrations: 50 mM, 100 mM, 250 mM. Eluted proteins were desalted (Econo-Pac 10DG columns, Bio-Rad; Hercules, CA) into lx phosphate- buffered saline (PBS, Teknova; pH 7.4, 137 mM NaCL, 1.4 mM KH₂P0₄, 4.3 mM Na₂HP0₄, 2.7 mM KC1) and concentrated to 1 mL. Contaminating proteins were removed by size- exclusion chromatography on Superdex™ 200 10/300 GL columns (GE Healthcare; Boston, MA) using an AKTAFPLC system (GE Healthcare; Boston, MA). Desired protein fractions were pooled, concentrated to <1 mL, and quantified by the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific; Cambridge, MA).

Protein concentration and purity (>95%) was verified by Coomassie-SDS-PAGE, and N-linked glycosylation was verified with PNGase F (New England Biolabs; Ipswich, MA) according to supplier's protocol. Proteins were stored at 4°C throughout the described process, ultimately stored as aliquots at -80°C, and thawed once before use. Only endotoxin- free reagents were used.

In vitro characterization of chimeric activator variants

A kinetic analysis of EPO-R binding by protein variants was performed using the BLItz® system (Pall ForteBio; Menlo Park, CA) according to the manufacturer's

instructions, and test proteins were added to measure association and dissociation constants.

The extracellular moiety of EPO-R N-terminally fused to Fc (R&D Systems;

Minneapolis, MN) was diluted to 50 ug/mL in lx BLItz® kinetic buffer (Pall ForteBio; Menlo Park, CA). Protein A Dip and Read™ Biosensors (Pall ForteBio; Menlo Park, CA) were hydrated in kinetic buffer for 10 min and a baseline read in kinetic buffer was taken for 30 s. 4 μϊ_^ of EPO-R-Fc protein was immobilized to biosensors for 120 s, followed by a 30 s wash in kinetic buffer to remove unbound protein. Biosensors were then exposed to 4 μL of varying concentrations of protein variants for 200 s to measure protein association, followed by a 200 s dissociation step in kinetic buffer. Binding constants were quantified using BLItz® software. (Pall ForteBio; Menlo Park, CA).

Stimulation of TF-1 cell proliferation by a given protein was tested as follows. 10⁴ cells per well were seeded in a 96-well plate in 100 of RPMI-1640 with serum and antibiotics (without GM-CSF). Cells were incubated with 10-fold serial dilutions (0.01 - 100 nM) of engineered proteins for 72 h at 37°C in 5% C0₂. Proliferation was determined by addition of 5 of WST-1 Cell Proliferation Reagent (Roche Diagnostics; Indianapolis, IN): after a 4 h incubation with dye, proliferation was measured by reading absorbance at 450 nm with background subtraction at 650 nm on a BioTek NEO HTS plate reader. Resulting data were plotted and fitted with nonlinear regression using GraphPad™, from which half maximal effective concentration (EC₅₀) values were obtained. Data represent the average ± standard error of three replicates.

Effects of chimeric activator variants on MCF-7 and HeLa cell proliferation were tested as follows. 10⁴ cells were seeded in a 96-well plate in 100 of growth medium per well. Cells adhered for 24 h, and medium was then replaced and supplemented with 0 or 50 nM of engineered proteins for 24 h at 37°C in 5% C0₂. Proliferation was measured as described above. Data represent the average ± standard error of three replicates.

Animal model

Human glycophorin A (huGYPA) transgenic FVB mice¹²³ were generously donated by the Hendrickson Lab (Emory University). This strain underwent embryo re-derivation at Charles River Laboratories (Wilmington, MA). Homozygous huGYPA transgene expression is embryonic lethal. Pups were screened for heterozygous transgene expression by measuring huGYPA expression on RBCs via flow cytometry. <10 lL of blood was collected by tail- nick in EDTA-coated capillary tubes (Sarstedt; Niimbrecht, Germany), and 1 iL was diluted 1: 1000 in FACS buffer (R&D Systems; Minneapolis, MN) at room temperature. On a 96- well plate, 100 \L of diluted blood was incubated with 5 lL of anti-huGYPA-PE antibody (R&D Systems) per well for 40 min in the dark at room temperature. Cells were washed 3x with FACS buffer and assayed immediately or stored at 4°C for <24 h prior to FACS analysis.

huGYPA expression on transgenic mouse and human RBCs was quantified using the BD QuantibriteTM Phycoerythrin (PE) Fluorescence Quantitation kit (BD Biosciences; San Jose, CA) with PE beads. Fresh whole human blood was obtained from Research Blood Components, LLC (Boston, MA) in EDTA vacutainers. Fresh whole huGYPA transgenic mouse blood was collected as described above. Cells were washed 2x with FACS buffer (R&D Systems; Minneapolis, MN). On a 96-well plate, 106 cells in 100 μΐ_^ volume were labeled with PE-conjugated anti-huGYPA antibody (R&D Systems; Minneapolis, MN) or PE-conjugated IgG isotype control (R&D Systems; Minneapolis, MN). Cells were washed 3x with FACS buffer, fixed in 1% paraformaldehyde, and stored at 4°C for <24 h prior to FACS analysis. Data represent the average + standard error of two biological replicates.

Pharmacodynamics and pharmacokinetics of chimeric activator variants

Four mice (2 male, 2 female, age 12-16 weeks) received one intraperitoneal (IP) injection with a given protein at varying concentrations in a 250 volume, and <50 of whole blood was collected by tail-nick in EDTA-coated tubes (Sarstedt; Niimbrecht, Germany) on days 0, 4, 7, and 11 post- injection. Within 3 h of collection, blood was analyzed. Total RBC and platelet counts were determined using a hematology analyzer (Hemavet 950, Drew Scientific; Waterbury, CT). Reticulocyte counts were determined by standard flow cytometry. A stock solution (1 mg/mL) of thiazole orange (Sigma- Aldrich; St. Louis, MO) was prepared in methanol and diluted 1: 10,000 in PBS. Whole blood was diluted 1:2000 in the working solution and incubated for 30 min in the dark at room temperature. 200 μL of stained blood per sample was added to a 96 well plate, fixed in 1% paraformaldehyde, and stored at 4°C for < 24 h prior to FACS analysis. Data represent the average ± standard error of four biological replicates.

In a separate experiment, reticulocytes and reticulated platelets were determined as described above by flow cytometry, including a co-stain with anit-CD41-PE antibody (BD Biosciences; San Jose, CA). Samples were examined by two-color analysis of thiazole orange and anit-CD41-PE (compared to the internal standard of platelets only labeled with CD41-PE). Data represent the average + standard error of four biological replicates.

Pharmacokinetics of 1 OF7-EPORI5OA Due to time constraints for sample processing, pharmacokinetic profiling of plasma concentrations and RBC-associated protein were performed in separate experiments. Four huGYPA transgenic or non-transgenic mice (2 males, 2 females, age 12-16 weeks) were dosed once by tail-vein injection with 100 μg of 10F7-EPO_RI_SOA, and <20 of whole blood was collected by tail-nick in EDTA-coated tubes (Sarstedt; Niimbrecht, Germany) in a 120 h time-course. Immediately after collection, blood was centrifuged at 6000 rpm for 5 min, and plasma was collected and transferred to a new tube on ice and ultimately frozen at -20°C until analysis. Plasma was analyzed using a human EPO ELISA kit (R&D Systems; Minneapolis, MN). Data was plotted on a logio scale, and terminal plasma half-lives were determined by fitting the terminal time-points to an exponential curve using GraphPad™. Data represent the average ± standard error of two biological replicates.

RBC-bound 10F7-EPO_RI_SOA was detected by flow cytometry. Whole blood was collected as described above in a 120 h time-course. Blood was diluted 100-fold into 10 mM EDTA/PBS and washed 2x with FACS buffer (R&D Systems; Minneapolis, MN). On a 96- well plate, 10⁶ cells in 100 volume were labeled using PE-conjugated anti-6x histidine tag antibody (Abeam; Cambridge, MA) or PE-conjugated IgG isotype control (R&D Systems: Minneapolis, MN). Cells were washed 3x with FACS buffer, fixed in 1% paraformaldehyde, and stored at 4°C for < 24 h prior to FACS analysis. Data was plotted on a logio scale, and terminal half-lives were determined by fitting the terminal time-points to an exponential curve using GraphPad™. Data represent the average ± standard error of two biological replicates.

Flow cytometry

Samples were analyzed on a BD LSRFortessa SORP flow cytometer equipped with an optional HTS sampler (BD Biosciences; San Jose, CA) using the following filter

configuration. FITC excitation: 488 nm/100 mW, emission filter BP 515/20; PE excitation: 561/50 mW, emission filter BP 582/15. Data were analyzed using FlowJo (Tree Star;

Ashland, OR).

Example 7: Pharmaceutically acceptable formulations of humanized 10F7 fusion proteins

Phosphate-buffered saline (PBS) was found to be a pharmaceutically acceptable formulation for fusion proteins comprising a glycophorin-binding element and an erythropoietin element. In particular, the experiments performed in Example 6 used PBS as the formulation buffer. For example, PBS from Teknova (Hollister, CA) (pH 7.4, 137 mM NaCL, 1.4 mM KH2P04, 4.3 mM Na2HP04, 2.7 mM KC1) was used.

Samples at a concentration of about 10-50 mcg/mL were prepared by two-step purification of the fusion protein to greater than 90% purity, and then stored in PBS frozen at -80°C for several months. Samples were thawed once before injection into a mammal, such as a mouse.

Other formulations are used as well. For example, "MRCT" is used: this

pharmaceutically acceptable formulation includes about 4% mannitol, 100 mM arginine, 5 mM citric acid, and 0.01% Tween 80. This solution is buffered to about pH 7, for example with hydrochoric acid, before use. The same buffer without citrate may also be used.

Pharmaceutically acceptable formulations may include human serum albumin as well.

Example 8: Additional humanized variants of 10F7 V regions.

Based on the results of Examples 3 and 4, the following additional humanized sequences were designed. The modifications were made based on insights of the inventors into the nature of antibody V region folding, expression and stability.

VH-5 (SEQ ID NO: 52)

QVQL VQ S GAE L VQP GG S VRL S C KAS GYTFNSYF MHWMKQAPGKGLEWVSM I RPNGGTTD YNEKFKGRVTLSVDKSKS TAYMQLNS LRAEDTAVYY CARWEGSYYA LDYWGQGTTVTVS S

VH-6 (SEQ ID NO: 53)

QVQL VQ S GAE L VQP GG S VRL S C KAS GYTFNSYF MHWMKQAPGKGLEWVSM I RPNGGTTD YAEKFKGRVTLSVDKSKS TAYMQLNS LRAEDTAVYY CARWEGSYYA LDYWGQGTTVTVS S

VH-7 (SEQ ID NO: 54)

QVQLVQSGGELVQPGGSLRL SCKAS GYTFNSYF MHWMRQAPGKGLEWVSM I RPNGGTTD YAESFKGRFT I SRDKSKS TLYMQLNS LRAEDTAVYY CARWEGSYYA LDYWGQGTTVTVS S

VH-8 (SEQ ID NO: 55)

QVQLVE SGGGLVQPGGSLRL SCAAS GYTFNSYF MHWVRQAPGKGLEWVSV I RPNGGTTD YADSVKGRFT I SRDNSKNTLYQMNNS LRAEDTAVYY CARWEGSYYA LDYWGQGTTVTVS S VH-9 (SEQ ID NO: 56)

QVKL QQ S GAE VVKP GAS VKL S C KAS GYTFNSYF MHWMKQAPGQGLEWI GM I RPNGGTTD YNEKFQGRATLTVDKSKS TAYMEL SRLRSGDSAVYY CARWEGSYYA LDYWGQGTTVTVS S

VL-4 (SEQ ID NO: 57)

DI QLTQSPS S LSASVGDRVT I TCRAS SNVKY MAWYQQKPGKAPKLWIY YTS

NLAS GVPSRF SGSGSGTDYTLT I S SLQPEDVATYYC QQFTS SPYTFGQGTKLE IK

VL-5 (SEQ ID NO: 58) DIQLTQSPSSLSASVGDRVTITCRAS SNVKY MAWYQQKPGKAPKLLIY YTS

NLQSGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK

VL-6 (SEQ ID NO: 59)

DIQMTQSPSSLSASVGDRVTITCRAS SNVKY MAWYQQKPGKAPKLLIY YTS

NLQSGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK

VL 7 (SEQ ID NO: 60)

DIQLTQSPSSLSASVGDRVTMTCRAS SNVKY MYWYQQKPGKAPKLWIY YTS NLASGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK

"G-blocks" of DNA corresponding to coding sequences for each of these segments were ordered from Integrated DNA Technologies.

The G-blocks corresponding to coding sequences of each of the preceding sequences were combinatorially combined into expression constructs and expressed in mammalian cells, as described in Examples 2 and 3. Expression levels were tested as in described in Example 3 and binding to 10F7 is tested as in Example 4.

Based on results from expression and binding assays, a humanized variant of a 10F7 scFv with improved folding, expression, stability and binding characteristics was obtained. Specifically, all of the new VH and VL chains showed rather poor expression in combination with each other or with the parental murine V regions, except for VH-9 and VL-7. For example, the following scFv sequence was obtained (from VH-9 and VL-7). This construct, expressed as an scFv with a myc epitope tag and a His6 tag in 293 cells, was expressed at levels almost as high as a corresponding scFv based on the murine 10F7 V regions:

QVQLVQSGAELVQPGGSVRLSCKASGYTFNSYFMHWMKQAPGKGLEWVSMIRPNGGTTDYNEKFKGRATLSVDKS KSTAYMQLNSLRAEDSAVYYCARWEGSYYALDYWGQGTTVTVSGGGGSGGGGSSGGGGSSDIQLTQSPSSLSASV GDRVTMTCRASSNVKYMYWYQQKPGKAPKLWIYYTSNLASGVPSRFSGSGSGTDYTLTI SSLQPEDVATYYCQQF TSSPYTFGQGTKLEIKGPEQKLI SEEDLNSAVDHHHHHH (SEQ ID NO: 68)

SEQ ID NO.61, below, shows the core V region sequences without tags:

QVKL QQ S GAE VVKP GAS VKL S C KAS GYTFNSYF MHWMKQAPGQGLEWI GM IRPNGGTTD YNEKFQGRATLTVDKSKSTAYMELSRLRSGDSAVYY CARWEGSYYA LDYWGQGTTVTVSS GGGGSGGGGSSGGGGSS DIQLTQSPSSLSASVGDRVTMTCRAS SNVKY MYWYQQKPGKAPKLWIY YTS NLASGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK (SEQ ID NO: 61)

The VH-9/VL-7-(myc-His6) protein was expressed by transient transfection of an expression vector into HEK 293 cells, and the ability to bind to human glycophorin A was assayed by incubating the transfection supernatant with TF-1 cells (a human leukemic cell line expressing human glycophorin A) followed by incubation with a phycoerythrin-coupled secondary antibody directed against the myc epitope tag. Specifically, about 100,000 TF-1 cells were placed in a 100 microliter binding reaction in FACS buffer (specified for a BD LSRFortessa flow cytometer; FACS buffer is a type of PBS). On ice in a 96-well round- bottom plate, about 50 microliters of supernatant containing the VH-9/VL-7-(myc-His6) protein was added to the cells, and incubated for 1 hour. The plate was then spun at 500xg, supernatants removed, cells washed with 200 microliters FACS buffer, and this process repeated once more. The cells were then resuspended in 100 microliters of FACS buffer, to which was added 100 microliters of a 1:50 dilution of an anti-His6 monoclonal antibody from Abeam (catalogue number Ab72467, lot number GR 168372-9). This mixture was incubated 30 minutes on ice, and cells were again washed twice in FACS buffer, resuspended in FACS buffer with 4% paraformaldehyde, incubated at 37°C for 10 minutes, and washed twice in FACS buffer. The stained cells were analyzed by flow cytometry using a BD LSRFortessa.

The VH-9/VL-7-(myc-His6) protein in the transfected cell supernatant showed binding to the TF-1 cells by this assay, as did VH-P/VL-P-(myc-His6) protein and VH-9/VL- P-(myc-His6) proteins. The signal by flow cytometry was roughly proportional to the amount of scFv protein in the transfected supernatants, as determined by Western blot.

A variant linker between the heavy and light chain elements may also be used within the scFv, as shown for example in the example sequence below.

QVKL QQ S GAE WKP GAS VKL S C KAS GYTFNSYF MHWMKQAPGQGLEWI GM IRPNGGTTD YNEKFQGRATLTVDKSKSTAYMELSRLRSGDSAVYY CARWEGSYYA LDYWGQGTTVTVSS GGGSSSGGGSSSGGGSS DIQLTQSPSSLSASVGDRVTMTCRAS SNVKY MYWYQQKPGKAPKLWIY YTS NLASGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK (SEQ ID NO: 62)

The scFv sequence above (SEQ ID NO: 62) is used to construct a fusion protein with an erythropoietin, such as the following, which is based on VH-9, VK-7, and EPO_RI_SOA- QVKL QQ S GAE WKP GAS VKL S C KAS GYTFNSYF MHWMKQAPGQGLEWI GM IRPNGGTTD YNEKFQGRATLTVDKSKSTAYMELSRLRSGDSAVYY CARWEGSYYA LDYWGQGTTVTVSS GGGSSSGGGSSSGGGSSDIQLTQSPSSLSASVGDRVTMTCRAS SNVKY MYWYQQKPGKAPKLWIY YTS NLASGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK GGSSGGGGSGGGSSGGGSSSGGGGSGGGGSSGGGSGGGS APPRLICDSRVLERYLLEAKEAEKITTGCAEHCSLNEKITVPDTKVNFYAWKRMEVG

QQAVEVWQGLALLSEAVLRGQALLVKSSQPWEPLQLHVDKAVSGLRSLTTLLRALGAQKEAI SPPDAA SAAPLRTITADTFRKLFRVYSNFLAGKLKLYTGEACRTGDR (SEQ ID NO: 63)

In addition, the scFv sequence of SEQ ID NO: 61 was used to construct EPO fusion proteins with the following sequences. These sequences (1) either possess or lack an epitope and His6 tag at the EPO C-terminus; (2) have one of two different variant linker sequences between the scFv and the EPO moiety. QVKLQQSGAEWKPGASVKLSCKASGYTFNSYFMHWMKQAPGQGLEWIGMIRPNGGTTDYNEKFQGRA TLTVDKSKSTAYMELSRLRSGDSAVYYCARWEGSYYALDYWGQGTTVTVSSGGGGSGGGGSSGGGGSS DIQLTQSPSSLSASVGDRVTMTCRASSNVKYMYWYQQKPGKAPKLWIYYTSNLASGVPSRFSGSGSGT DYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIKGGSSGGGGSGGGSSGGGSSSGGGGSGGGGS SGGGSGGGSAPPRLICDSRVLERYLLEAKEAENITTGCAEHCSLNENITVPDTKVNFYAWKRMEVGQQ AVEVWQGLALLSEAVLRGQALLVNSSQPWEPLQLHVDKAVSGLRSLTTLLRALGAQKEAI SPPDAASA APLRTITADTFRKLFRVYSNFLAGKLKLYTGEACRTGDRGPEQKLI SEEDLNSAVDHHHHHH (SEQ ID NO: 69)

QVKLQQSGAEWKPGASVKLSCKASGYTFNSYFMHWMKQAPGQGLEWIGMIRPNGGTTDYNEKFQGRA TLTVDKSKSTAYMELSRLRSGDSAVYYCARWEGSYYALDYWGQGTTVTVSSGGGGSGGGGSSGGGGSS DIQLTQSPSSLSASVGDRVTMTCRASSNVKYMYWYQQKPGKAPKLWIYYTSNLASGVPSRFSGSGSGT DYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIKGGSSGGGGSGGGSSGGGSSSGGGGSGGGGS SGGGSGGGSAPPRLICDSRVLERYLLEAKEAENITTGCAEHCSLNENITVPDTKVNFYAWKRMEVGQQ AVEVWQGLALLSEAVLRGQALLVNSSQPWEPLQLHVDKAVSGLRSLTTLLRALGAQKEAI SPPDAASA APLRTITADTFRKLFRVYSNFLAGKLKLYTGEACRTGDR (SEQ ID NO: 70)

QVKLQQSGAEWKPGASVKLSCKASGYTFNSYFMHWMKQAPGQGLEWIGMIRPNGGTTDYNEKFQGRA TLTVDKSKSTAYMELSRLRSGDSAVYYCARWEGSYYALDYWGQGTTVTVSSGGGSSGGGGSSGGGGSS DIQLTQSPSSLSASVGDRVTMTCRASSNVKYMYWYQQKPGKAPKLWIYYTSNLASGVPSRFSGSGSGT DYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIKGGSSGGGSSGGGSSGGGSSSGGGSSGGGGS SGGSSGGGSAPPRLICDSRVLERYLLEAKEAENITTGCAEHCSLNENITVPDTKVNFYAWKRMEVGQQ AVEVWQGLALLSEAVLRGQALLVNSSQPWEPLQLHVDKAVSGLRSLTTLLRALGAQKEAI SPPDAASA APLRTITADTFRKLFRVYSNFLAGKLKLYTGEACRTGDRGPEQKLI SEEDLNSAVDHHHHHH (SEQ ID NO: 71)

QVKLQQSGAEWKPGASVKLSCKASGYTFNSYFMHWMKQAPGQGLEWIGMIRPNGGTTDYNEKFQGRA TLTVDKSKSTAYMELSRLRSGDSAVYYCARWEGSYYALDYWGQGTTVTVSSGGGSSGGGGSSGGGGSS DIQLTQSPSSLSASVGDRVTMTCRASSNVKYMYWYQQKPGKAPKLWIYYTSNLASGVPSRFSGSGSGT DYTLTI SSLQPEDVATYYCQQFTSSPYTFGQGTKLEIKGGSSGGGSSGGGSSGGGSSSGGGSSGGGGS SGGSSGGGSAPPRLICDSRVLERYLLEAKEAENITTGCAEHCSLNENITVPDTKVNFYAWKRMEVGQQ AVEVWQGLALLSEAVLRGQALLVNSSQPWEPLQLHVDKAVSGLRSLTTLLRALGAQKEAI SPPDAASA APLRTITADTFRKLFRVYSNFLAGKLKLYTGEACRTGDR ( SEQ ID NO: 72)

The first of these protein sequences was encoded by the following nucleic acid sequence. The other protein sequences were encoded by variations of the following sequence in which, for example, the element encoding the myc and His6 tag was deleted. These nucleic acid sequence were inserted into pSecTag2 for expression in mammalian cells.

CAAGTTAAGTTGCAGCAATCTGGTGCTGAAGTGGTGAAGCCAGGTGCTTCTGTTAAGTTGTCTTGTAA GGCTTCTGGTTACACCTTCAACTCTTACTTTATGCATTGGATGAAGCAAGCACCAGGTCAAGGTTTGG AATGGATCGGTATGATTAGACCAAACGGTGGTACTACCGATTACAATGAGAAGTTTCAGGGCAGGGCT ACTTTGACTGTTGATAAATCCAAGTCCACTGCTTACATGGAATTGTCCAGGTTGAGGTCTGGTGATAG TGCTGTTTACTACTGTGCTAGATGGGAAGGTTCTTACTACGCTTTGGATTACTGGGGTCAAGGTACCA CTGTTACTGTTTCTTCCGGTGGAGGTGGATCTGGTGGTGGAGGATCTTCAGGAGGTGGTGGATCTTCC GATATTCAGTTGACTCAATCTCCATCTAGTTTGTCTGCTAGTGTGGGTGACAGGGTTACTATGACTTG TAGAGCTTCATCTAACGTTAAGTACATGTACTGGTACCAACAGAAGCCTGGTAAGGCTCCAAAGTTGT GGATTTACTACACTTCTAACTTGGCTTCTGGTGTTCCAAGTAGATTTTCTGGTTCAGGTTCTGGTACT GACTACACTTTGACTATTTCCTCTCTTCAACCTGAAGATGTCGCTACTTACTACTGTCAACAATTCAC TTCTTCCCCATACACTTTTGGACAAGGTACTAAGTTGGAAATCAAGGGTGGAAGTAGTGGTGGAGGTG GTTCCGGAGGAGGAAGTTCCGGTGGTGGATCTTCTTCTGGAGGTGGAGGATCCGGTGGTGGAGGATCA TCTGGTGGAGGATCTGGTGGTGGTTCCGCTCCACCTAGATTGATTTGTGATTCCAGAGTTTTGGAAAG ATACTTGTTGGAAGCTAAGGAGGCTGAAAATATTACTACTGGTTGTGCTGAACATTGTTCTTTGAACG

AGAATATTACTGTTCCAGATACTAAGGTTAACTTTTACGCTTGGAAGAGAATGGAAGTTGGTCAGCAA

GCTGTTGAAGTTTGGCAAGGTTTGGCTTTGTTGTCTGAAGCTGTTTTGAGAGGTCAAGCTTTGTTGGT

TAATTCTTCTCAACCATGGGAACCATTGCAATTGCATGTTGATAAGGCTGTTTCTGGTTTGAGATCTT

TGACTACCTTGTTGAGAGCTTTGGGTGCTCAAAAGGAAGCTATTTCTCCTCCAGATGCTGCTTCTGCC

GCTCCATTGAGAACTATTACTGCTGATACTTTTAGAAAGTTGTTTAGAGTTTACTCTAACTTCTTGGC

CGGTAAGTTGAAGTTGTACACTGGTGAAGCTTGTAGAACTGGTGATCGGGGGCCCGAACAAAAACTCA

TCTCAGAAGAGGATCTGAATAGCGCCGTCGACCATCATCATCATCATCATTGA (SEQ ID NO:

73)

Example 9: 10F7-EPO fusion proteins with additional glycosylation sites in the EPO moiety

An insight of the invention is that it is often advantageous to fuse anti-glycophorin A variable regions of an scFv to a mutant form of erythropoietin that has a reduced on-rate for its receptor. However, the vast majority of loss-of-function mutations in hormones, cytokines and similar ligands have the effect of primarily enhancing the off-rate only or both reducing the on-rate and enhancing the off-rate. For example, the mutation R150A in EPO enhances the off-rate of EPO but has no effect on the on-rate, as described in Example 6. A further insight of the invention is that mutations in EPO that create additional N-linked glycosylation sites that do not sterically interfere with receptor binding will generally tend to decrease the on-rate, and will thus be useful in this context. For example, the following mutations in the EPO moiety may be useful, either individually or in combination:

Ala30Asn and His32Thr

Pro87Val, Trp88Asn, and Pro90Thr

Arg53Asn and Glu55Thr

Accordingly, the chimeric activators provided by the following sequences may be particularly useful. The first sequence is based on (VH-9, VK-7, and EPO containing the following mutations: Ala30Asn, His32Thr, Pro87Val, Trp88Asn, and Pro90Thr. The mutations in EPO relative to the wild type sequence of EPO are indicated in boldface and underlined.

Humanized 10F7 with highly glycosylated EPO #1:

QVKLQQSGAEWKPGASVKLSCKAS GYTFNSYF MHWMKQAPGQGLEWIGM IRPNGGTTD YNEKFQGRATLTVDKSKSTAYMELSRLRSGDSAVYY CARWEGSYYA LDYWGQGTTVTVSS GGGSSSGGGSSSGGGSS DIQLTQSPSSLSASVGDRVTMTCRAS SNVKY MYWYQQKPGKAPKLWIY YTS NLASGVPSRFSGSGSGTDYTLTISSLQPEDVATYYC QQFTSSPYTFGQGTKLEIK GGSSGGGGSGGGSSGGGSSSGGGGSGGGGSSGGGSGGGS

apprlicdsrvlerylleakeaenittgene cslnenitvpdtkvnfya krmevgqqavev qgla llseavlrgqallvnssqvnetlqlhvdkavsglrslttllralgaqkeaisppdaasaaplrtitad tfrklfrvysnflrgklklytgeacrtgdr (SEQ ID NO: 64)

The following sequence also includes Arg53Asn and Glu55Thr mutations in the EPO element.

Humanized 10F7 with highly glycosylated EPO #2.

apprlicdsrvlerylleakeaenittgene cslnenitvpdtkvnfyawknmtvgqqavevwqgla llseavlrgqallvnssqvnetlqlhvdkavsglrslttllralgaqkeaisppdaasaaplrtitad tfrklfrvysnflrgklklytgeacrtgdr

(SEQ ID NO: 65)

Example 10: Clinical use of the chimeric activators

A patient with kidney failure and anemia is treated as follows with an engineered protein/chimeric activator described herein. Typical kidney failure patients undergo dialysis to remove toxins such as urea from the blood. The dialysis process involves an intravenous line, so after completion of dialysis, a 10F7-EPO chimeric activator is administered through this intravenous line. In some embodiments, intravenous administration is a preferred mode of administration because potential immunogenicity is further reduced.

In one treatment paradigm, a naive 70 kg patient with a hemoglobin of <10 is given a dose of about between 0.01 and 100 mcg/kg, more preferably between 0.1 and 10 mcg/kg, still more preferably between 0.3 and 3 mcg/kg, and most preferably about 1 mcg/kg. The hematocrit and/or hemoglobin levels are assessed over the course of four weeks. The target hematocrit is typically set at 31, 39 or 42 depending on the patient's status and needs. After 4 weeks, a new dose is administered based on the response to the original dose, with the goals of ultimately achieving the target hematocrit and/or hemoglobin level, but also with the goal of not exceeding an increase of 1.0 g/dL hemoglobin (corresponding to a hematocrit increase of 3) in any 2-week period. Intravenous iron may be administered during the same patient visit or during the same general treatment period as the chimeric activator. Dosage is typically once per two weeks or once per 4 weeks. It is understood that the response to a 10F7-EPO chimeric activator may be variable from patient to patient, because underlying inflammation or other disease states may blunt the response to the product. Example 11; 10F7-EPO chimeric activators with mutations in the EPO moiety affecting a distinct interaction surface.

EPO generally binds to a homodimer of the EPO receptor on cells that control RBC formation and clotting, but may also bind to a heterodimer of the EPO receptor and the more common receptor CD 131, for example on cardiac cells and neurons (FIGs. 18A-18C).

Binding of the EPO-R/CD131 heterodimer promotes survival of these and other cells during hypoxia. Others have identified mutations in EPO, such as Serl04Ile, that sterically block binding of EPO to the EPO receptor homodimer but still allow binding to EPO-R/CD131 heterodimeric receptors.

When EPO binds to the EPO receptor dimer, EPO presents two different surfaces to each EPO receptor. The structure of EPO binding to its receptors is described by Syed et al. Nature 395:511 (1998) and the Protein DataBase structure file 1CN4. The interaction surface that includes helix 4 of EPO mediates a relatively strong contact, while the interaction surface that includes helix 3 of EPO, from amino acids Glu89 to Leu 112 mediates a weaker interaction; helix 1 contributes amino acids to both receptor interactions.

It is an insight of the invention that when EPO binds to the EPO-R/CD131 heterodimer, CD 131 replaces the EPO-R that contacts helix 3 and makes the weaker interaction. It is observed that certain mutations in EPO helix 3, helix 1, and nearby residues have the effect of reducing but not abolishing the 'weak' interaction with EPO-R in a way that can be rescued by fusion of EPO to a glycophorin-binding element, while also not affecting the interaction with CD 131. Thus, fusions of an EPO bearing such mutations to a 10F7 scFv have the effect of enhancing RBC production, not enhancing thrombosis, and enhancing hypoxia resistance (FIGs. 18D-18F). The table below illustrates representative mutations and their effects.

Table 4: The cellular effects of representative EPO mutations

Preferred mutations on the weak face of EPO are in the residues Leu5, Argl4, Leu93, Asp96, Lys97, Ser 100, Argl03, Serl04, Thrl07, and Leul08. For example, Leu5Val, Leu5Ala, Argl4Lys, Argl4Met, Argl4Ile, Arg 14Ala, Argl4Ser, Leu93Ala, Asp96Arg, Asp96Lys, Lys97Ala, Lys97Ser, SerlOOArg, Argl03Ser, Argl03Met, Argl03ne,

Argl03Lys, Argl03Glu, Serl04Ala, Serl04Gly, Thrl07Ser, Thrl07Ala, Leul08Val, Leul08Ala, and Leul08Ser.

References

1. Jelkmann, W. & Wagner, K. Beneficial and ominous aspects of the pleiotropic action of erythropoietin. Ann Hematol 83, 673-686 (2004).

2. Bunn, H. F. Erythropoietin. Cold Spring Harb Perspect Med 3, (2013).

3. Hoffman, R. et al. Hematology. (Elsevier Health Sciences, 2013)

doi: 10.1016/j.celrep.2012.12.021

4. Shander, A., Cappellini, M. D. & Goodnough, L. T. Iron overload and toxicity: the hidden risk of multiple blood transfusions. Vox Sanguinis 97, 185-197 (2009).

5. Pisoni, R. L. et al. Anemia management and outcomes from 12 countries in the Dialysis Outcomes and Practice Patterns Study (DOPPS). Am J Kidney Dis 44, 94-111 (2004).

6. Bohlius, J. et al. Recombinant human erythropoietins and cancer patients: Updated meta-analysis of 57 studies including 9353 patients. JNCI Journal of the National Cancer Institute 98, 708-714 (2006).

7. Besarab, A. et al. The effects of normal as compared with low hematocrit values in patients with cardiac disease who are receiving hemodialysis and epoetin. N. Engl. J. Med.

339, 584-590 (1998).

8. Fishbane, S. & Besarab, A. Mechanism of increased mortality risk with

erythropoietin treatment to higher hemoglobin targets. Clin J Am Soc Nephrol 2, 1274-1282 (2007).

9. Hedenus, M. et al. Efficacy and safety of darbepoetin alfa in anaemic patients with lymphoproliferative malignancies: a randomized, double-blind, placebo-controlled study. Br. J. Haematol. 122, 394-403 (2003). 10. Leyland- Jones, B. et al. Maintaining normal hemoglobin levels with epoetin alfa in mainly nonanemic patients with metastatic breast cancer receiving first- line chemotherapy: a survival study. J. Clin. Oncol. 23, 5960-5972 (2005).

11. Driieke, T. B. et al. Normalization of hemoglobin level in patients with chronic kidney disease and anemia. N. Engl. J. Med. 355, 2071-2084 (2006).

12. Singh, A. K. et al. Correction of anemia with epoetin alfa in chronic kidney disease. N. Engl. J. Med. 355, 2085-2098 (2006).

13. Wright, J. R. et al. Randomized, double-blind, placebo-controlled trial of erythropoietin in non-small cell lung cancer with disease-related anemia. Journal of Clinical Oncology 25, 1027-1032 (2007).

14. Smith, R. E. et al. Darbepoetin alfa for the treatment of anemia in patients with active cancer not receiving chemotherapy or radiotherapy: Results of a phase III, multicenter, randomized, double-blind, placebo-controlled study. Journal of Clinical Oncology 26, 1040- 1050 (2008).

15. Pfeffer, M. A. et al. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N. Engl. J. Med. 361, 2019-2032 (2009).

16. Stowell, C. P., Jones, S. C, Enny, C, Langholff, W. & Leitz, G. An open-label, randomized, parallel group study of perioperative epoetin alfa versus standard of care for blood conservation in major elective spinal surgery: safety analysis. Spine 34, 2479-2485 (2009).

17. Solomon, S. D. et al. Erythropoietic response and outcomes in kidney disease and type 2 diabetes. N. Engl. J. Med. 363, 1146-1155 (2010).

18. Lewis, E. F. et al. Darbepoetin alfa impact on health status in diabetes patients with kidney disease: A randomized trial. Clin J Am Soc Nephrol 6, 845-855 (2011).

19. Untch, M. et al. PREPARE trial: a randomized phase III trial comparing preoperative, dose-dense, dose-intensified chemotherapy with epirubicin, paclitaxel and CMF versus a standard-dosed epirubicin/cyclophosphamide followed by paclitaxel darbepoetin alfa in primary breast cancer— results at the time of surgery. Annals of Oncology 22, 1988-1998 (2011).

20. Ross, S. D. et al. Clinical benefits and risks associated with epoetin and darbepoetin in patients with chemotherapy-induced anemia: a systematic review of the literature. Clin Ther 28, 801-831 (2006).

21. Glaspy, J. et al. Erythropoiesis- stimulating agents in oncology: a study-level metaanalysis of survival and other safety outcomes. Br J Cancer 102, 301-315 (2010). 22. Seliger, S. L. et al. Erythropoiesis-stimulating agents increase the risk of acute stroke in patients with chronic kidney disease. Kidney International 80, 288-294 (2011).

23. Henke, M. et al. Erythropoietin to treat head and neck cancer patients with anaemia undergoing radiotherapy: randomised, double-blind, placebo-controlled trial. The Lancet 362, 1255-1260 (2003).

24. Bennett, C. L. et al. Venous thromboembolism and mortality associated with recombinant erythropoietin and darbepoetin administration for the treatment of cancer- associated anemia. JAMA 299, 914-924 (2008).

25. Bohlius, J. et al. Recombinant human erythropoiesis-stimulating agents and mortality in patients with cancer: a meta-analysis of randomised trials. Lancet 373, 1532-

1542 (2009).

26. Tonelli, M. et al. Benefits and harms of erythropoiesis-stimulating agents for anemia related to cancer: a meta-analysis. Canadian Medical Association 180, E62-E71

(2009) .

27. Hess, G., Nordyke, R. J., Hill, J. & Hulnick, S. Effect of reimbursement changes on erythropoiesis stimulating agent utilization and transfusions. Am. J. Hematol. 85, 838-843

(2010) .

28. Rizzo, J. D. et al. American Society of Clinical Oncology/ American Society of Hematology Clinical practice guideline update on the use of epoetin and darbepoetin in adult patients with cancer. Journal of Clinical Oncology 28, 4996-5010 (2010).

29. Adamson, J. W. Erythropoietic-stimulating agents: the cancer progression controversy and collateral damage to the blood supply. Transfusion 49, 824-826 (2009).

30. Vekeman, F. et al. Impact of limiting erythropoiesis-stimulating agent use for chemotherapy-induced anemia on the United States blood supply margin. Transfusion 49, 895-902 (2009).

31. Vainchenker, W., Bouguet, J., Guichard, J. & Breton-Gorius, J. Megakaryocyte colony formation from human bone marrow precursors. Blood 54, 940-945 (1979).

32. Ishibashi, T., Koziol, J. A. & Burstein, S. A. Human recombinant erythropoietin promotes differentiation of murine megakaryocytes in vitro. J. Clin. Invest. 79, 286-289 (1987).

33. Degos, L., Linch, D. C. & Lowenberg, B. Textbook of Malignant Hematology. (CRC Press, 2004). doi: 10.1093/emboj/21.7.1616 34. Berridge, M. V., Fraser, J. K., Carter, J. M. & Lin, F. K. Effects of recombinant human erythropoietin on megakaryocytes and on platelet production in the rat. Blood 72, 970-977 (1988).

35. Tsukada, J. et al. The effect of high doses of recombinant human erythropoietin on megakaryocytopoiesis and platelet production in splenectomized mice. Br. J. Haematol. 76,

260-268 (1990).

36. McDonald, T. P., Clift, R. E. & Cottrell, M. B. Large, chronic doses of erythropoietin cause thrombocytopenia in mice. Blood 80, 352-358 (1992).

37. Wolf, R. F. et al. Erythropoietin administration increases production and reactivity of platelets in dogs. Thromb Haemost 78, 1505-1509 (1997).

38. Wolf, R. F. et al. Erythropoietin potentiates thrombus development in a canine arterio-venous shunt model. Thromb Haemost 77, 1020-1024 (1997).

39. Loo, M. & Beguin, Y. The effect of recombinant human erythropoietin on platelet counts is strongly modulated by the adequacy of iron supply. Blood 93, 3286-3293 (1999).

40. Kirkeby, A. et al. High-dose erythropoietin alters platelet reactivity and bleeding time in rodents in contrast to the neuroprotective variant carbamyl-erythropoietin (CEPO). Thromb Haemost (2008). doi: 10.1160/TH07-03-0208

41. Eschbach, J. W., Egrie, J. C, Downing, M. R., Browne, J. K. & Adamson, J. W. Correction of the anemia of end-stage renal disease with recombinant human erythropoietin. Results of a combined phase I and II clinical trial. N. Engl. J. Med. 316, 73-78 (1987).

42. Eschbach, J. W. et al. Recombinant human erythropoietin in anemic patients with end-stage renal disease. Results of a phase III multicenter clinical trial. Ann. Intern. Med. 111, 992-1000 (1989).

43. Fabris, F. et al. Effect of human recombinant erythropoietin on bleeding time, platelet number and function in children with end- stage renal disease maintained by haemodialysis. Pediatr. Nephrol. 5, 225-228 (1991).

44. Cases, A. et al. Recombinant human erythro-poietin treatment improves platelet function in uremic patients. Kidney Int 42, 668-672 (1992).

45. Taylor, J. E., McLaren, M., Henderson, I. S., Belch, J. J. & Stewart, W. K.

Prothrombotic effect of erythropoietin in dialysis patients. Nephrol. Dial. Transplant. 7, 235- 239 (1992).

46. Roger, S. D. et al. Enhanced platelet reactivity with erythropoietin but not following transfusion in dialysis patients. Nephrol. Dial. Transplant. 8, 213-217 (1993). 47. Beguin, Y. et al. Effect of recombinant human erythropoietin on platelets in patients with anemia of renal failure: correlation of platelet count with erythropoietic activity and iron parameters. Eur. J. Haematol. 53, 265-270 (1994).

48. Malyszko, J. S., Malyszko, J., Pawlak, K. & Mysliwiec, M. Erythropoietin and uremic platelet aggregation in vivo and in vitro. Int. J. Clin. Lab. Res. 26, 199-202 (1996).

49. Dale, G. L. & Alberio, L. Is there a correlation between raised erythropoietin and thrombotic events in sickle-cell anaemia? Lancet 352, 566-567 (1998).

50. Beguin, Y. Erythropoietin and platelet production. Haematologica 84, 541-547

(1999).

51. Stohlawetz, P. J. et al. Effects of erythropoietin on platelet reactivity and thrombopoiesis in humans. Blood 95, 2983-2989 (2000).

52. Ando, M. et al. Circulating platelet-derived microparticles with procoagulant activity may be a potential cause of thrombosis in uremic patients. Kidney International 62, 1757-1763 (2002).

53. Chuang, Y. C, Chen, J. B., Yang, L. C. & Kuo, C. Y. Significance of platelet activation in vascular access survival of haemodialysis patients. Nephrology Dialysis Transplantation 18, 947-954 (2003).

54. Vaziri, N. D. & Zhou, X. J. Potential mechanisms of adverse outcomes in trials of anemia correction with erythropoietin in chronic kidney disease. Nephrology Dialysis Transplantation 24, 1082-1088 (2008).

55. Heinisch, B. B. et al. The effect of erythropoietin on platelet and endothelial activation markers: A prospective trial in healthy volunteers. Platelets 23, 352-358 (2012).

56. Szczech, L. A. et al. Secondary analysis of the CHOIR trial epoetin-a dose and achieved hemoglobin outcomes. Kidney International 74, 791-798 (2008).

57. Bitonti, A. J. et al. Pulmonary delivery of an erythropoietin Fc fusion protein in non-human primates through an immunoglobulin transport pathway. Proc Natl Acad Sci USA 101, 9763-9768 (2004).

58. Way, J. C. et al. Improvement of Fc-erythropoietin structure and pharmacokinetics by modification at a disulfide bond. Protein Engineering Design and Selection 18, 111-118 (2005).

59. Long, D. L. et al. Design of homogeneous, monopegylated erythropoietin analogs with preserved in vitro bioactivity. Experimental Hematology 34, 697-704 (2006). 60. Sathyanarayana, P. et al. CNTO 530 functions as a potent EPO mimetic via unique sustained effects on bone marrow proerythroblast pools. Blood 113, 4955-4962 (2009).

61. Bouman-Thio, E. et al. A phase I, single and fractionated, ascending-dose study evaluating the safety, pharmacokinetics, pharmacodynamics, and immunogenicity of an erythropoietin mimetic antibody fusion protein (CNTO 528) in healthy male subjects. The Journal of Clinical Pharmacology 48, 1197-1207 (2013).

62. Pastan, I., Hassan, R., FitzGerald, D. J. & Kreitman, R. J. Immunotoxin therapy of cancer. Nat Rev Cancer 6, 559-565 (2006).

63. Schrama, D., Reisfeld, R. A. & Becker, J. C. Antibody targeted drugs as cancer therapeutics. Nat Rev Drug Discov 5, 147-159 (2006).

64. Teicher, B. A. Antibody-drug conjugate targets. Current cancer drug targets 9, 982

(2009).

65. Winkler, J. Nanomedicines based on recombinant fusion proteins for targeting therapeutic siRNA oligonucleotides. Therapeutic delivery 2, 891-905 (2011).

66. Vugmeyster, Y. Pharmacokinetics and toxicology of therapeutic proteins:

Advances and challenges. WJBC 3, 73 (2012).

67. Chen, X., Lee, H.-F., Zaro, J. L. & Shen, W.-C. Effects of receptor binding on plasma half-life of bifunctional transferrin fusion proteins. Mol. Pharmaceutics 8, 457-465 (2011).

68. Arthur J Atkinson, J., Shiew-Mei Huang, P. D., Lertora, J. J. L. & Sanford P Markey, P. D. Principles of Clinical Pharmacology. (2012).

69. Robinson-Mosher, A., Shinar, T., Silver, P. A. & Way, J. Dynamics simulations for engineering macromolecular interactions. Chaos: An Interdisciplinary Journal of

Nonlinear Science 23, 025110-025110-13 (2013).

70. Cironi, P., Swinburne, I. A. & Silver, P. A. Enhancement of cell type specificity by quantitative modulation of a chimeric ligand. Journal of Biological Chemistry 283, 8469- 8476 (2008).

71. Taylor, N. D., Way, J. C, Silver, P. A. & Cironi, P. Anti-glycophorin single-chain Fv fusion to low affinity mutant erythropoietin improves red blood cell-lineage specificity.

Protein Engineering Design and Selection 23, 251-260 (2010).

72. Bigbee, W. L., Vanderlaan, M., Fong, S. S. & Jensen, R. H. Monoclonal antibodies specific for the Mand N-forms of human glycophorin A. Mol. Immunol. 20, 1353- 1362 (1983). 73. Okumura, N., Tsuji, K. & Nakahata, T. Changes in cell surface antigen expressions during proliferation and differentiation of human erythroid progenitors. Blood 80, 642-650 (1992).

74. Southcott, M. J., Tanner, M. J. & Anstee, D. J. The expression of human blood group antigens during erythropoiesis in a cell culture system. Blood 93, 4425-4435 (1999).

75. Elliott, S., Lorenzini, T., Chang, D., Barzilay, J. & Delorme, E. Mapping of the active site of recombinant human erythropoietin. Blood 89, 493-502 (1997).

76. Cheetham, J. C. et al. NMR structure of human erythropoietin and a comparison with its receptor bound conformation. Nature Structural & Molecular Biology 5, 861-866 (1998).

77. Kontos, S. & Hubbell, J. A. Improving protein pharmacokinetics by engineering erythrocyte affinity. Mol. Pharmaceutics 7, 2141-2147 (2010).

78. Agre, P. Red Blood Cell Membranes. (CRC Press, 1989).

79. Kontos, S. & Hubbell, J. A. Drug development: longer-lived proteins. Chem. Soc. Rev. 41, 2686 (2012).

80. Kontos, S., Kourtis, I. C, Dane, K. Y. & Hubbell, J. A. Engineering antigens for in situ erythrocyte binding induces T-cell deletion. Proceedings of the National Academy of Sciences 110, E60-8 (2013).

82. Merry, A. H., Hodson, C, Thomson, E., Mallinson, G. & Anstee, D. J. The use of monoclonal antibodies to quantify the levels of sialoglycoproteins alpha and delta and variant sialoglycoproteins in human erythrocyte membranes. Biochem J 233, 93-98 (1986).

83. Loken, M. R., Shah, V. O., Dattilio, K. L. & Civin, C. I. Flow cytometric analysis of human bone marrow: I. Normal erythroid development. Blood 69, 255-263 (1987).

84. Holt, L. J. et al. Anti-serum albumin domain antibodies for extending the half- lives of short lived drugs. Protein Eng. Des. Sel. 21, 283-288 (2008).

85. Wrighton, N. C. et al. Increased potency of an erythropoietin peptide mimetic through covalent dimerization. Nat Biotechnol 15, 1261-1265 (1997).

86. Casadevall, N. et al. Pure red-cell aplasia and anti-erythropoietin antibodies in patients treated with recombinant erythropoietin. N. Engl. J. Med. 346, 469-475 (2002).

87. Fan, Q. et al. Preclinical evaluation of Hematide, a novel erythropoiesis stimulating agent, for the treatment of anemia. Experimental Hematology 34, 1303-1311 (2006). 88. Stead, R. B. Evaluation of the safety and pharmacodynamics of Hematide, a novel erythropoietic agent, in a phase 1, double-blind, placebo-controlled, dose-escalation study in healthy volunteers. Blood 108, 1830-1834 (2006).

89. Macdougall, I. C. et al. A peptide-based erythropoietin-receptor agonist for pure red-cell aplasia. N. Engl. J. Med. 361, 1848-1855 (2009).

90. Kaushik, T. & Yaqoob, M. M. Lessons learned from peginesatide in the treatment of anemia associated with chronic kidney disease in patients on dialysis. BTT 7, 243-246 (2013).

91. Ferguson, T. A., Choi, J. & Green, D. R. Armed response: how dying cells influence T-cell functions. Immunological Reviews 241, 77-88 (2011).

92. Griffith, T. S. & Ferguson, T. A. Cell Death in the maintenance and abrogation of tolerance: The Five Ws of Dying Cells. Immunity 35, 456-466 (2011).

93. Nimmerjahn, F. & Ravetch, J. V. Fey receptors as regulators of immune responses. Nature Reviews Immunology 8, 34-47 (2008).

94. Pluckthun, A. & Pack, P. New protein engineering approaches to multivalent and bispecific antibody fragments. Immunotechnology 3, 83-105 (1997).

95. Sou, K., Goins, B., Oyajobi, B. O., Travi, B. L. & Phillips, W. T. Bone marrow- targeted liposomal carriers. Expert Opin. Drug Deliv. 8, 317-328 (2011).

96. Miyazaki, Y. et al. Therapeutic impact of erythropoietin-encapsulated liposomes targeted to bone marrow on renal anemia. Mol. Pharmaceutics 141006162142007 (2014). doi: 10.1021/mp500453a

97. Sasaki, R., Yanagawa, S., Hitomi, K. & Chiba, H. Characterization of

erythropoietin receptor of murine erythroid cells. Eur J Biochem 168, 43-48 (1987).

98. Lee, L. G., Chen, C. H. & Chiu, L. A. Thiazole orange: a new dye for reticulocyte analysis. Cytometry 7, 508-517 (1986).

99. Ritchie, J. L., Alexander, H. D. & Rea, I. M. Flow cytometry analysis of platelet P-selectin expression in whole blood— methodological considerations. Clin Lab Haematol 22, 359-363 (2000).

100. Kamath, S. Platelet activation: assessment and quantification. European Heart Journal 22, 1561-1571 (2001).

101. Honegger, A., Malebranche, A. D., Rothlisberger, D. & Pluckthun, A. The influence of the framework core residues on the biophysical properties of immunoglobulin heavy chain variable domains. Protein Engineering Design and Selection 22, 121-134 (2008). 102. Du, J. et al. Structural basis for recognition of CD20 by therapeutic antibody rituximab. J Biol Chem 282, 15073-15080 (2007).

103. Epogen [Package insert]. Thousand Oaks, CA: Amgen; 2005.

104. Besarab, A. et al. Efficacy and tolerability of intravenous continuous erythropoietin receptor activator: a 19-week, phase II, multicenter, randomized, open-label, dose-finding study with a 12-month extension phase in patients with chronic renal disease. Clin Ther 29, 626-639 (2007).

105. Bennett, C. L. et al. Venous thromboembolism and mortality associated with recombinant erythropoietin and darbepoetin administration for the treatment of cancer- associated anemia. JAMA 299, 914-924 (2008).

106. Social Security: Official Social Security Website.

www.ssa.gov/oact/STATS/table4c6.html

107. Liu Y, et al. (2006) The Antimelanoma Immunocytokine scFvMEL/TNF Shows Reduced Toxicity and Potent Antitumor Activity against Human Tumor Xenografts.

Neoplasia 8(5):384-393.

108. Bagshawe KD (2009) Targeting: the ADEPT story so far. Current drug targets.

109. Kreitman RJ, Pastan I (2011) Antibody Fusion Proteins: Anti-CD22

Recombinant Immunotoxin Moxetumomab Pasudotox. Clinical Cancer Research

17(20):6398-6405.

110. Gillies SD (2013) A new platform for constructing antibody-cytokine fusion proteins (immunocytokines) with improved biological properties and adaptable cytokine activity. Protein Engineering Design and Selection 26(10):561-569.

111. Sadelain M, Brentjens R, Riviere I (2013) The Basic Principles of Chimeric

Antigen Receptor Design. Cancer Discovery 3(4):388-398.

112. Huehls AM, Coupet TA, Sentman CL (2014) Bispecific T-cell engagers for cancer immunotherapy. Immunology and Cell Biology 93(3):290-296.

113. Moraga I, et al. (2015) Tuning Cytokine Receptor Signaling by Re-orienting Dimer Geometry with Surrogate Ligands. Cell: 1-14.

114. Lauffenburger DA, Fallon EM, Haugh JM (1998) Scratching the (cell) surface: cytokine engineering for improved ligand/receptor trafficking dynamics. Chemistry &

Biology 5(10):R257-63.

115. Inoue M, Nakayama C, Noguchi H (1996) Activating mechanism of CNTF and related cytokines. Mol Neurobiol 12(3): 195-209. 116. Garcin G, et al. (2014) High efficiency cell-specific targeting of cytokine activity. Nature Communications 5:3016.

117. Anagnostou A, et al. (1994) Erythropoietin receptor mRNA expression in human endothelial cells. Proc Natl Acad Sci USA 91(9):3974-3978.

118. Jelkmann W, Elliott S (2013) Erythropoietin and the vascular wall: The controversy continues. Nutrition, Metabolism and Cardiovascular Diseases 23(S 1):S37-S43.

119. Thurber GM, Schmidt MM, Wittrup KD (2008) Antibody tumor penetration: transport opposed by systemic and antigen-mediated clearance. Advanced Drug Delivery Reviews 60(12): 1421-1434.

120. Catimel B, Wilson KM, Kemp BE (1993) Kinetics of the autologous red cell agglutination test. J Immunol Methods 165(2): 183-192.

121. Rahuel C, et al. (1988) Alteration of the genes for glycophorin A and B in glycophorin-A-deficient individuals. Eur J Biochem 177(3):605-614.

122. Landschulz KT, Boyer SH, Noyes AN, Rogers OC, Frelin LP (1992) Onset of erythropoietin response in murine erythroid colony-forming units: assignment to early S- phase in a specific cell generation. Blood 79(10):2749-2758.

123. Auffray I (2001) Glycophorin A dimerization and band 3 interaction during erythroid membrane biogenesis: in vivo studies in human glycophorin A transgenic mice. Blood 97(9):2872-2878.

124. Rearden A (1986) Evolution of glycophorin A in the hominoid primates studied with monoclonal antibodies, and description of a sialoglycoprotein analogous to human glycophorin B in chimpanzee. J Immunol 136(7):2504-2509.

125. Egrie JC, Dwyer E, Browne JK, Hitz A, Lykos MA (2003) Darbepoetin alfa has a longer circulating half-life and greater in vivo potency than recombinant human

erythropoietin. Experimental Hematology 31(4):290-299.

126. Way JC, et al. (2005) Improvement of Fc -erythropoietin structure and pharmacokinetics by modification at a disulfide bond. Protein Engineering Design and Selection 18(3): 111-118.

127. Trost N, et al. (2013) Recombinant human erythropoietin alters gene expression and stimulates proliferation of MCF-7 breast cancer cells. Radiology and Oncology

47(4):382-389.

128. Lester RD, Jo M, Campana WM, Gonias SL (2005) Erythropoietin Promotes MCF-7 Breast Cancer Cell Migration by an ERK/Mitogen-activated Protein Kinase- dependent Pathway and Is Primarily Responsible for the Increase in Migration Observed in Hypoxia. J Biol Chem 280(47):39273-39277.

129. Rinder HM, et al. (1998) Correlation of thrombosis with increased platelet turnover in thrombocytosis. Blood 91(4): 1288-1294.

130. Wiczling P, Krzyzanski W (2007) Method of determination of the reticulocyte age distribution from flow cytometry count by a structured-population model. Cytometry 71A(7):460-467.

131. McBane RD, Gonzalez C, Hodge DO, Wysokinski WE (2013) Propensity for young reticulated platelet recruitment into arterial thrombi. J Thromb Thrombolysis

37(2): 148-154.

132. Macdougall IC (2002) Optimizing the use of erythropoietic agents—

pharmacokinetic and pharmacodynamic considerations. Nephrol Dial Transplant 17 Suppl 5:66-70.

133. Khandelwal S, Saxena RK (2006) Assessment of survival of aging erythrocyte in circulation and attendant changes in size and CD147 expression by a novel two step biotinylation method. Exp Gerontol 41(9): 855-861.

134. Swedberg K, et al. (2013) Treatment of Anemia with Darbepoetin Alfa in Systolic Heart Failure. N Engl J Med 368(13): 1210-1219.

135. Singhal A, Gupta CM (1986) Antibody-mediated targeting of liposomes to red cells in vivo. FEBS Lett 201(2):321-326.

136. Vogel J (2003) Transgenic mice overexpressing erythropoietin adapt to excessive erythrocytosis by regulating blood viscosity. Blood 102(6):2278-2284.

137. Jeong S-K, Cho YI, Duey M, Rosenson RS (2010) Cardiovascular Risks of Anemia Correction with Erythrocyte Stimulating Agents: Should Blood Viscosity Be Monitored for Risk Assessment? Cardiovasc Drugs Ther 24(2): 151-160.

EQUIVALENTS

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the following claims. All references, including patent documents, disclosed herein are incorporated by reference in their entirety.

What is claimed is:

Claims

1. A Glycophorin A-binding protein comprising a sequence selected from the group consisting of SEQ ID NOs: 2-5, 7-9, and 52-60.

2. The Glycophorin A-binding protein of claim 1, wherein the protein comprises an antibody or a glycophorin-A binding fragment thereof.

3. The Glycophorin A-binding protein of claim 2, wherein the protein comprises an scFv.

4. The Glycophorin A-binding protein of any one of claims 1-3, wherein the protein comprises a protein domain that does not bind to Glycophorin A.

5. The Glycophorin A-binding protein of claim 4, wherein the protein comprises a cytokine or hormone.

6. The Glycophorin A-binding protein of claim 5, wherein the cytokine or hormone is a four-helix -bundle protein.

7. The Glycophorin A-binding protein of any one of claims 4-6, wherein the protein comprises an antigen.

8. A nucleic acid sequence encoding a Glycophorin A-binding protein, wherein the nucleic acid sequence comprises a sequence selected from the group consisting of SEQ

ID NOs: 41-49 and 73.

9. A Glycophorin A-binding protein comprising a sequence selected from the group consisting of the sequences in Table 1, wherein the sequence is not one of the murine sequences listed in Table 1.

10. The Glycophorin A-binding protein of any one of claims 5-7, wherein the cytokine or hormone is Erythropoietin (EPO).

11. The Glycophorin A-binding protein of claim 10, wherein EPO contains one or more mutations relative to wild type human EPO.

12. The Glycophorin A-binding protein of claim 11, wherein the one or more mutations relative to wild type human EPO are at an amino acid selected from the group consisting of R150, A30, H32, P87, W88, P90, R53, and E55.

13. The Glycophorin A-binding protein of claim 11 or 12, wherein EPO contains the mutation R150A.

14. The Glycophorin A-binding protein of claim 12 or 13, wherein EPO contains one or more mutation selected from the group consisting of A30N, H32T, P87V, W88N, P90T, R53N, and E55T.

15. The Glycophorin A-binding protein of any one of claims 10-14, wherein the

Glycophorin A-binding protein comprises a linker.

16. The Glycophorin A-binding protein of claim 15, wherein the linker is located between the EPO and a portion of the Glycophorin A-binding protein that binds Glycophorin A.

17. An antibody that selectively binds to Glycophorin A, wherein the antibody comprises a sequence selected from the group consisting of SEQ ID NOs: 2-5,7-9, and 52-60.

18. The antibody of claim 15, wherein the antibody is an scFv.

19. A method comprising administering a therapeutically effective amount of a Glycophorin A-binding protein of claims 1-6 or 8-14 or the antibody of any one of claims 14- 15 to a subject in need thereof.

20. The method of claim 19, wherein the subject has anemia or kidney failure.

21. The method of claim 19 or 20, wherein the therapeutically effective amount of the Glycophorin A-binding protein is administered to the subject more than once.

22. A Glycophorin A-binding protein comprising a sequence selected from the group consisting of SEQ ID NO:39, 63-65, and 68-72.

23. An engineered protein comprising

an scFv that binds to glycophorin A and

an erythropoietin (EPO) moiety that contains at least four N-linked glycosylation sites.

24. An engineered protein comprising

an scFv that binds to glycophorin A, and

an erythropoietin (EPO) moiety that contains one or more mutations that, in the absence of the scFv, reduce binding to a homodimeric EPO receptor by at least about twofold but have less than a two-fold effect on binding to an EPO receptor/CD 131 heterodimer.

25. An engineered protein comprising

an scFv that binds to glycophorin A, and

an erythropoietin (EPO) moiety comprising (i) at least four N-linked glycosylation sites, and (ii) one or more mutations relative to wild type human EPO that, in the absence of the scFv, reduce binding to a homodimer of erythropoietin receptors by at least about twofold but have less than a two-fold effect on binding to a heterodimer of an erythropoietin receptor and CD 131.

26. The engineered protein of claim 24 or 25, wherein the one or more mutations are at an amino acid selected from the group consisting of L5, R14, L93, D96, K97, S 100,

R103, S 104, T107, and L108.

27. The engineered protein of claim 26, wherein the one or more mutations are selected from the group consisting of L5V, L5A, R14K, R14M, R14I, R14A, R14S, L93A, D96R, D96K, K97A, K97S, S 100R, R103S, R103M, R103I, R103K, R103E, S 104A, S 104G, T107S, T107S, T107A, L108V, L108A, L108S. 9/21