Method for producing long-acting EPO-fc fusion protein

Abstract

The present invention relates to a method for producing an Fc fusion protein comprising an EPO-hybrid Fc fusion protein (i.e., epoetin alpha) having a C-terminal lysine content of about 1–10% and a sialic acid content of about 17–27 mol / mol. The method comprises culturing host cells transformed with DNA encoding the Fc fusion protein and DNA encoding human α-2,3-sialyltransferase under optimized culture conditions, and purifying the Fc fusion protein using an optimized purification process including anion exchange chromatography.

Description

[Technical Field] [1] This application is a continuation of U.S. application Ser. No. 13 / 493,508 filed on June 11, 2012, now Patent No. US 8,586,531 B2, Patent No: US 7,867,491 B2 issued on Jan 11, 2011, and Patent No: AU2016334954 B2 filed on Oct 06, 2016. The entire teachings of the above applications are incorporated herein by reference. [2] The present invention relates to methods for producing a long-acting erythropoietin (EPO) Fc fusion protein, and more specifically, to an EPO-Fc-producing cell line for producing a highly sialylated form of long-acting EPO-Fc fusion protein and methods for producing a long-acting erythropoietin (EPO) Fc fusion protein using the cell line. [3] [Background Art] [4] Endogenous erythropoietin (EPO) is a glycoprotein hormone naturally produced by the peritubular cells of the kidney that stimulates red blood cell production, also known as erythropoiesis. Renal cortex peritubular cells produce most EPO in the human body, although in a fetus, the liver is the primary site of production (Jelkmann, 2016). EPO production also occurs in the spleen, liver, bone marrow, lung, and brain in small quantities. The partial pressure of oxygen (pO2) directly regulates EPO production. The lower the pO2, the greater the production of EPO. Indirectly, low hemoglobin levels stimulate EPO production. [5] Erythropoiesis stimulating agents (ESAs) are recombinant versions of EPO produced pharmacologically via recombinant DNA technology in cell cultures. Examples of ESAs include epoetin alfa, darbepoetin alfa, and methoxy polyethylene glycol-epoetin beta (Jelkmann, 2016). The latter two examples are EPO with extended half-life; they carry additional carbohydrate side chains and additional pegylation, respectively. [6] ESAs are generally indicated in conditions where there is impaired red blood cell production. The two primary FDA-approved indications for ESAs are anemia due to chronic kidney disease (CKD) and chemotherapy-induced anemia in patients with cancer (Noxon et al, 2017). Other approved indications are anemia secondary to zidovudine treatment in HIV infection, support in patients receiving autologous blood transfusions, anemic patients undergoing elective surgery (both pre-op and post-op) anemia in preterm infants. [7] In CKD, where there is damage to the kidneys thus a limited EPO production by the kidney peritubular cells, ESA treatment alleviates the anemia condition (Jelkmann, 2013). In CKD, ESAs are effective in patients who are receiving dialysis as well as in non-dialysis patients. [8] Endogenous EPO and ESAs act by binding to its cell surface receptor, the EPO receptor (EPOR), which is expressed at the highest level on erythroid progenitor cells. The binding of EPO triggers the activation of EPOR including dimerization and / or reorientation of the receptor followed by JAK / STAT and PI3K-AKT signaling cascades (Watowich, 2011 and Cokic et al., 2012). The activation of EPOR on erythroid progenitor cells promotes its survival, cell proliferation, and differentiation into reticulocytes. [9] Polypeptides such as EPO generally have low stability and are easily denatured. Further, these polypeptides have a short in vivo half-life because they are degraded by proteases in the blood and are easily eliminated through the kidney or liver. Therefore, in order to maintain the blood concentration and potency of protein therapeutics containing polypeptides as pharmacological ingredients, it is necessary to frequently administer these protein therapeutics to patients. However, in the case of protein therapeutics that are mostly administered to patients in the form of injections, frequent injections cause pain to the patients.

[10] Besides additional glycans and attachment of PEG, fusion of the protein therapeutic with Fc fragment derived from immunoglobulin molecule is known to increase stability and serum half-life. Immunoglobulin is a major component of blood. Human immunoglobulin includes various classes such as IgG, IgM, IgA, IgD, and IgE.

[11] Immunoglobulins are comprised of four polypeptide chains, two heavy chains and two light chains, which are associated via disulfide bonds to form tetramers. Each chain is composed of a variable region and a constant region. The heavy chain constant region is further divided into three or four regions (CH1, CH2, CH3, and CH4), depending on the isotypes. The Fc region of the heavy chain constant region, depending on the Ig isotype, includes hinge, CH2, CH3, and / or CH4 domains.

[12] Since it was discovered that sialic acid-degrading enzyme (sialidase) reduces the in vivo biological activity of EPO, the relationship between sialic acid and the biological activity of EPO have been actively studied. Subsequently, studies have shown that attachment of sialic acids to the surface of recombinant EPO protein prevent protein degradation, extend the in vivo half-life and ultimately promotes the efficacy. Sialic acid affects the in vivo activity of EPO by protecting the galactose residue on EPO surface from binding to the asialoglycoprotein receptors which are highly expressed on the surface of hepatocytes and remove the target glycoprotein from circulation.

[13] Many publications have demonstrated that sugar residues play an essential role in maintaining the inherent efficacy of glycoproteins, and especially in the case of human-derived EPO, the role of sugars is absolute. In other words, the activity of EPO disappears in the absence of sugar residues, and the in vivo activity of EPO is greatly influenced by the composition of the sugars, particularly the content of sialic acid.

[14] Meanwhile, N-acetylmannosamine may be added to the culture medium to increase the sialic acid content of EPO-Fc fusion protein. EPO-Fc protein obtained through this process is usually a mixture of EPO-Fc fusion protein having a high sialic acid content and those with a low sialic acid content. Since EPO-Fc fusion protein with high sialic acid content has a longer in vivo half-life than those with low sialic acid content, there is a need to develop a technology capable of selectively producing EPO-Fc fusion protein having a high sialic acid content.

[15] The production of EPO-Fc fusion protein requires the processes of culturing transfected host cells to express a large amount of EPO-Fc fusion protein, followed by purification of EPO-Fc fusion protein. Thus, to mass-produce high-quality EPO-Fc fusion protein, optimization of the cell culture and purification technologies is very important.

[16] EPO-hybrid Fc fusion protein, called Efepoetin alfa is a homodimer with each subunit comprising of EPO and hybrid Fc derived from human IgD and human IgG4 (FIG. 1). The complete amino acid sequence of Efepoetin alfa is shown in FIG. 2 (SEQ ID NO:2 patent no. US 8,586,531 B2). EPO amino acid sequence in Efepoetin alfa is identical to the native EPO which has a half-life in the blood around 5 hours.

[17] The function of EPO fragment in Efepoetin alfa structure is confirmed by: (1) binding to EPOR by SPR (surface plasmon resonance) assay (FIG. 3a - FIG. 3d), and (2) the ability to induce proliferation of UT-7 cell line in dose dependent manner (FIG. 4). UT-7 cell line is a human leukemia cell line with megakaryocytic features that was established from the bone marrow of a patient with acute megakaryoblastic leukemia. The UT-7 cell line shows strict dependency on IL-3, GM-CSF, or EPO (Komatsu et al., 1991).

[18] The fusion of EPO fragment to hybrid Fc is designed to prolong the half-life of EPO by (1) increasing the molecular weight of the molecule to avoid renal clearance and (2) binding to FcRn (FIG. 5) to avoid cellular degradation. By combining the CH2 and CH3 domains of IgG4 with the CH2 domain and hinge region of IgD, the hybrid Fc construct lacks immune effector activities such as antibody-dependent cellular cytotoxicity (ADCC), antibodydependent cellular phagocytosis (ADCP) (FIG. 6), and complement-dependent cytotoxicity (CDC) (FIG. 7). It also provides maximum hinge flexibility for the EPO to bind its receptor and eliminates the Fab-arm exchange in the blood which is an intrinsic property of human IgG4 (Rispens et al. 2011).

[19] In vivo proof of concept experiment for Efepoetin alfa was conducted in rat model of renal failure to mimic the human CKD (FIG. 8). In brief, Sprague-Dawley rats had 5 / 6 part of the kidneys surgically removed to induce renal failure condition. Eight rats per group were randomized and injected subcutaneously with a single dose of test articles including GX-E2 (Efepoetin alfa produced by GX-E2 cell line), Aranesp (darbepoetin alfa), Mircera (methoxy polyethylene glycol-epoetin beta), Epokine (epoetin alfa), and GX-E2 formulation buffer (as vehicle control). Blood was collected and blood parameters (i.e. reticulocyte, RBC, hematocrit, and hemoglobin) were measured using automatic hematology analyzer twice a week for 60 days after treatment.

[20] There was no treatment related death during the course of the experiment. There was no difference in body weight gain of the treated and untreated animals, indicating that the treatments were well tolerated.

[21] Percent reticulocytes are proportion of immature red blood cells in the whole blood; it is used as indicator of bone marrow function and the body's ability to produce RBC. On the other hand, percent hematocrit represents percentage of RBC in the whole blood and hemoglobin level indicates the oxygen-carrying capacity of the blood. Therefore, hematocrit and hemoglobin levels correspond well with the RBC count. As shown in FIG. 8, following the ESA injection, there was a rapid increase of reticulocyte level with the peak in day 5 which indicates the ESA activity on the erythroid progenitor cells. As the reticulocyte matures into RBC, the reticulocyte level decreases and RBC, hemoglobin, and hematocrit level increases. Since the half-life of RBC in circulation is around 120 days, as the effect of ESA disappears, the RBC level reaches a steady state. Compared to the untreated control group, high and medium dose of GX-E2, Aranesp, and Mircera showed statistically significant therapeutic effect based on these parameters. The high dose of GX-E2 had a more prolonged effect compared to other treatments.

[22] The invention provided herein is a robust commercial manufacturing process that can produce EPO Fc fusion protein such as Efepoetin alfa with consistent quality attributes that have been proven safe and effective for treatment of anemia in CKD patients, and with acceptable yield. The three main indicators of consistent product quality are C-terminal lysine truncation, sialic acid content, and glycosylation profiles.

[23] C-terminal lysine (CTL) truncation during bioproduction of monoclonal antibodies is a common phenomenon. This can lead to variable amounts of charge variants as lysine carries one net positive charge. The impact of C-terminal lysine truncation has been studied and it has been demonstrated that human IgG1s with zero, one and two C-terminal lysine(s) show no difference in the binding to FcRn and FcYRIIIa, thus it has no impact on the in vivo halflife and immune effector activity of human IgGs (Faid et al. 2021). Furthermore, Cai et al. shows that CTL is rapidly removed from injected antibodies in vivo by endogenous circulating carboxypeptidases in the blood stream (Cai et al. 2011). Although CTL has little, if any, impact on product safety and efficacy, the content of CTL in recombinant antibodies or Fc fusion proteins can be used as a measure of manufacturing consistency and product homogeneity.

[24] In contrast to CTL, as mentioned above, negatively charged sialic acids that are present as N-acetylneuraminic acid attached to the terminus of glycan chains have been shown to affect half-life of many glycoproteins including epoetin alfa, the recombinant EPO (Chen et al. 2022). One of the mechanisms is most likely inhibition of glycoprotein binding to mannose and asialoglycoprotein receptors that are known for clearance of glycoproteins from circulation. Our study has also shown that despite having Fc fragment, high content of sialic acid in the EPO-hybrid Fc fusion protein unexpectedly significantly contribute to the extended half-life in vivo, hence its long-acting pharmacodynamic effect (FIG. 12). Due to the importance of sialic acid in the product efficacy, sialic acid content should be kept within the range that represents clinical trial material that has been tested in the target indication.

[25] Glycosylation of protein is one characteristic of post translational modifications in mammalian cells. The glycosylation profile is mostly dependent on the production cell line and cell culture condition. Although biosynthetic production process generates several molecular variants, the level of molecular variation is tightly controlled in the upstream manufacturing process as well as the downstream process that can enrich for a certain molecular variant. Similar as CTL and sialic acid content, glycosylation profile can also be used as a measure of manufacturing consistency.

[26]

[27] Cai B., et al. (2011). "C-terminal lysine processing of human immunoglobulin G2 heavy chain in vivo". Biotechnol Bioeng., 108, 404-412.

[28] Chen B., et al. (2022). "Impact of N-linked glycosylation on therapeutic proteins". Molecules, 27, 8859.

[29] Cokic VP., et al. (2012). "JAK-STAT and AKT pathway-coupled genes in erythroid progenitor cells through ontogeny". J Transl Med., 10, 116.

[30] Faid V., et al. (2021). "C-terminal lysine clipping of IgG1: impact on binding to FcgRIIIa and neonatal Fc receptors". Eur J Pharm Sci., 159, 105730.

[31] Jelkmann W. (2016). "Erythropoietin". Front Horm Res. 47, 115-127.

[32] Jelkmann W. (2013). "Physiology and pharmacology of erythropoietin". Transfus Med Hemother. 40, 302-309.

[33] Komatsu N., et al. (1991). "Establishment and characterization of a human leukemic cell line with megakaryocytic features: dependency on granulocyte-macrophage colonystimulating factor, interleukin 3, or erythropoietin for growth and survival". Cancer Res., 51, 341-348.

[34] Lewis AM., et al. (2016). "Understanding and controlling sialylation in a CHO Fc-fusion process". PLoS ONE. 11(6), e0157111. doi:10.1371 / journal.pone.0157111.

[35] Miyazaki Y., et al. (1997). "Establishment and characterization of a new erythropoietindependent acute myeloid leukemia cell line, AS-E2". Leukemia, 11, 1941-1949.

[36] Noxon V., et al. (2017). "Tale of two erythropoiesis-stimulating agents: utilization, dosing, litigation, and costs of Darbopoetin and Epoetin among South Carolina medicaid-covered patients with cancer and chemotherapy-induced anemia". J Oncol Pract. 13, e562-e573.

[37] Pyzik M., et al. (2015). "FcRn: the architect behind the immune and non-immune functions of IgG and albumin". J Immunol. 194(10), 4595-4603.

[38] Rispens T., et al. (2011). "Mechanism of immunoglobulin G4 Fab-arm exchange". J Am Chem Soc. 133, 26, 10302-10311.

[39] Shields RL., et al. (2001). "High resolution mapping of the binding site on human IgG1 for FcyRI, FcyRII, FcyRIII, and FcRn and design of IgG1 variants with improved binding to the FcyR". J Bio Chem. 276, 6591-6604.

[40] Torres M., et al. (2018). "Process and metabolic engineering perpectives of lactate production in mammalian cell cultures". Chem Eng. 22, 184-190.

[41] Watowich SS. (2011). "The erythropoietin receptor: molecular structure and hematopietic signaling pathways". J Investig Med., 59(7), 1067-1072.

[42] Xu X. (2017). "Safety, pharmacokinetics and sialic acid production after oral administration of N-acetylmannoseamine (ManNAc) to subjects with GNE myopathy". Mol Genet Metab. 122, 126-134.

[43]

[44] [Summary of the Invention 1 [Technical Probleml

[45] An object of the present invention is to provide an EPO-Fc-producing cell line for producing EPO-Fc fusion protein having a desired CTL content, sialic acid content, isoelectric point (pI) range, and specific glycosylation profile required for the long-acting activity in vivo, and a method for producing the same.

[46] Another object of the present invention is to provide a method for producing an EPO-Fc fusion protein.

[47] Objects to be achieved by the present invention are not limited to the objects mentioned above, and other objects not mentioned above will be clearly understood by those skilled in the art to which the present invention pertains from the following description.

[48] [Technical Solution1

[49] To achieve the above objects, one embodiment of the present invention provides an EPO-Fc fusion protein producing cell line transfected with an a-2,3-sialyltransferase gene.

[50] In one embodiment of the present invention, the a-2,3-sialyltransferase may comprise an nucleotide sequence of SEQ ID NO: 1. In another embodiment of the present invention, the a-2,3-sialyltransferase may consist of the nucleotide sequence of SEQ ID NO: 1.

[51] In one embodiment of the present invention, the EPO-Fc fusion protein may bind and activate the endogenous EPO receptor at about the same level or better than the EPO BRP.

[52] In one embodiment of the present invention, the EPO-Fc fusion protein may induce the proliferation and differentiation of erythroid progenitor cells into reticulocytes at about the same level or better than EPO BRP.

[53] In one embodiment of the present invention, the EPO-Fc fusion protein may induce the proliferation of EPO-dependent cell line such as UT-7 at about the same level or better than EPO BRP.

[54] In one embodiment of the present invention, the Fc may lack the immune effector activities including, but not limited to, ADCC (antibody-dependent cellular cytotoxicity), ADCP (antibody-dependent cellular phagocytosis), and CDC (complement-dependent cytotoxicity) activities that harm the EPO receptor-expressing cells.

[55] In one embodiment of the present invention, the Fc may extend the half-life of the Fc fusion protein by binding FcRn only at pH range that is representative of the endosomal compartment in mammalian cells and dissociating from FcRn at neutral pH, thus avoiding the cellular degradation.

[56] In one embodiment of the present invention, the Fc may extend the half-life of the Fc fusion protein by increasing its molecular weight to avoid renal clearance.

[57] In one embodiment of the present invention, the EPO-Fc may comprise an amino acid sequence of SEQ ID NO: 2.

[58] In one embodiment of the present invention, the EPO-Fc fusion protein may comprise an EPO sequence and hybrid Fc sequence corresponding to the amino acid residues 1-193 and amino acid residues 194-438 of amino acid sequence SEQ ID NO: 2, respectively.

[59] In one embodiment of the present invention, the Fc may comprise 30 amino acid residues of the hinge region and 8 amino acid residues of CH2 domain N-terminus of human IgD (amino acid residues 194-231 of SEQ ID NO: 2), and 100 amino acid residues of CH2 domain C-terminus and 107 amino acid residues of CH3 domain of human IgG4 (amino acid residues 232-438 of SEQ ID NO: 2).

[60] Characterization and methods of making and using the EPO Fc fusion proteins are disclosed and described in detail in, for example "Erythropoietin immunoglobulin fusion proteins", Sehwan Yang, Young Chul Sung, Patent No: US 8,586,531 B2, filed on June 11, 2012, issued on November 19, 2013 and "Immunoglobulin fusion proteins", Sehwan Yang, Young Chul Sung, Patent No: US 7,867,491 B2, filed on May 30, 2008, issued on Jan 11, 2011 (all incorporated herein by reference in its entirety).

[61] In some embodiments of the invention, the EPO-Fc fusion expressing host cell is a Chinese Hamster's Ovary (CHO) derived cell line including DHFR-deficient CHO-DG44 cell and CHO DG44 a-2,3 ST cell line (accession number: KCLRF-BP-00525). Some embodiments of the invention include GX-E4 cell lines clone number 8P-4 and 10P-1. The GX-E4 cell line clone number 8P-4 is stored at Biovian Oy, Genexine and Kalbio Global Medika. The GX-E4 cell line clone number 10P-1 is stored at Genexine.

[62] To achieve the above objects, another embodiment of the present invention provides a method for producing an EPO-Fc-producing cell line, comprising steps of: introducing an EPO-Fc gene into host cells, and then selecting EPO-Fc-producing host cells expressing EPO-Fc fusion protein; re-transfecting the selected EPO-Fc-producing host cells with a hygromycin-resistant vector into which an a-2,3-sialyltransferase gene has been introduced; and selecting hygromycin-resistant host cells as an EPO-Fc-producing cell line among the re-transfected EPO-Fc-producing host cells, wherein the EPO-Fc-producing cell line overexpresses a-2,3-sialyltransferase.

[63] To achieve the above objects, still another embodiment of the present invention provides a method for producing an EPO-Fc fusion protein having following quality attributes: CTL content of about 1% to about 10%, sialic acid content of about 17 mol / mol to about 27 mol / mol and pI of about 4.0 to about 6.0. In another embodiment, the method producing Fc fusion protein thereof comprises of fed-batch culturing a host cell transfected with DNA encoding the Fc fusion protein and DNA encoding human a2,3 sialyltransferase, isolating and purifying an Fc fusion protein produced by the same cell line.

[64] In one embodiment of the present invention, a method for producing an EPO-Fc fusion protein, may comprise: (a) fed-batch culturing the EPO-Fc fusion protein producing cell line; and (b) isolating and purifying an EPO-Fc fusion protein produced in the cell line.

[65] In one embodiment of the present invention, the EPO of the EPO-Fc fusion protein may further comprise modification for additional glycosylation.

[66] In one embodiment of the present invention, the EPO-Fc fusion protein may further comprise PEG.

[67] In one embodiment of the present invention, the EPO-Fc fusion protein may have specific quality attributes including at least one of the following: C-terminal lysine content, N- and O-glycan profiles, sialic acid content, and pI range.

[68] In some embodiments of the invention, the CTL content of the EPO-Fc fusion protein may be about 1% to about 10%, particularly, about 2% to about 9% e.g. about 3% to about 7.0%, or about 5.0%. In some embodiments, the sialic acid content of the EPO-Fc fusion protein may be about 17 mol / mol to about 27 mol / mol, particularly, about 20 mol / mol to about 25 mol / mol. In some embodiments, the pI of the EPO-Fc fusion protein may be about 4.0 to about 6.0, particularly, about 4.5 to about 5.3.

[69] In one embodiment of the present invention, the N- and O-glycan profiles may include the non-glycosylated content on position N24 about 0-20%, on position N38 about 0-5%, on position N83 about 10-17%, on position N261 about 0-10%, and on position S126 about 5070%.

[70] In one embodiment of the present invention, the N- and O-glycan profiles may include the glycan containing sialic acid on position N24 about 50-90%, on position N38 about 55-95%, on position N83 about 40-80%, on position N261 about 0-30%, and on position S126 about 15-40%.

[71] In an embodiment of the present invention, the fed-batch culturing may be performed in the presence of N-acetyl-D-mannosamine (ManNAc) and galactose.

[72] In an embodiment of the present invention, the N-acetyl-D-mannosamine may be present at a concentration of about 1 mM to about 20 mM, and the galactose may be present at a concentration of about 1 mM to about 40 mM.

[73] In an embodiment of the present invention, the fed-batch culturing is performed under conditions of about 34°C to about 38°C and about pH 6.7 to about pH 7.5.

[74] In an embodiment of the present invention, the fed-batch culture may comprise culturing the cell line for 9 to 11 days.

[75] In an embodiment of the present invention, the fed-batch culture may be cultured at a temperature lowered by 2 to 5°C on day 3 or 4 days after inoculation of the cell line

[76] In an embodiment of the present invention, the fed-batch culturing may be performed at a temperature changed to about 29°C to about 36°C on day 3-4 of the culturing.

[77] In an embodiment of the present invention, the fed-batch culturing may be performed until the viability of the cells in culture medium reaches 70% or less.

[78] In an embodiment of the present invention, the EPO-Fc fusion protein isolation and purification may be performed at room temperature.

[79] In an embodiment of the present invention, the EPO-Fc fusion protein isolation and purification, except for viral inactivation, may be performed at room temperature.

[80] In an embodiment of the present invention, the EPO-Fc fusion protein isolation and purification may be performed at 17°C to 27°C.

[81] In an embodiment of the present invention, the isolating and purifying may be performed using affinity chromatography or ion exchange chromatography. In another embodiment ion exchange chromatography is an anion exchange chromatography.

[82] In an embodiment of the present invention, the ion exchange chromatography may include an anion exchange chromatography process using equilibration buffers with pH of about 6.00 to about 6.50 to facilitate the binding of acidic species of EPO-Fc fusion protein to the resin. In another of the present invention, the ion exchange chromatography may include an anion exchange chromatography process using equilibration buffers with pH of about 6.10 to about 6.30 to facilitate the binding of acidic species of EPO-Fc fusion protein to the resin.

[83] In an embodiment of the present invention, the ion exchange chromatography may include an anion exchange chromatography process using equilibration buffer with conductivity of about 6.00 to about 10.50 mS / cm, and elution buffer with conductivity of about 11.50 to about 16.00 mS / cm to facilitate the isolation of EPO-Fc fusion protein with the quality attributes thereof. In another embodiment of the present invention, the ion exchange chromatography includes an anion exchange chromatography process using equilibration buffer with conductivity of about 8.00 to about 10.00 mS / cm, and elution buffer with conductivity of about 12.00 to about 15.00 mS / cm to facilitate the isolation of EPO-Fc fusion protein with the quality attributes thereof.

[84] In another embodiment the anion exchange chromatography step uses, an equilibration buffer with pH of about 6.00 to about 6.50, and conductivity of about 6.00 to about 10.50 mS / cm, and an elution buffer with pH of about 6.00 to about 6.50 and conductivity of about 11.50 to about 16.00 mS / cm.

[85] In some embodiments of the invention, the equilibration and elution buffers' pH in the anion exchange step is about 6.10 to about 6.30, the equilibration buffer's conductivity in the anion exchange step is about 8.00 to about 10.00 mS / cm, and the elution buffer's conductivity in the anion exchange step is about 12.00 to 15.0 mS / cm.

[86] In some embodiments of the invention, the Fc fusion protein is an EPO-Fc fusion protein, wherein the Fc fragment (1) lacks the immune effector activities including, but not limited to, ADCC (antibody-dependent celllular cytotoxicity), ADCP (antibody-dependent cellular phagocytosis), and CDC (complement-dependent cytotoxicity) activities that harm the EPO receptor expressing cells, (2) and extends the half-life of the Fc fusion molecule by (2a) increasing the molecular weight of the Fc fusion protein to avoid renal clearance and (2b) binding FcRn only at pH range that is representative of endosomal compartment in mammalian cells and dissociating from FcRn at neutral pH to avoid cellular degradation.

[87] The Fc fragment thereof comprises of immunoglobulin constant region and hinge region which are derived from IgG1, IgG2, IgG3, IgG4, IgD, IgA, IgE, IgM, or their combination (e.g. IgG4 and IgD) sequences and modified sequence of the immunoglobulin subtype thereof.

[88] In some embodiments of the invention, the EPO-Fc fusion protein can (1) bind and activate the endogenous EPOR at least at the same level as EPO BRP, (2) and induce the proliferation and differentiation of erythroid progenitor cells into reticulocytes that matures into red blood cells. The EPO fragment thereof can be derived from native EPO or modified EPO sequence and / or EPO-related structure including that with additional glycan structure or additional PEG.

[89] The method of this invention, as described above, may be used in a process for manufacturing Efepoetin alfa, which is an EPO-hybrid Fc fusion protein, in a commercial manufacturing process, and may in particular contribute to increasing the efficacy of the EPO-hybrid Fc fusion protein. Efepoetin alfa is a protein used, for example, for the treatment of anemia in humans.

[90] The method disclosed in this specification may be used to improve the sialic acid content associated with a longer in vivo half-life and batch homogeneity in, for example, a commercial manufacturing process, for example, a manufacturing process used to produce efepoetin alfa, which is an EPO hybrid Fc fusion protein used for the treatment of anemia in humans.

[91] [Advantageous Effects!

[92] The present invention relates to a method for producing a long-acting EPO fusion protein. According to the method of the present invention, it is possible to efficiently produce a highly sialylated form of a long-acting erythropoietin fusion protein with an increased in vivo halflife through optimization of the cell line, culture method, culture conditions, purification conditions, etc.

[93] The effects of the present invention are not limited to the above-mentioned effects, and it is to be understood that the effects of the present invention include all effects that may be deduced from the configuration of the present invention as described in the description of the invention or the appended claims.

[94] [ BRIEF DESCRIPTION OF THE FIGURES!

[95] FIG. 1 shows the structure of Efepoetin alfa protein. Efepoetin alfa is a homodimer. Each subunit contains erythropoietin (EPO), IgD's hinge and partial CH2 domain, and IgG4's partial CH2 and CH3 domains.

[96] FIG. 2 shows the full-length amino acid sequence of Efepoetin alfa (SEQ ID NO:2). Black: EPO sequence. Underlined black: signal peptide. Bold: partial human IgD heavy chain sequence (30 aa of the hinge region + 8 aa of CH2 domain N-terminus). Italic: partial human IgG4 heavy chain sequence (100 aa of CH2 domain C-terminus + 107 of CH3 domain).

[97] FIG. 3a to FIG 3d show binding to EPO receptor (EPOR) measured by surface plasmon resonance (SPR) assay. BSA and Rituxan (anti-CD20 antibody) are negative control. EPO-BRP is EPO reference standard, a positive control. GX-E2 is Efepoetin alfa produced by GX-E2 cell line.

[98] FIG. 4 shows the ability of GX-E2 (Efepoetin alfa) to induce proliferation of EPO dependent UT-7 cell line in dose dependent manner. EPO-BRP is used as positive control.

[99] FIG. 5 shows binding of Efepoetin alfa to human FcRn measured by SPR assay.

[100] FIG. 6 shows FcyRI ELISA result. Rituxan and Enbrel (human TNFR2 and human IgG1 Fc fusion protein) are positive controls. GX-E2 is Efepoetin alfa produced by GX-E2 cell line.

[101] FIG. 7 shows C1q ELISA result. Rituxan and Enbrel are positive controls.

[102] FIG. 8 shows erythropoiesis stimulating activity of GX-E2 (Efepoetin alfa) in rat model of chronic kidney disease (CKD). RBC, red blood cell. HGB, hemoglobin. HCT, hematocrit. RET, reticulocyte. Rats with surgically induced CKD were injected with a single dose of test article at the indicated dose. The blood parameters were measured twice a week for 60 days.

[103] FIG. 9 shows the representative IEF gel electrophoresis profile of Efepoetin alfa. RS, reference standard. S1-S3, Efepoetin alfa samples.

[104] FIG. 10 shows the representative peptide mapping result of Efepoetin alfa. Red chromatogram, formulation buffer. Green chromatogram, Efepoetin alfa reference material.

[105] FIG. 11 shows the representative in vivo bioassay result.

[106] FIG. 12 shows the ability of GX-E2 (Efepoetin alfa) with the indicated sialic acid content (number in parenthesis) to induce change in hemoglobin level in normal mice. Sialic acid content unexpectedly affects the efficacy of Efepoetin alfa.

[107] FIG. 13 shows the amino acid sequence of a-2,3 sialyltransferase with HA tag (SEQ ID NO. 3).

[108] FIG. 14 shows the detail construct of pCI-hygro-a 2,3 st-IRES-pERP plasmid that express a2,3 sialyltransferase.

[109] FIG. 15 shows schematic diagram of GX-E4 cell line generation.

[110] FIG. 16 shows Western blot analysis of the a-2,3 sialyltransferase tranfected cell pools using anti HA antibody. C is control. Lanes 6-10 correspond to the indicated a-2,3 sialyltransferase transfected cell pools.

[111] FIG. 17 shows the IEF gel electrophoresis from a-2,3 sialyltransferase transfected GX-E2 cell pools. Lane 2 is the purified GX-E2 (Efepoetin alfa). Lanes 3-7 are cell culture from the indicated a-2,3 sialyltransferase transfected GX-E2 cell pools. Lane 8 is GX-E2 cell culture.

[112] FIG. 18 shows the IEF gel electrophoresis of GX-E4 single subclones. GX-E4 cell line is GX-E2 cell line that has been transfected with a-2,3 sialyltransferase.

[113] FIG. 19 shows growth curve of cells cultured in the presence of the indicated feed.

[114] FIG. 20 shows IEF gel electrophoresis result of cells cultured in the presence of the indicated feed.

[115] FIG. 21 shows IEF gel electrophoresis result of the DOE (design of experiment) study for ManNAc and galactose concentrations.

[116] FIG. 22 shows the feed condition - prediction profiler plot.

[117] FIG. 23 shows schematic of upstream manufacturing process of Efepoetin alfa. IPC, in process control (critical control parameter that needs to be fulfilled according to the acceptance criteria before the process can continue). IPM, in process monitoring (control parameter for information purpose). VCD, viable cell density.

[118] FIG. 24 shows IEF gel electrophoresis of cell culture from the fed-batch bioreactor at the indicated culture period. Compared to the purified GX-E2 reference standard, cell culture taken at day 10 contain more acidic species compared to those taken at day 14, 15, or 16.

[119] FIG. 25 shows schematic of downstream manufacturing process of Efepoetin alfa.

[120] FIG. 26 shows prediction profiler plots obtained from DOE study of the production bioreactor step.

[121] FIG. 27 shows viable cell density and lactate concentration curves of 6 batches of GX-E4 during the 10-day production bioreactor culture. Batch 1-3: pH 7.0 and temperature shift on day 3. Batch 4-6: pH 6.80 and temperature shift on day 4.

[122] FIG. 28 shows the N-glycan profile at N24 position.

[123] FIG. 29 shows the N-glycan profile at N38 position.

[124] FIG. 30 shows the N-glycan profile at N83 position.

[125] FIG. 31 shows the N-glycan profile at N261 position.

[126] FIG. 32 shows the O-glycan profile at S126 position.

[127] [ DETAILED DESCRIPTION OF THE INVENTION]

[128] The followings describe the example embodiments of the invention and the definition of terms used herein.

[129] The present invention relates to an EPO-Fc-producing cell line.

[130] The EPO-Fc-producing cell line of the present invention is a cell line transfected with an a-2,3-sialyltransferase gene.

[131] The EPO-Fc-producing cell line may have an enhanced function due to additional introduction of a-2,3-sialyltransferase thereinto and overexpression of a-2,3-sialyltransferase therein.

[132] The a-2,3-sialyltransferase gene may consist of, for example, SEQ ID NO: 1, without being limited thereto.

[133] The transfection may be performed using the method, conditions, and means known in the art. For example, the cell line from the existing EPO-Fc Master Cell Bank (MCB) may be re-transfected with a-2,3-sialyltransferase by electroporation.

[134] The present invention also relates to a method for producing an EPO-Fc-producing cell line.

[135] The method of the present invention comprises steps of: introducing an EPO-Fc gene into host cells, and then selecting EPO-Fc-producing host cells expressing EPO-Fc fusion protein; re-transfecting the selected EPO-Fc-producing host cells with a hygromycin-resistant vector into which an a-2,3-sialyltransferase gene has been introduced; and selecting hygromycin-resistant host cells as an EPO-Fc-producing cell line among the re-transfected EPO-Fc-producing host cells.

[136] Here, the EPO-Fc, the EPO-Fc-producing cell line, the a-2,3-sialyltransferase gene, and re-transfection are as described above.

[137] The host cells may be, for example, CHO cells, more specifically, CHO dhfr(-) cells or CHO DG44 cells.

[138] The EPO-Fc-producing cell line produced by the method of the present invention has an enhanced function by overexpressing a-2,3-sialyltransferase, and is able to produce a highly sialylated EPO-Fc fusion protein.

[139] The EPO-Fc-producing cell line is a cell line used for expression of the EPO-Fc fusion protein. For example, the EPO-Fc-producing cell line is one obtained by re-transfecting a CHO host cell, transfected to express EPO-Fc, with the a-2,3-sialyltransferase gene. As a more specific example, the EPO-Fc-producing cell line may be a CHO DG44 a-2,3 ST cell line (accession number: KCLRF-BP-00525).

[140] The present invention provides methods for producing an Fc fusion protein including EPO-Fc fusion protein that has a C-terminal lysine (CTL) content of about 1% to about 10%, sialic acid content of about 17 mol / mol to about 27 mol / mol and pI of about 4.0 to about 6.0, comprising the upstream process that includes fed-batch culturing the host cell transfected with DNA. encoding the Fc fusion protein and DNA. encoding human a-2,3-sialyltransferase in a culture medium containing N-acetyl-D-mannosamine and galactose, and the downstream process that includes controlling the anion exchange chromatography step using the equilibration buffer with pH of about 6.00 to about 6.50 and with conductivity of about 6.00 to about 10.50 mS / cm, and the elution buffer with pH of about 6.00 to about 6.50 and with conductivity of about 11.50 to about 16.00 mS / cm.

[141] In another embodiment, the medium may contain N-acetyl-D-mannosamine at a concentration of is about 1 mM to about 20 mM, preferably about 5 mM to about 15 mM, more preferably about 10 mM, and galactose at a concentration of about 1 mM to about 40 mM, preferably about 20 mM to about 35 mM, more preferably about 32 mM.

[142] Additionally, the medium may further contain glutamine, methotrexate, and hygromycin.

[143] For example, the fed-batch culturing may be performed at a temperature of 34°C to 38°C at the initiation of the culture. Further, the fed-batch culturing may be performed at pH 6.7 to pH 7.5 at the initiation of the culture.

[144] In addition, the fed-batch culturing may be performed at a temperature changed to 29°C to 36°C lowering the temperature by 2-5°C on 3-4 days after cell inoculation.

[145] By changing the temperature as described above, that is, maintaining the temperature at 34 to 38°C in the cell growth phase and lowering the temperature to 29 to 36°C by 2-5°C in the protein production phase, it is possible to increase protein production. In addition, as the temperature is lowered, the enzymatic activity of sialidase in the CHO cells is lowered, so that the sialic acid of the protein is not degraded and may be maintained stably.

[146] The culturing may be performed, for example, until the viability of the cells in the culture medium reaches 80% or less or 70% or less, without being limited thereto. More specifically, the culturing may be performed for the shortest possible time until the viability of the cells in the culture medium reaches 70% or less.

[147] The term "about" used herein refers to a range of ±20%, ±15%, ±10%, ±5%, or ±1% of the stated value

[148] The term "Fc fusion protein" or "EPO-Fc fusion protein" used herein refers to a protein resulted from a fusion or joining between Fc fragment and another protein including a biologically active molecules such as EPO. In some embodiments of this invention, the Fc fragment of the EPO-Fc fusion protein is a hybrid of human IgD sequence and human IgG4 sequence as explained in FIG. 1, FIG. 2, and SEQ ID NO:2. The Fc fragment thereof is also referred as hybrid Fc fragment.

[149] The term "Fc fragment" or "Fc", as used herein, refers to a protein fragment that contains the heavy chain constant region 1 (CH1), the heavy chain constant region 2 (CH2), and the heavy chain constant region 3 (CH3) of an immunoglobulin molecule, and not the variable regions of the heavy and light chains, and the light chain constant region 1 (CL1) of the immunoglobulin. It may further include the hinge region at the heavy chain constant region.

[150] In some embodiments, the hybrid Fc fragment of the invention binds neonatal Fc receptor (FcRn) but lacks binding ability to Fc gamma receptors (FcyR) and C1q (complement component 1q).

[151] FcRn that preferentially binds IgG under acidic condition was first discovered for its role in maternal IgG transfer to the offspring through placenta. Later, it was discovered that FcRn protects IgG from cellular degradation. Proteins including IgGs are continously internalized by vascular endothelial cells and immune cells via non-specific pinocytosis. Once inside the cells, IgGs bind FcRn as the pH in the endosomes becomes acidic. FcRn bound IgGs are then recycled back to the cell surface where they dissociate from FcRn at higher pH. Whereas proteins that do not bind FcRn would be transported to lysosome to be degraded. This mechanism prolongs the half-life of IgG (reviewed in Pyzik et al., 2015).

[152] Human IgG, particularly IgG1, can mediate strong immune effector activities including ADCC, ADCP, and CDC that can harm or kill the cells where it binds. The ADCC is mediated by the interaction of FcYRIIIa, which is mainly expressed on NK cells, with the Fc fragment of the IgG. The ADCP is mediated by the FcYRIIa which is mainly expressed on macrophages. While CDC is mediated by the interaction of Fc fragment with C1q that triggers classical complement pathway. In some embodiments, assessment of the immune effector activity of EPO-Fc fusion protein is performed using ELISA with FcyRI and C1q. In this assay, FcyRI was selected to replace FcYRIIIa and FcYRIIa since FcyRI binds similar epitope of the Fc fragment (Shields et al., 2001) much more strongly compared to the FcyRIIIa and FcyRIIa. Other cell-based assays known by one skilled in the art using primary human NK cells or NK cell line, human macrophages or macrophage cell line, and appropriate target cell lines can be used to assess the immune effector activities of Fc fusion protein such as EPO-Fc fusion protein.

[153] In other embodiments, Fc fragment of the EPO-Fc fusion protein lacks Fab-arm exchange. Human IgG4 is an unusually dynamic antibody having the capacity to undergo half-molecule exchange or Fab-arm exchange resulting in asymmetrical, bispecific antibodies with two different antigen binding sites (Rispens et al., 2011). In some embodiments, combining the hinge region of IgD and the 8 amino acid residues of CH2 domain of human IgD with partial CH2 and CH3 domains of IgG4 eliminates the Fab arm exchange ability of the human IgG4 Fc fragment.

[154] The term "Efepoetin alfa" refers to EPO-Fc fusion protein which structure and amino acid sequence are shown in FIG. 1 and FIG. 2, respectively. In some embodiments, GX-E2 protein is Efepoetin alfa that is produced by GX-E2 cell line. In other embodiments, GX-E4 protein is Efepoetin alfa that is produced by GX-E4 cell line. Yet in another embodiment, GX-E4 cell line is a GX-E2 cell line that has been co-transfected with DNA encoding human a-2,3-sialyltransferase.

[155] CTL content is expressed herein as the percentage of C-terminal peptide possessing a CTL and can be calculated using the following equation: CTL content (%) = (number of C-terminal peptide possessing a CTL) / (total number of C-terminal peptide) x 100. The number of C-terminal peptide corresponds to the number of the Fc fusion subunit since each C-terminal peptide is derived from a distinct Fc fusion subunit. In some embodiments of the methods described herein, CTL content of the Fc fusion protein or EPO-Fc fusion protein is about 1% to about 10%, specifically, about 2% to about 9%, more specifically, about 3% to about 7%.

[156] Sialic acid content is expressed herein as the mole amount of sialic acid per mole amount of Fc fusion protein, wherein the sialic acid mole amount is determined using resorcinol and UV spectrometer as described in the Example 2 below. Other methods that measure sialic acid mole amount such as High-performance liquid chromatography (HPLC) can be used in lieu of the resorcinol method. In some embodiments of the methods described herein, sialic acid content of the Fc fusion protein or EPO-Fc fusion protein is about 17 mol / mol to about 27 mol / mol, more specifically, about 20 mol / mol to about 25 mol / mol.

[157] Isoelectric point (pI) range used herein refers to the pI range of the purified Fc fusion protein or EPO-Fc fusion protein that is determined by IEF gel electrophoresis as explained in the Example 1 below. Other methods that separate glycoprotein by electrical charge such as ion exchange chromatography can be used in lieu of IEF gel electrophoresis. In some embodiments of the methods described herein, the pI range of the Fc fusion protein or EPO-Fc fusion protein is about 4.0 to about 6.0, more specifically, about 4.5 to about 5.3. In other embodiment, the pI of a glycoprotein represents the summation of all electrical charges that are associated with CTL, sialic acid, glycans, as well as amino acid residues and post translational modifications including deamidation.

[158] The term "upstream process" used herein refers to initial phases of biomanufacturing process that include all steps related to growing the host cells, producing the desired protein by the host cells in a cell culture system, harvesting or collecting the desired protein from the cell culture and removing the cells and cell debris from the cell culture which is also known as clarification step.

[159] The term "downstream process" used herein refers to later phases of biomanufacturing process that include all steps related to purifying the desired protein, removing the product-related and process-related impurities, inactivating and removing the potential adventitious agents, and formulating the desired protein which is also known as (a.k.a.) active pharmaceutical ingredient (API).

[160] Related to the biomanufacturing process, the term "drug substance" or "DS" is used herein to describe API, which is part of "drug product" or "DP" that refers to the finished dosage form.

[161] The term "product-related impurities" refers to impurities deriving from the product / active pharmaceutical ingredient (e.g. degradation and aggregation products, structural heterogeneity). This type of impurities is common for biological products due to its inherent heterogeneity and physicochemical property. These impurities may have reduced efficacy, safety, and / or PK, hence they need to be removed or reduced.

[162] The term "process-related impurities" refers to impurities deriving from the manufacturing process (e.g. protein A that leaches from the resin of affinity chromatography column, host cell DNA and host cell protein). These impurities are inherent for biologic product manufacturing process. Since these impurities are likely to cause unwanted serious side effects in patients, sufficient reduction during purification process needs to be ensured.

[163] Methods of producing Fc fusion protein having at least one of the defined quality attributes thereof include using any host cells known in the art for producing glycoprotein, including mammalian cells, such as Chinese Hamster Ovary (CHO) cell lines, e.g. DHFR deficient CHO-DG44 cell, CHO-K1SV cell (Lonza Biologics, Cambridge, MA, USA), and CHO-K1 cell (American Tissue Culture Collection (ATCC), Manassas, VA, USA). Other exemplary host cells are SP2 / 0-Ag14 cell (ATCC) and NS0 (European Collection of Authenticated Cell Cultures (ECACC, Salisbury, Wilshire, UK). Cells can be cultured in a single-cell suspension or in an anchorage-dependent fashion using culture medium and feeds that are appropriate for the host cells. The choice of a suitable culture medium is within the knowledge of one skilled in the art. In some embodiments, the culture medium is a serum-free medium or animal component-free medium.

[164] In some embodiments on the invention, the host cell is CHO-DG44 cell line with DHFR (dihydrofolate reductase) gene mutation. This cell line was developed in Dr. Lawrence Chasin laboratory (Columbia University, NY, USA). Due to its defects in DHFR gene, the CHO-DG44 cell line requires hypoxanthine and thymidine for growth. The strategy for clone selection using DHFR gene is a common knowledge of one skilled in the art.

[165] In some embodiments, gene encoding the protein of interest is amplified using an increasing concentration of methotraxate (MTX) which inhibits DHFR. This strategy to increase the productivity of the transfected cell line is also a common knowledge of one skilled in the art.

[166] A variety of culture system are known in the art, including shake flasks, roller bottles, and bioreactors, e.g. stirred-tank bioreactors. Culture medium can be added in a single batch process or in a fed-batch process, in which small batches of culture medium are added periodically. The host cells can also be cultured in a perfusion culture, which involves continuous removal of a certain volume of culture medium from the culture and replacing the removed volume with a comparable volume of fresh medium. Typically, perfused cultures can be operated to achieve higher cell densities than batch cultures and can be maintained for longer periods of time. Perfused cultures also enable repeated harvest. However, the present invention also provides evidence that there is an inverse correlation between total amount of protein produced in the cell culture and the sialylation level of the protein product. Therefore, the major goal of the upstream process is not high protein yield but high level of acidic protein which represents proteins with high sialic acid content.

[167] The present invention further provides methods to increase sialylation level of the biomolecule that comprises the use of host cells transfected with DNA encoding sialyltransferase enzymes including human a-2,3-sialyltransferase and the other 19 sialyltransferases in the same class that are involved in the biosynthesis of sialoglycan or sialic acid-containing glycans. The methods thereof further include culturing the said host cells in the presence of N-acetyl-D-mannosamine, the precursor of sialic acid.

[168] The term "N-acetyl-D-mannosamine" refers to uncharged monosaccharide and the first committed precursor in the sialic acid biosynthetic pathway (Xu et al., 2017). In some embodiments of this invention, N-acetyl-D-mannosamine is added to the cell culture to boost the sialic acid biosynthesis.

[169] Methods of producing Fc fusion protein having at least one of the defined quality attributes thereof include the control of the cell culture's pH between about 6.70 to about 7.50, more specifically between about 6.80 to about 7.00. The pH in the culture can be controlled using several approaches that are known in the art. These include the use of buffering agents in the culture medium, CO2 sparging into the culture medium to reduce the pH, base addition to the culture to increase the pH, acid addition to the culture to reduce the pH, removal of the accumulated CO2 by sparging N2 and increasing agitation and headspace aeration, and the combination approaches thereof. The base solutions that are commonly used to increase culture pH include sodium bicarbonate, sodium hydroxide, potassium hydroxide, and ammonium hydroxide. The acid solutions that are used to reduce the culture pH include phosphoric acid and sulfuric acid.

[170] Methods of producing Fc fusion protein having at least one of the defined quality attributes thereof can further comprise controlling the harvest time of the Fc fusion protein. In some embodiments, the Fc fusion protein can be harvested within 14 days after cell inoculation, specifically within 13 days after inoculation, preferably within 10 days after inoculation. In one embodiment of the present invention, controlling the harvest time maximizes the production of the glycoprotein's acidic species that represent proteins with high sialic acid content as shown in the Example 9 below.

[171] The term "harvest" used herein means collect, particularly, the desired protein from the cell culture. The term "controlling the harvest time" refers to adjusting the time period, in which the host cells produce the desired protein, for collection of the desired protein to be purified in subsequent stages.

[172] Methods of producing Fc fusion protein having at least one of the defined quality attributes thereof includes the downstream process control. The downstream process aims mainly at removing components other than the EPO-Fc fusion protein contained in the culture medium, and may be performed using the method, conditions, and means known in the art. For example, host cell-derived impurities may be removed by subjecting the culture medium to several stages of filtration and purification. Host cell-derived impurities may include protein impurities, including abnormal peptides such as various aggregates and fragments, DNA impurities, endogenous viruses, exogenous viruses, and other particles, which may be derived from host cells and supplementary materials.

[173] In addition, in the purification process, the EPO-Fc fusion protein may be purified by, for example, protein A purification, hydroxyapatite (HAP) purification, etc.

[174] In the present invention, the isolation and purification are performed at room temperature except for the low-pH virus inactivation step described below.

[175] Since EPO is a hormone and is unstable at room temperature, purification is generally performed at cold temperature (5 ± 3°C). However, in the method of the present invention, purification may be performed at room temperature (22 ± 5°C) as the stability of the EPO-Fc fusion protein is high.

[176] In addition, when purification is performed at cold temperature, problems arise such as high costs and worker safety issues, but in the present invention, these problems may be overcome because isolation and purification is performed at room temperature.

[177] The purification may further include a low-pH virus inactivation step after protein A purification, wherein the low-pH virus inactivation step may be performed under conditions of 10°C to 15°C and pH 3.5 to pH 3.8.

[178] The isolation and purification may be performed using, for example, affinity chromatography or ion exchange chromatography. The ion exchange chromatography may include anion exchange chromatography. The anion exchange chromatography process described herein depends on the reversible adsorption of charged solute molecules (in this case, negatively charged) to immobilized groups of opposite charge. The process is divided into 4 stages: equilibration, sample loading and wash, elution, and regeneration. The process thereof is critically dependent on the pH and conductivity (or ionic strength) of the buffers used in the process. The buffer pH determines the electrical charge status of the proteins; thus, it determines the binding property of the molecules to the immobilized resin. For example, at pH 6.0, the glycoprotein having pI of 4.0 would have negative charge while the glycoprotein having pI of 8.0 would have positive charge. At this condition, only the glycoprotein having pI 4.0 binds to the positively charged anion exchange resin while the glycoprotein with pI 8.0 would pass through the column. The buffer conductivity determines the exchange capacity during the binding and elution process step e.g. low conductivity allows binding of the target molecule while high conductivity would elute the glycoprotein from the column. In some embodiments, the pH of equilibration, wash, and elution buffers is about 6.00 to about 6.50, more specifically, about 6.10 to about 6.30. In other embodiments, the equilibration / wash buffer conductivity is about 6.00 to about 10.50 mS / cm, more specifically, about 8.00 to about 10.0 mS / cm. In another embodiments, the elution buffer conductivity is about 11.50 to about 16.00 mS / cm, more specifically, about 12.00 to about 15.00 mS / cm.

[179] The term "quality attribute", as used herein, means a physical or chemical characteristics that are related to the product quality. Since pI range determined by IEF gel electrophoresis and sialic acid content confirmed by the resorcinol assay are directly correlated with the PK, efficacy, and safety of the product, they are categorized as "critical quality attributes (CQA)" which are physical or chemical characteristics that should be within an appropriate limit, range, or distribution to ensure the desired product quality. Whereas CTL is not known to impact the PK, immunogenicity, efficacy, and safety of the product, thus it is not part of CQA. Nevertheless, CTL can be used as a measure of manufacturing consistency and product homogeneity.

[180] The term "controlling" or "control", as used herein, means monitor and adjust the process parameter or condition to give a desired output. Controlling includes measuring a substance or a parameter being controlled directly or indirectly and using those measurements to make corrections toward the desired result.

[181] In some embodiments, in order to reduce the risk of adventitious agent contamination, raw materials used in the manufacturing of Fc fusion protein are preferably animal component-free and of GMP grade, that are qualified for biologic manufacturing.

[182] The term "erythropoietin" or "EPO" refers to a glycoprotein hormone naturally produced by the peritubular cells of the kidney that stimulates red blood cell production, also known as erythropoiesis. Erythropoiesis starts with the binding of EPO to the EPO receptor (EPOR) on the erythroid progenitor cells which triggers receptor dimerization and / or reorientation of the receptor followed by JAK / STAT and PI3K / AKT signaling cascades (Watowich, 2011 and Cokic et al., 2012). JAK / STAT pathway refers to Janus kinase and signal transducer and activator of transcription pathway. And PI3K / AKT pathway refers to phosphatidylinositol 3-kinase and Ak strain transforming pathway. Both JAK / STAT and PI3K / AKT pathways are well characterized membrane-to-nucleus signaling modules that induce the expression of various genes which, in the case of EPO and erythroid progenitor cells, promote erythroid progenitor cell survival, proliferation, and differentiation into reticulocytes. Reticulocytes are immature red blood cells that mature into red blood cells within 1 or 2 days after being released into peripheral blood. EPOR is highly expressed on erythroid progenitor cells, but it is also expressed at lower level in multiple tissues such as brain, heart, and bone. Thus, in addition to erythropoiesis, EPO also causes other biological effects on multiple tissues.

[183] The term "chronic kidney disease" or "CKD" refers to a type of kidney (a.k.a. renal) disease in which a gradual loss of kidney function occurs over a period of months to years. Since the main function of the kidney is to remove waste from the blood and return the cleaned blood back to the body through a so called "renal clearance" or "renal / kidney filtration" mechanism, CKD patients eventually needs dialysis to remove the waste from their blood when their kidneys stop working. The renal clearance system has a molecular weight cutoff of around 45 kDa making the molecules with molecular weight larger than the cutoff has a longer half-life in the peripheral blood and vice versa. Thus, one approach to increase the half-life of a therapeutic molecule is to simply increase the molecular weight through fusion with another molecule (e.g. Fc fragment) as EPO-Fc fusion protein referred herein. Although in the case of Fc fusion protein, the Fc fragment can also bind to FcRn to avoid cellular degradation that further increase the lifetime in the circulation.

[184] In addition to the waste removal, another function of kidney cells particularly renal cortex peritubular cells is to produce EPO (Jelkmann, 2016). In fact, kidney produces most EPO in the human body. Due to reduced EPO level, patients with renal failure due to CKD or other causes such as treatment of unrelated disease using toxic small molecule drugs e.g. chemotherapy, often suffer from anemia.

[185] The term "rat model of renal failure" used herein refers to the animal disease model of renal failure using rat. Animal disease model is developed to investigate the biology of the disease and to evaluate the effects of a potential therapeutic agent. Rat model of renal failure can be induced by removal of the kidney through surgery or chemotherapy.

[186] The term "erythropoiesis stimulating agent" or "ESA", as used herein, refers to recombinant versions of EPO produced pharmacologically via recombinant DNA technology and bioprocess. ESA is used to treat anemia related to renal failure. Examples of ESAs include epoetin alfa with the brand name of Epokine, darbepoetin alfa with the brand name of Aranesp or NESP, and methoxy polyethylene glycol-epoetin beta with the brand name of Mircera. Aranesp and Mircera are EPO with extended half-life; they carry additional carbohydrate side chains or additional pegylation, respectively.

[187] The term "EPO BRP" refers to European Pharmacopoeia EPO biological reference preparation which is the EPO reference material or reference standard. EPO BRP is used as positive controls in several experiments presented in this invention. The term "reference standard" or "reference material" refers to a highly purified and well-characterized material suitable to test the identity, strength, quality, and purity of substances for pharmaceutical products.

[188] The term "in vivo bioassay" as used herein refers to an assay that measures the potency of API using animals. In some embodiments, the experiment is done using normocythemic mice which refers to mice that have normal number of red blood cells. The potency of test articles are compared directly with that of reference material or reference standard.

[189] In some embodiments of the present invention, the capacity of EPO-Fc fusion protein to stimulates erythroid progenitor cells can be assessed by a simple binding assay such as surface plasmon resonance (SPR) technique using recombinant EPOR. In other embodiment, the same assessment can be done with cell proliferation assay using cell lines that are dependent on EPO such as UT-7 cell line (Komatsu et al., 1991) and AS-E2 cell line (Miyazaki et al, 1997; Riken BioResource Research Center, Tsukuba, Ibaraki, Japan).

[190] The term "SPR" refers to an optical technique for detecting the interaction of two different molecules in which one is mobile and the other one is fixed on a thin gold film. The interaction between two molecules produces a small change in refractive index at the gold surface, which can be quantified using instrument such as Biacore 3000 (Biacore, Uppsala, Sweden).

[191] Glycosylation of protein is one characteristic of post translational modifications in mammalian cells. The glycosylation profile is mostly dependent on the production cell line and cell culture condition. Although biosynthetic production process generates several molecular variants, the level of molecular variation is tightly controlled in the upstream manufacturing process as well as the downstream process that can enrich for a certain molecular variant. Similar as CTL and sialic acid content, glycosylation profile can also be used as a measure of manufacturing consistency. In some embodiments, glycosylation profile of Efepoetin alfa is evaluated using method explained in Example 13 below and the results of N-glycan and O-glycan profiles demonstrate high similarity between EPO-Fc fusion proteins produced by GX-E2 cell line and GX-E4 cell line, indicating the manufacturing process of EPO-Fc fusion protein is robust and capable of producing drug substance with consistent quality attributes.

[192] The term "process control" used herein refers to a set of operation or process parameters and detection parameters that regulate and monitor the manufacturing process with the objective to deliver products that meet predefined acceptance criteria for each quality attributes to ensure the quality of the product. In some embodiments of the invention, the acceptance criteria for each quality attributes, the criticality of each quality attributes, the range of process parameter, the criticality of the process parameter, the control strategy including in-process control (IPC) and in-process monitoring or testing (IPM or IPT), and detection parameters are defined based on the process and analytical development results, the process characterization results, and manufacturing experience. The term "IPC" used herein means critical control parameter that needs to be fulfilled according to the acceptance criteria before the process can continue. The term "IPM" or "IPT" used herein refers to parameters that is collected during the manufacturing process for information purpose only.

[193] The term "Design of Experiment" or "DOE" refers to a systematic, efficient method that enables scientists and engineers to study the relationship between multiple input variables (a.k.a. factors) and key output variables (a.k.a. responses). It is a structured statistical approach for collecting data and making discoveries. DOE is a common approach in process characterization which one of the goals is to establish the functional relationship between input process parameters and output quality attributes. The results of the process characterization experiments are the identification of process parameters that have an impact on critical quality attributes (CQAs) of the product and their acceptance range or limit. The process parameters that have impacts on CQAs are called critical process parameters in which the acceptance range or limit need to be met in order to ensure the quality of the products.

[194] One of the embodiments of the invention is peptide mapping method that ensures the identity of the product. In this method, sample is denatured and N-glycans which are the main glycans in Efepoetin alfa are removed. The sample is then digested with trypsin and the resulting peptides are separated using UPLC. The unique chromatogram profile of the test sample is then compared with that of reference material. Since this method covers all tryptic peptides of Efepoetin alfa including the N- and C-terminal peptides, the same method can be used to assess the CTL content.

[195] The term "cell bank" used herein refers to a cell line that is preserved and stored under control conditions for future use in various biotechnological applications including creation of another cell bank and production of biologic drug substance. The term "Research Cell Bank" or "RCB" is cell bank that is used for research and development purpose. RCB is the end result of cell line development which is a process to develop a cell line that is intended for the production of biologic product. RCB is often used to generate a "Master Cell Bank (MCB)" which refers to an aliquot of a single pool of cells that has been prepared from the selected cell clone under defined conditions, dispensed into multiple containers, and stored under defined conditions, and serves as the primary source of cells for biologic drug production. The term "Working Cell Bank (WCB)" refers to cell bank prepared from MCB under defined culture conditions and used directly to manufacture biologic product. MCB and WCB are subjects to quality control testing, genetic stability assessment, and cell recovery assessment.

[196]

[197] EXEMPLIFICATION

[198]

[199] Example 1: Isoelectric focusing (IEF) gel electrophoresis method

[200] C-terminal lysine residue carries a one net positive charge while sialic acid carries one net negative charge. Since sialic acid constitutes the glycosylation profile, it also affects the electrical charge of the protein. The electrical charge property of Efepoetin alfa is measured using IEF gel electrophoresis in which the electrophoresis profile of the test sample is compared with Efepoetin alfa reference material. IEF gel electrophoresis is a method that separates proteins based on pI, the isoelectric point which is the pH at which a molecule carries zero net electrical charge. The experiment is conducted using IEF gel with pH gradient from 3.0 to 7.0. Samples are mixed with sample buffers 1:1 ratio and loaded onto the IEF gel. Electrophoresis running condition: 100 V for 1 hour, 200V for 1 hour, and 400 V for 45 min. The gel is then stained with Coomassie blue and destained using destaining solution. The representative example of IEF gel electrophoresis profile of Efepoetin alfa is shown in FIG. 9. The result shows heterogeneous bands between pI 4.5 and 5.3, indicating that Efepoetin alfa is acidic protein that has negative net charge at neutral pH. The IEF gel electrophoresis result reflects the CTL and sialic acid content.

[201]

[202] Example 2: Sialic acid content

[203] Sialic acid content which is part of the glycosylation profile is directly measured by calorimetric method using resorcinol and UV spectrophotometer. The samples are first treated with mild acid to liberate sialic acid, followed by reaction with resorcinol. The mixture is boiled at 100-105°C for 30 min which gives reactive intermediate that reacts with resorcinol to produce chromogen that can be extracted into butanol and butyl acetate to give chromophore. The resulting chromophore is finally measured at 580 nm using UV spectrophotometer. The amount of sialic acid is determined using a standard curve and the molar content of sialic acid per mole of Efepoetin alfa is calculated. A representative example of sialic acid content measurement is shown in Table 1. GX-E2 and GX-E4 are Efepoetin alfa produced by GX-E2 cell line and GX-E4 cell line, respectively. See Example 6 to Example 8 for explanation of the two cell lines.

[204]

[205] [Table 1]

[206] Sialic Acid Content of Efepoetin alfa

[207] Batch No. 747R16005 (GX-E2) EF18002-T (GX-E4) EF18003-T (GX-E4) EF19001-T (GX-E4) Average SD Sialic Acid Content (mol / mol) 23 22 24 24 23.25 0.96

[208]

[209] Example 3: Peptide mapping method

[210] CTL content can be detected using peptide mapping method. In brief, the sample is first concentrated. The sample is then mixed with denaturation buffer and heated at 95°C for 10 min. After that, the sample is treated with PNGase F to remove N-linked glycan at 37°C for 3 hours. The protein is precipitated using acetone at -20°C. The remaining free N-glycan is removed using 60% acetonitrile at -20°C for 1 hour. The sample is then denatured with 1M dithiotreitol at 37°C for 30 min. One M iodoacetic acid is added to the sample and incubated at 37°C for 30 min for carboxymethylation. Buffer exchange is performed, and the sample is digested with trypsin at 37°C for 4 hours. The trypsin digestion is stopped by adding trifluoroacetic acid. The digested peptides are separated using UPLC and detected with UV detector at 214 nm. The chromatogram profile is compared with the reference material. Representative example of peptide mapping result is shown in FIG. 10.

[211]

[212] Example 4: In vivo bioassay method

[213] Potency of Efepoetin alfa is measured using in vivo bioassay using normocythemic mice. In brief, ~8-week-old B6D2F1 mice (F1 cross between C57BL6 female and DBA2 male mice) are injected subcutaneously with 0.5 mL of 800, 400, and 200 ng / mL sample or reference standard diluted in 0.1% BSA solution. Four days after injection, blood samples are collected and analyzed by flow cytometer. The relative potency of the test article is compared with that of reference standard. Representative example is shown in FIG. 11.

[214]

[215] Example 5: The effect of sialic acid content on efficacy of Efepoetin alfa

[216] Sialic acid content is known to determine the half-life of Efepoetin alfa in vivo. To determine the impact of sialic acid content on Efepoetin alfa efficacy in vivo, Efepoetin alfa with various sialic acid content were produced using GX-E2 cell line and purified. Five batches of Efepoetin alfa with sialic acid content ranging from 19.1 to 22.6 mol / mol were injected into five groups of rats. The PK profile and the change of hemoglobin level in each group are monitored, and the results are shown in Table 2 and FIG. 12, respectively.

[217]

[218] [Table 2]

[219] PK Profiles of Efepoetin Alfa with Various Sialic Acid Content

[220] Test Article Sialic Acid Content (mol / mol) AUEC last (g*hr / dL) P-value (one-way ANOVA) Vehicle - 4.06 ± 12.67 0.008 GX-E2 reference standard 22.6 32.05 ± 25.57 - GX-E2 #1 19.1 6.67 ± 6.98 0.023 GX-E2 #2 20.0 20.95 ± 15.70 0.561 GX-E2 #3 22.0 31.75 ± 12.36 1.000 GX-E2 #4 21.9 33.41 ± 9.13 1.000

[221]

[222] As shown in Table 2, GX-E2 reference standard, GX-E2 #3, and GX-E2 #4 that had sialic acid content between 21.9 and 22.6 mol / mol had very similar exposure. They also induced similar hemoglobin responses (FIG. 12). On the other hand, the GX-E2 #1 with sialic acid content of 19.1 mol / mol had a significantly lower exposure compared to the GX-E2 reference standard (Table 2) and this is correlated with a significantly lower hemoglobin response (FIG. 12). The GX-E2 #2 with sialic acid content of 20.0 mol / mol also had ~30% lower exposure and had a slightly reduced hemoglobin response compared to the reference standard, but the changes were not significant. The vehicle control group showed no change in the hemoglobin level. Altogether, these results suggest that despite having Fc fragment, Efepoetin alfa efficacy in vivo is unexpectedly dependent on sialic acid content.

[223]

[224] Example 6: Transfection of a-2,3-sialyltransferase to GX-E2 cell line

[225] Considering that high sialic acid content is required for the extended half-life and long-acting activity of Efepoetin alfa, it is important to have the Efepoetin alfa production cell line that can produce highly sialylated proteins.

[226] In CHO cells, efepoetin alfa undergoes N-glycosylation through post-translational modification (PTM). Since N-glycosylation occurs only at the aspartic acid position of the protein, additional N-glycosylation will not take place unless a mutation occurs. When the target protein is overexpressed in CHO cells, the process that typically facilitates sufficient N-glycosylation through PTM may not proceed adequately, resulting in insufficient sialic acid production at the desired level. Therefore, it is crucial to maintain the proper level of existing N-glycosylation. To address this, a-2,3 sialyltransferase has been added. The basic principle is that Neu5Ac (N-acetylneuraminate) and CTP are synthesized into CMP-Neu5Ac by an enzyme in the nucleus. In the cytoplasm, CMP-Neu5Ac is hydroxylated to form CMP-Neu5Gc. Subsequently, it is transported into the Golgi apparatus through the sialic acid transporter located in the Golgi membrane. In the Golgi, it is incorporated into glycoproteins by sialyltransferases. a-2,3 sialyltransferase facilitates the addition of sialic acid to sugars that are attached to galactose. Therefore, a-2,3 sialyltransferase plays an important role in promoting the production of sialic acid in efepoetin alpha (Patrick et al, 2009, Optimal and consistent protein glycosylation in mammalian cell culture Figure 2).

[227] Specifically, to increase the level of protein sialylation, ST3GAL4 gene encoding a-2,3-sialyltransferase which is an enzyme that catalyzes the transfer of sialic acid to carbohydrate groups of glycoproteins or glycolipids is transfected into the Efepoetin alfa producing GX-E2 cell line. Amino acid sequence of a-2,3-sialyltransferase with HA tag is shown in FIG. 13, SEQ ID NO.3). To generate stable cell lines expressing a-2,3-sialyltransferase, GX-E2 master cell bank (MCB) cells were transfected with a-2,3-sialyltransferase-HA expression plasmid, pCI-hygro-a-2,3-st-IRES-pERP (FIG. 14), by the electroporation method. The schematic diagram of GX-E4 cell line generation is shown in FIG. 15. Briefly, GX-E2 MCB cells were thawed and subcultured in Ex-Cell CHO DHFR- media containing 4 mM glutamine, 1 g / L poloxamer 188, and 200 nM methotraxate (MTX) for several passages until the cell growth was stable. Cells with >95% viability was mixed with the pCI-hygro-a-2,3-st-IRES-pERP plasmid and the mixture was subjected to electroporation at voltage of 1400 V, 1600 V, and 1700 V. The transfected cells were cultured in 6-well plate and subjected to hygromycin selection.

[228]

[229] Example 7: Hygromycin selection and characterization of the transfected cell pools

[230] Forty-eight hours post-transfection, the transfected cells were seeded at a viable cell density of 4.5 x 104 cells / mL in 4-5 mL EX-CELL CHO DHFR- medium containing 4 mM L-glutamine, 200 nM MTX, 1 g / L poloxamer 188, and 100 ug hygromycin. The cells were cultured for 16 days at 37°C and 5% CO2. The hygromycin selection was completed when the cell viability exceeded 90%. Five subpools of transfected cells (ie. #6P, 7P, 8P, 9P, and 10P) were tested to confirm the presence of a-2,3-sialyltransferase by Western blot using anti HA antibody (FIG. 16) and to confirm the increase of sialylation using IEF gel electrophoresis (FIG. 17). All five transfected pools express a-2,3-sialyltransferase and produce more sialylated proteins (lanes 3 to 7, FIG. 17) compared to the untransfected GX-E2 cells (lane 8, FIG. 17) as indicated by the increase in intensity of bands with lower pI.

[231]

[232] Example 8: Single cell subcloning and characterization of GX-E4 subclones

[233] The stable clone of a-2,3-sialyltransferase transfected GX-E2 cell line is called GX-E4 cell line. Single cell subcloning of the transfected subpools #8P, 9P, and 10P was performed using limiting dilution (FIG. 15). The cells were seeded 0.5-1 cell per well in 96-well plates (10 plates for each subpool). The clonality of the subclones was documented with Clone Select Imager at 4 hours (Day 0), 7 days and 14 days after seeding. After 14 days of incubation, culture medium was changed to fresh medium, and the specific productivity was assessed by the Fc ELISA the next day. Based on the result, a total 9 clones with high productivity were selected and expanded. Five subclones were successfully expanded and tested for mycoplasma, specific productivity, and sialylation level. All samples were negative for mycoplasma. Compared to GX-E2 cell line, the GX-E4 subclones did not show any significant difference in specific productivity. However, GX-E4 subclones #8P-4 and #10P-1 produced more acidic proteins compared to the GX-E2 WCB as shown in FIG. 18. Therefore, these two subclones were selected as lead subclones and used to manufacture research cell bank (RCB). These subclones showed a good stability in term of cell growth and Efepoetin alfa production up to 34 passages in shake flask culture as well as fed-batch bioreactor culture at passages 7, 14, 21, 28, and 34. The result of long-term stability is shown in Table 3.

[234]

[235] [Table 3]

[236] Long term stability of two lead RCBs

[237] Clone Media condition Td(hr) Productivity Reduction rate Candidate 3 W + MTX 200nM + Hygromycin 100ug / ml + Poloxamer 188 1g / L 30.7 - 29.3 7.28% W / O + Poloxamer 188 1g / L 28.5 - 26.0 19.79% Candidate 5 W + MTX 200nM + Hygromycin 100ug / ml + Poloxamer 188 1g / L 31.5 - 30.6 9.45% W / O + Poloxamer 188 1g / L 31.2 - 26.6 34.32%

[238] Candidate 3: GX-E4 clone #8P-4, Candidate 5: GX-E4 clone #10P-1.

[239] Td: doubling time in hour.

[240] Productivity reduction rate: comparison of specific productivity between passage 1 and 34.

[241] Media condition W: with methotrexate and hygromycin,

[242] Media condition W / O: without methotrexate and hygromycin.

[243]

[244] As shown in Table 3, candidate 3 (clone #8P-4) shows slightly better doubling time and lower reduction in productivity over time. Based on this result, GX-E4 #8P-4 RCB was selected to generate GX-E4 master and working cell banks (MCB and WCB). All cell banks have been thoroughly evaluated and the results met the acceptance criteria according to regulatory guidelines.

[245]

[246] Example 9: Upstream manufacturing process of Efepoetin alfa

[247] The goals of upstream process are to express Efepoetin alfa protein in cell culture and harvest it in sufficient quantity for subsequent purification and formulation. The upstream process starts with the thawing of WCB followed by several passages of cell expansion in shake flasks. The content of shake flasks is then pooled and inoculated into a fed-batch bioreactor. The bioreactor culture is harvested and clarified using depth filtration or continuous centrifugation before downstream processing. The schematic of the upstream process is shown in FIG. 23. The upstream process is performed using single-use or sanitized product contact equipment in a controlled environment. Each process step is carefully controlled by defined process parameters and input as well as output process control parameters. This control strategy is derived from thorough process development and process characterization experiments as well as manufacturing experience at lab scale, pilot scale and commercial scale.

[248] The present invention includes the upstream process presented in FIG. 23. This new cell culture mode was established and used to produce clinical trial material for phase 3 clinical study. The perfusion culture that was used previously has several disadvantages including difficulty to find optimal protein production condition, high risk of contamination due to long culture duration, and difficulty in scaling up. Accordingly, in the present invention, the culture mode is changed to a fed-batch mode that minimizes the risk of contamination through the usage of bioreactor in a shorter duration between 10-15 days.

[249] The production bioreactor step is described as follows. The bioreactor is filled with culture medium (EX-CELL CHO DHFR(-), 4 mM glutamine, 1.0 g / L Poloxamer 188) and preheated before cell inoculation. Seed cells from passage 8 are inoculated into the production bioreactor with initial seeding density of 400,000 to 700,000 cells / mL. The viability after inoculation should be > 90%.

[250] Culture in the bioreactor is maintained at 34.0 to 38.0°C and pH of 6.7 to 7.5 with dissolved oxygen (DO) at 50%. The pH, cell viability, cell density, and concentrations of metabolite such as glucose, lactate, glutamine, glutamate, and ammonia are measured regularly. Other parameters such as temperature, stirring speed, DO concentration, and pH are controlled throughout the culture period. On day 3 or 4 after inoculation, the temperature is changed to 29.0 to 36.0°C lowering 2 to 5°C . During the culture period, various feeds are added at predetermined concentration and time.

[251] The content of the production bioreactor is harvested when the cell viability drops less than 70%. The cell culture harvest is clarified using depth filtration or continuous centrifugation to remove the cells and cell debris. The titer and purity of the target protein are monitored. Compared to the previous process with GX-E2 cell line, the new process which is the subject of the current invention increase the titer 7.4 times (Table 4).

[252]

[253] [Table 4]

[254] Comparison betwee GX-E2 and GX-E4 upstream processes

[255] Cell line Culture mode Titer (mg / L) Purity (%) GX-E2 Perfusion 126.5 99.9 GX-E4 Fed-batch 931 93.8

[256]

[257] To enhance glycosylation (including sialic acid content), various feeds were screened. Table 5 shows the culture condition used in this experiment. Table 6 lists various feeds and concentrations that were tested in this experiment.

[258]

[259] [Table 5]

[260] Culture condition

[261] Basal media EX-CELL CHO DHFR(-) + glutamine 4 mM + methotrexate 200 nM + hygromycin 100 ^g / ml Cell line GX-E4 Culture mode Fed-batch Temperature 29.0 to 38.0°C Working volume 60 mL / 250 mL SF Feeding Feed A, B, C Test feed addition time Day 3 Culture duration Cell viability > 80%

[262]

[263] [Table 6]

[264] Test feeds

[265] No. Feed Concentration 1 None NA 2 Dexamethasone 1 uM 3 Galactose 10 mM 4 Glucosamine 10 mM 5 Glucosamine + Uridine 10 mM + 5 mM 6 ManNAc (N-acetyl-D-manoseamine) 20 mM 7 ManNAc + Cytidine 10 mM + 10 mM

[266]

[267] [Table 7]

[268] Results of various feeds to enhance glycosylation

[269] No. Feed Purity (%) Titer (g / L) Culture duration (days) 1 None 87.6 1.2 11 2 Dexamethasone 83.8 1.3 12 3 Galactose 84.2 1.3 12 4 Glucosamine 61.1 0.7 13 5 Glucosamine + Uridine 48.7 0.2 9 6 ManNAc 88.8 0.9 9 7 ManNAc + Cytidine 89.5 0.7 9

[270]

[271] The results of feed screening are shown in Table 7, FIG. 19, and FIG. 20. The results in Table 7 show that ManNAc or ManNAc + cytidine addition increased the purity, while dexamethasone or galactose addition increased the titer. In regard to cell growth, dexamethasone, galactose, or ManNAc addition did not change the growth curve (FIG. 19). In regard to IEF gel electrophoresis, addition of glucosamine, glucosamine + uridine, or ManNAc + cytidine increased the basic species compared to the control group, while the addition of galactose or ManNAc increased the acidic species (FIG. 20). Based on these results, galactose and ManNAc were selected as feeds in the production bioreactor step.

[272] Next, feeding conditions of galactose and ManNAc were optimized using design of experiment (DOE) (see Example 11 for detailed description of DOE). Sixteen experimental runs were conducted using ManNAc concentration ranging from 0 to 20 mM and galactose concentration ranging from 0 to 40 mM. The results are shown in Table 8, FIG. 21, and FIG. 22. It was confirmed that as the ManNAc concentration increased, the purity increased but the titer decreased (Table 8). In addition, it was confirmed that lane 6 and 8 in FIG. 21 had lower pI values compared to other experimental results. Based on the DOE results, prediction profiler plot was generated. Based on the profiler plot (FIG. 22), productivity, IEF gel profile, etc., ManNAc 10 mM and galactose 32 mM were found to be the optimal feeding concentrations.

[273]

[274] [Table 8]

[275] DOE results of galactose and ManNAc feeding conditions

[276] Run ManNAc(mM) Galactose(mM) Titer(%) Purity(%) Culture days 1 20 0 85.0 82.9 11 2 8.7 22.8 126.4 74.3 15 3 0 16.2 110.1 80.3 13 4 0 0 100.0 80.8 12 5 8.7 22.8 99.5 81.9 12 6 8.7 22.8 95.5 79.5 11 7 13.9 0 86.1 83.0 11 8 3.8 30.4 107.5 77.6 13 9 11.1 40 98.0 78.6 13 10 20 40 76.3 82.6 11 11 20 12 79.6 82.3 11 12 0 40 99.2 78.7 12 13 8.7 22.8 93.8 80.9 12 14 20 40 77.4 82.4 11 15 20 12 77.4 83.5 11 16 7 7.4 105.7 80.4 13

[277]

[278] Another key finding in the upstream process development is that the duration of the production bioreactor step should be limited to 10 days. Many bioprocesses use a 14-day period to maximize cell growth and production of protein of interest. However, Efepoetin alfa is a unique product that relies on high sialic acid content thus the process needs to maximize the production of acidic species. As shown in Table 9, the total protein titer and viable cell density increased significantly when the culture period was extended to 14, 15, and 16 days.

[279]

[280] [Table 9]

[281] Process development result: Duration of bioreactor culture

[282] Duration of Bioreactor Culture (days) Titer (g / L) Integral Viable Cell Density (x 106 cells-day / mL) 10 1.2 25.7 14 2.4 42.4 15 2.6 45.6 16 2.9 48.4

[283]

[284] However, as shown in FIG. 24, the cell cultures of 14- to 16-day period contain less acidic species compared to the cell culture of 10-day period. Because of this result, the production bioreactor culture period is set for 10 days with the feed schedule as shown in FIG. 23.

[285]

[286] Example 10: Downstream manufacturing process of Efepoetin alfa

[287] The downstream process goals are to purify Efepoetin alfa by removing process-related and product-related impurities and potential adventitious agents, as well as enrich for the acidic species that is high in sialic acid content. The Efepoetin alfa drug substance (DS) is purified from cell culture supernatant with the following unit operations (refer to FIG. 25):

[288] • Protein A affinity chromatography that captures the Efepoetin alfa using Protein A. ligand that binds the Fc fragment of the protein.

[289] • Virus inactivation using low pH buffer at 10 - 15°C.

[290] •  Depth filtration to remove large impurities such as particulates / precipitates that form during the previous process steps.

[291] • Ion exchange chromatography (IEX) 1 using ceramic hydroxyapatite type II resin to remove product-related impurities such as oligomers and process-related impurities such as host cell DNA.

[292] • Diafiltration for buffer exchange.

[293] •  IEX2 chromatography using Q Sepharose Fast Flow (QFF) resin which is a strong anion exchanger. This chromatography step purifies the acidic species of Efepoetin alfa. At the same time, it removes the basic species and other process-related impurities.

[294] • Virus filtration to remove any remaining viruses.

[295] • Ultrafiltration and diafiltration (UF / DF) 2 to concentrate and formulate Efepoetin alfa DS.

[296] • Add polysorbate 20 to the formulated DS, adjust the final concentration, and store in freezer.

[297] The downstream process is performed using single-use or sanitized product contact equipment in a controlled environment. Each process step is carefully controlled by defined process parameters and in process control parameters to ensure product quality during the manufacturing process. This control strategy is derived from thorough process development and process characterization experiments as well as manufacturing experience at lab scale, pilot scale and commercial scale.

[298] One distinguishing characteristic of the downstream process of Efepoetin alfa is the purification of acidic species by anion exchange chromatography with positively charged QFF resin at the IEX2 step. This step utilizes buffer with pH of 6.10-6.30 to facilitate the binding of protein species with pI < 5.3. After extensive washing, the bound acidic species is eluted with buffer having a conductivity of 12.10 - 14.9 mS / cm (FIG. 25).

[299] Compared to GX-E2 process, the downstream purification process of current invention contains the following changes:

[300] • All purification steps except the virus inactivation step are performed at controlled room temperature.

[301] • Due to increased production scale, larger columns are used.

[302] • The order of the virus filtration process and the ultrafiltration / diafiltration process has been changed. (Ultrafiltration / diafiltration is performed after virus filtration)

[303]

[304] The details of downstream purification process of Efepoetin alfa are described below:

[305] (1) Protein A Affinity Chromatography

[306] Affinity chromatography is the first step in a series of purification steps of the downstream process. The strategy used in affinity chromatography is to bind protein A ligands to the Fc region of the target protein. The target protein is selectively isolated from the protein pool present in the clarified harvest. After the equilibration step, the clarified harvest is loaded onto the column. Thereafter, re-equilibration and washing steps are performed to remove impurities weakly bound to the column. Finally, the Fc fusion protein is eluted from the column using low pH buffer.

[307] (2) Virus Inactivation - Low-pH Treatment

[308] The eluate from affinity chromatography is adjusted to a temperature of 12.5 ± 2.5°C using a temperature control mixer. The pH is adjusted to 3.65 ± 0.1 using pre-cooled Protein A elution buffer while stirring at 50 - 70 rpm. After reaching the target pH, the solution is incubated for 120 ± 10 minutes. The solution is then neutralized with 1M Tris buffer until the pH reached 6.9 ± 0.1.

[309] (3) Deep Filtration

[310] Depth filtration is performed to remove host cell proteins (HCPs) remaining in the solution. This step is also performed to remove particulates / sediment formed during the pH adjustment step in virus inactivation. The depth filter is wetted with water for injection (WFI) and buffer. After sufficient wetting, the solution is passed through the depth filter. Thereafter, post-washing with a buffer is performed to wash out the remaining material in the depth filter. Upon completion of depth filtration, the solution is filtered through a 0.2-pm filter.

[311] (4) Ion Exchange Chromatography 1 (IEX1)

[312] Ion exchange chromatography is the second chromatography step. It is operated in bind-elute mode to remove oligomeric impurities. The column is then washed to reduce impurities and endotoxin levels. After the equilibration step, the depth filtration solution is loaded onto the column. Thereafter, a re-equilibration step is performed to remove impurities that remain in the column. An elution step is then performed. The target protein is eluted first, and impurities remained on the column. The remaining impurities bound to the column were washed out with column wash buffer. Before washing, an equilibration step of reducing phosphate levels is performed to prevent crystal formation that could interfere with the resin performance.

[313] (5) Ultrafiltration / Diafiltration 1 (UF / DF 1)

[314] The diafiltration step is performed in TFF mode using a polyethersulfone (PES) membrane with a molecular cutoff of 30 kDa. Before sample loading, the system (skid and UF filter) is cleaned with 0.5M NaOH and washed with WFI to remove the cleaning solution. After equilibrating the membrane with equilibration buffer, each sample is then loaded into the system and the buffer is replaced with loading buffer for the next chromatography step. The resulting material is trapped by the filter and remained as retentate, and the impurities flowed through the filter and were discarded as permeate. When the pH and conductivity of the solution from the permeate are the same as the pH and conductivity of the exchange buffer, the diafiltration process is completed.

[315] (6) Ion Exchange Chromatography 2 (IEX2)

[316] Ion exchange chromatography (anion exchange) is the final chromatography step. It is operated in bind-elute mode for the purpose of removing basic protein impurities and fragment proteins. After the equilibration step, the diafiltration retentate is loaded onto the column. Thereafter, a re-equilibration step is performed to remove impurities that could remain in the column. Next, an elution step is performed. The target protein is eluted first, and impurities remains on the column. The remaining impurities bound to the column are washed out with column wash buffer. Once the process iss complete, the column is washed again to reduce impurities and endotoxin levels.

[317] (7) Virus Filtration

[318] Virus filtration is performed to remove virus particles present in the solution. It is operated in normal flow filtration mode at constant pressure. Pressure vessels are used as containers to hold liquid, and nitrogen gas is used to apply a constant pressure in order to facilitate virus filtration. In tandem mode, a pre-filter with a pore size of 0.1 pm is used before the virus filter. After sufficiently wetting with WFI and buffer, the solution is passed through a virus filter. Once virus filtration is completed, the solution is filtered through a 0.2- pm filter.

[319] (8) Formulation Exchange (UF / DF-2)

[320] The formulation exchange (UF / DF-2) step is performed in TFF mode using a PES membrane with a molecular cutoff of 30 kDa. Before sample loading, the system (skid and UF filter) is cleaned with 0.5M NaOH and washed with WFI to remove the cleaning solution. After the equilibration step, each sample was loaded into the system. The resulting material is trapped by the filter and remained as retentate, and impurities flowed through the filter and are discarded as permeate. After reaching the target concentration, the buffer is replaced. When the pH and conductivity of the solution from the permeate are the same as the pH and conductivity of the exchange buffer, the buffer exchange process is completed.

[321] The results of purification and analysis are shown in Tables 10 and 11 below. It was confirmed that when the temperature condition is changed from cold to room temperature, and the buffer and pH in each purification step is optimized accordingly, the yield increases by 2.7 times (based on 10 days culture harvest).

[322]

[323] [Table 10]

[324] Comparison between purification process conducted at cold and room temperature

[325] GX-E4 Cold purification Room temperature purification NEW_ST 10 day NEW_ST 11 day NEW_ST 10 day NEW_ST 11 day Step recovery (%) Purity (%) Step recovery (%) Purity (%) Step recovery (%) Purity (%) Step recovery (%) Purity (%) Input 100 84.0 100 82.6 100 84 100 82.6 PA 85.1 90.4 83.6 87.3 91.8 87.4 89.2 85.5 CHT II 84.8 95.1 84.2 94.0 84.0 95.1 84.2 93.8 UF / DF 94.9 95.7 94.2 93.7 95.8 95.3 95.8 93.9 Q 10.3 98.6 9.3 96.4 25.2 98.6 23.1 98.7 Final 7.1 95 6.2 93.0 18.6 97.3 16.6 97.2

[326]

[327] [Table 11]

[328] Quality parameters of product purified at cold and room temperature

[329] ST10 day, cold ST11 day, cold ST10 day, RT ST11 day, RT UV 1.48 mg / mL 1.51 mg / mL 1.54 mg / mL 1.77 mg / mL SE-HPLC 95.01 93.04 97.31 97.16 RP-HPLC 95.77% 95.78% 95.27% 95.00% SA content 20.6 mol / mol 21.1 mol / mol 20.2 mol / mol 21.5 mol / mol

[330]

[331] Example 11: DOE study for production bioreactor step

[332] One of the goals of process characterization experiments is to establish the functional relationship between input process parameters and output quality attributes. These experiments would identify critical process parameters (CPP) which are the process parameters that have an impact on critical quality attributes (CQAs) of the product. CQAs are the product key quality attributes that affect safety, efficacy, PK / PD, and immunogenicity in patients.

[333] Process characterization was initiated after finalization of the process at the intended manufacturing scale. The process characterization experiments were conducted in scale down models that had been qualified so that they are representative of the manufacturing scale. Failure Model Effect Analysis (FMEA, referred in ICH Q8[R2]) tool is first used to identify all potential CPPs (pCPPs). Using the pCPP, process characterization experiments were conducted using a univariate or multivariate experimentation approach. Based on the outcome of these experiments, each pCPP is categorized into either CPP or non-CPP. The acceptance range for these CPPs (input process parameters) are derived based on the output attributes acceptance range. The acceptance range for non-CPPs was derived from process development / characterization and manufacturing batches.

[334] Besides the product quality, the process characterization experiments can also identify process parameters that influence process performance such as yield. Described below is the multivariate design of experiment (DOE) results that were used to improve the yield of the manufacturing process.

[335] To further characterize the production bioreactor process, DOE with Response Surface Methodology (RSM) were conducted using a qualified scale down model at 1L scale. In this experiment, we focused on 4 factors (i.e. pH, shift temperature, feed timing, and galactose timing) with other process parameters fixed as in the manufacturing process described in FIG. 25. Since the final yield depends on the small proportion of GX-E4 with high sialic acid content, which is hard to discriminate at the clarification step, the product is purified up to IEX2 step. This experiment measured 7 responses as follows:

[336] • Responses for control strategy: Purity at IEX2 step (%)

[337] • Responses for information: Final Viable Cell Number, Final Cell Viability (%), Yield at bioreactor step (mg), Purity at bioreactor step (%), IEX2 step yield (%), and Total Yield (%)

[338] The lower screening range (LSR) and upper screening range (USR) for the three factors are shown in Table 12.

[339]

[340] [Table 12]

[341] Production Bioreactor - Screening Range of the Four Factors

[342] No. Factor LSR USR Type 1 PH 6.6 7.0 Continuous 2 Shift Temperature 32°C 34°C Categorical 3 Feed Timing day 3, 5, 7, 9 (d3) day 4, 6, 8 (d4) Categorical 4 Galactose Timing day 5, 7, 9 (d5) day 4, 6, 8 (d4) Categorical

[343]

[344] A total of 18 runs were conducted based on central composite design with 4 factors and 2 central points. This experiment enables a response surface modelling of all main factors and two-factor interactions. The result is shown in Table 13.

[345]

[346] [Table 13]

[347] Production Bioreactor - Response Data for Each DOE Run

[348] Eik PH Shift temp Feed timing Galactose timing Viable Cell Number Cell Viability (%) Yield at Biorx (mg) Purity at Biorx (%) IEX2 Yield (mg) IEX2 - Purity (%) Total Yield (%) 1 6.6 34 d4 d4 2460500000 75.14 190.92 89.93 24.9 98.15 13 2 6.6 32 d4 d4 2371950000 81.03 179.4 89.9 21.78 97.85 12 3 7 34 d3 d4 1519406250 43.71 354.996 90 40.59 98 11 4 6.8 32 d4 d5 1855000000 70 258.24 89.1 24.6 100 10 5 6.8 34 d3 d5 3569000000 85.89 392.756 95 56.74 98 14 6 7 32 d3 d5 1023750000 38.4 263.52 84.78 19.46 98 7 7 7 34 d4 d5 1515628125 49.83 271.375 83 44.82 98 17 8 6.6 32 d3 d5 1522125000 71.05 149.896 85.06 14.12 96 9 9 6.8 34 d4 d4 1726312500 55.63 262.764 85.26 40.47 98 15 10 6.8 32 d3 d4 1893437500 77.66 216.796 89 26.84 97 12 11 6.6 32 d3 d4 1114762500 63.83 126.684 84 9.7 96 8 12 6.8 34 d4 d4 2976750000 85.84 290.628 90 36.09 99 12 13 6.8 32 d4 d5 1476656250 82.37 236.052 91 30.5 98 13 14 6.6 34 d3 d4 1077359375 57.19 148.732 85 11.42 97 8 15 6.6 34 d4 d5 1151325000 62.63 134.057 83 7.88 97 6 16 7 32 d4 d4 1272375000 42.39 194.58 89 22.7 98 12 17 6.8 34 d3 d5 3198125000 84.64 364.704 91 61.42 98 17 18 6.8 32 d3 d4 2651187500 92.62 274.456 91 30.5 98 11

[349]

[350] The statistical analysis of the data produces prediction profiler plots shown in FIG. 26. Based on the prediction profiler plots, pH 7.0 condition which was the manufacturing process setting at that time has a lower viable cell number, % cell viability and slightly less total yield compared to pH 6.9 and pH 6.8 conditions.

[351]

[352] Example 12: Modification of upstream process to improve final yield

[353] Based on the process characterization results, we reduced the pH in the production bioreactor step from 7.0 to 6.8. The temperature shift was also delayed from day 3 to day 4 to allow the cells achieve optimal growth. These minor changes resulted in significantly meaningful outcomes.

[354] First, they improved the final viable cell density (VCD) and significantly reduced the lactate concentration (FIG. 27). The new culture condition allowed cells to grow steadily and reach the peak VCD of around 5,000,000 cells / mL on day 10 (see batch # 4-6 in FIG. 27). Whereas, in the old culture condition, the VCD reached the peak around day 7 and declined afterwards (see batch # 1-3 in FIG. 27). Similarly, the lactate concentration in the new culture condition was kept low (< 2g / L) over the entire 10-day culture period, while it steadily increased reaching > 4g / L at day 10 in the old culture condition. In CHO cell systems, lactate concentration exceeding 40 mM (or 3.6g / L) severely impair cell growth (Torres et al., 2018).

[355] Second and most importantly, in the 50L scale, the new condition increased the final yield by more than 20% from an average of 6.69 grams to 8.11 grams (Table 14). The percent yield also improved significantly from an average of 13.10% to 31.73%. The difference in percent yield is mainly due to the increase of acidic species production rather than the total protein production. The total protein produced by the new culture condition was ~25.6 grams while the old culture condition yielded ~51 grams.

[356]

[357] [Table 14]

[358] Yield comparison between the two production bioreactor conditions

[359] Condition pH 7.0 pH 6.8 Batch no. 1 2 3 4 5 6 Final yield amount (gram) 7.06 6.32 6.68 8.6 8.05 7.68 Average final yield amount 6.69 8.11 SD yield amount 0.37 0.46 Total protein at bioreactor harvest 53.28 47.38 52.60 25.89 26.17 24.61 % yield 13.25 13.34 12.7 33.22 30.76 31.21 Average % yield 13.10 31.73 SD % yield 0.35 1.31

[360]

[361] The results shown in Table 3 and Table 14 suggest that there is an inverse correlation between total protein produced in the bioreactor and the sialylation level of Efepoetin alfa. This finding is supported by Lewis et al. (Lewis et al., 2016) in which they discovered that the sialic acid level decreased as the protein titer increased over the 14-day period of fed-batch culture of Fc-fusion protein.

[362]

[363] Example 13: Consistent product quality of Efepoetin alfa

[364] The impact of production cell line change (from GX-E2 to GX-E4) and upstream process change (from pH 7.0 to pH 6.8) to the quality attributes of Efepoetin alfa has been thoroughly evaluated. The studies include the following attributes:

[365] • Appearance, pH

[366] • Content: protein content, excipient concentration, sialic acid content

[367] • Product purity / Product-originated impurities: by SE-HPLC (Size Exclusion - High Performance Liquid Chromatography) that separates the molecules by size, by RP-HPLC (Reverse Phase - High Performance Liquid Chromatography) that separates molecules by electric charge.

[368] • Functional attributes: in vivo bioassay.

[369] • Host and Process-derived impurities: host cell DNA (HCD), host cell proteins (HCP), and residual protein A.

[370] •  Safety: Sterility, Endotoxin

[371] •  Structural analysis:

[372] - Primary structure: amino acid composition, peptide mapping, amino acid sequencing, N and C terminal integrity, Post-transcriptional modifications (deamidation and oxidation).

[373] - Higher order structure: disulfide bonds, free thiols, and circular dichroism.

[374] - Glycosylation: monosaccharide composition, sialic acid content, N-glycan profile, and O-glycan profile.

[375] - Molecular weight: Reduced / deglycosylated protein, non-reduced protein

[376] Due to the unique biosynthetic production process and molecular characteristics of biological products, the biologic drug substance can include several molecular variants. The degree of molecular variance is controlled by a robust manufacturing process which is product specific. While most of the quality attributes mentioned above (e.g. process-derived impurities, purity, and general attributes) are controlled in the downstream purification process, the structure and functional attributes are mainly dependent on the production cell line and the upstream process. Three characteristic quality attributes of Efepoetin alfa that represent manufacturing consistency are sialic acid content, % CTL truncation, and unique glycosylation profiles.

[377] The N and C-terminal integrity was evaluated by endoprotease digestion followed by RP HPLC in combination with UV detection and mass spectrometry identification. The quantitation of the N and C terminal variants was based on the extracted ion chromatograms. While the N-terminus of Efepoetin alfa produced by GX-E2 and GX-E4 cell lines is intact, the C-terminal peptide contains two species, the intact peptide and peptide that lacks CTL. The ratio of the modified and unmodified peptides is provided in table 15. Majority of Efepoetin alfa protein has a truncated C-terminus. The frequency of C-terminal truncation is largely similar between the 4 batches of Efepoetin alfa.

[378]

[379] [Table 15]

[380] Frequency of C-terminal Lys truncation

[381] Batch no. Intact (%) C-terminal Lys truncation (%) 747R16005 (GX-E2) 3.3 96.7 EF18002-T (GX-E4) 4.9 95.1 EF18003-T (GX-E4) 5.0 95.0 EF19001-T (GX-E4) 6.9 93.1

[382]

[383] To determine the glycosylation sites, Efepoetin alfa was first digested with Glu-C or trypsin, the resulting peptides were separated using RP-UPLC and analyzed using Q-TOF MS. The extracted ion chromatogram of oxonium ion that is present in the N-acetylhexosamine of the glycopeptide is then compared with the corresponding total ion chromatogram (TIC) to confirm the identity of the glycopeptides and identify the glycosylation sites. There are 4 N-glycosylation sites (N24, N38, N83, and N261)) and 1 O-glycosylation site (S126) found in Efepoetin alfa. The N-linked glycosylation sites match the respective consensus sequence, while there is no consensus sequence of O-linked glycosylation site.

[384] To determine the N- and O-linked glycan profile, the deconvoluted masses of the respective glycopeptide were measured and compared to the theoretical masses calculated from the amino acid sequence plus various forms of glycan. The relative content of various forms of N-glycan was calculated from the signal intensity and the results of the analysis are summarized in FIG. 28 to FIG. 32. The results show that Efepoetin alfa produced by GX-E2 and GX-E4 cell lines including the GX-E4 cell line produced by the modified upstream process (i.e. batch EF19001-T) are structurally comparable. Majority of N-glycosylation sites are glycosylated while ~60% of O-glycosylation site does not have any glycan attached.

[385] Altogether the evaluation results of the structure and quality of Efepoetin alfa including the sialic acid content (Table 1), % CTL truncation (Table 15), and the glycosylation profile (FIG. 28 - 32), wherein the sialic acid content and the glycosylation profile can be determined by IEF gel electrophoresis, demonstrate that the manufacturing process described herein is robust and capable to produce Efepoetin alfa drug substance with consistent quality attributes.

[386]

[387] SEQ ID NO: 1. a-2,3-sialyltransferase (nucleotide sequence)

[388] ATGGTCAGCA AGTCCCGCTG GAAGCTCCTG GCCATGTTGG CTCTGGTCCT 50

[389] GGTCGTCATG GTGTGGTATT CCATCTCCCG GGAAGACAGT TTTTATTTTC 100

[390] CCATCCCAGA   GAAGAAGGAG   CCGTGCCTCC   AGGGTGAGGC AGAGAGCAAG 150

[391] GCCTCTAAGC TCTTTGGCAA CTACTCCCGG GATCAGCCCA TCTTCCTGCG 200

[392] GCTTGAGGAT TATTTCTGGG TCAAGACGCC ATCTGCTTAC GAGCTGCCCT 250

[393] ATGGGACCAA GGGGAGTGAG GATCTGCTCC TCCGGGTGCT AGCCATCACC 300

[394] AGCTCCTCCA TCCCCAAGAA CATCCAGAGC CTCAGGTGCC GCCGCTGTGT 350

[395] GGTCGTGGGG AACGGGCACC GGCTGCGGAA CAGCTCACTG GGAGATGCCA 400

[396] TCAACAAGTA CGATGTGGTC ATCAGGTTGA ACAATGCCCC AGTGGCTGGC 450

[397] TATGAGGGTG ACGTGGGCTC CAAGACCACC ATGCGTCTCT TCTACCCTGA 500

[398] ATCTGCCCAC TTCGACCCCA AAGTAGAAAA CAACCCAGAC ACACTCCTCG 550

[399] TCCTGGTAGC TTTCAAGGCA ATGGACTTCC ACTGGATTGA GACCATCCTG 600

[400] AGTGATAAGA AGCGGGTGCG AAAGGGTTTC TGGAGACAGC CTCCCCTCAT 650

[401] CTGGGATGTC AATCCTAAAC AGATTCGGAT TCTCAACCCC TTCTTCATGG 700

[402] AGATTGCAGC TGACAAACTG CTGAGCCTGC CAATGCAACA GCCACGGAAG 750

[403] ATTAAGCAGA AGCCCACCAC GGGCCTGTTG GCCATCACGC TGGCCCTCCA 800

[404] CCTCTGTGAC TTGGTGCACA TTGCCGGCTT TGGCTACCCA GACGCCTACA 850

[405] ACAAGAAGCA GACCATTCAC TACTATGAGC AGATCACGCT CAAGTCCATG 900

[406] GCGGGGTCAG GCCATAATGT CTCCCAAGAG GCCCTGGCCA TTAAGCGGAT 950

[407] GCTGGAGATG GGAGCTATCA AGAACCTCAC GTCCTTC      987

[408]

[409] SEQ ID NO: 2. EPO-Fc (Efepoein alfa) (amino acid sequence)

[410] MGVHECPAWL WLLLSLLSLP LGLPVLGAPP RLICDSRVLE RYLLEAKEAE 50

[411] NITTGCAEHC SLNENITVPD TKVNFYAWKR MEVGQQAVEV WQGLALLSEA 100

[412] VLRGQALLVN SSQPWEPLQL HVDKAVSGLR SLTTLLRALG AQKEAISPPD 150

[413] AASAAPLRTI TADTFRKLFR VYSNFLRGKL KLYTGEACRT GDRRNTGRGG 200

[414] EEKKKEKEKE EQEERETKTP ECPSHTQPLG VFLFPPKPKD TLMISRTPEV 250

[415] TCVVVDVSQE DPEVQFNWYV DGVEVHNAKT KPREEQFNST YRVVSVLTVL 300

[416] HQDWLNGKEY KCKVSNKGLP SSIEKTISKA KGQPREPQVY TLPPSQEEMT 350

[417] KNQVSLTCLV KGFYPSDIAV EWESNGQPEN NYKTTPPVLD SDGSFFLYSR 400

[418] LTVDKSRWQE GNVFSCSVMH EALHNHYTQK SLSLSLGK

[419]

[420] SEQ ID NO: 3. a -2,3-sialyltransferase - HA (amino acid sequence)

[421] MVSKSRWKLL AMLALVLVVM VWYSISREDS FYFPIPEKKE PCLQGEAESK 50

[422] ASKLFGNYSR DQPIFLRLED YFWVKTPSAY ELPYGTKGSE DLLLRVLAIT 100

[423] SSSIPKNIQS LRCRRCVVVG NGHRLRNSSL GDAINKYDVV IRLNNAPVAG 150

[424] YEGDVGSKTT MRLFYPESAH FDPKVENNPD TLLVLVAFKA MDFHWIETIL 200

[425] SDKKRVRKGF WRQPPLIWDV NPKQIRILNP FFMEIAADKL LSLPMQQPRK 250

[426] IKQKPTTGLL AITLALHLCD LVHIAGFGYP DAYNKKQTIH YYEQITLKSM 300

[427] AGSGHNVSQE ALAIKRMLEM GAIKNLTSFY PYDVPDYAGY PL

[428]

[429] Deposit Information

[430] Depositary Institution: Korean Cell Line Research Foundation (KCLRF)

[431] Address: 101 Daehak-ro, Jongno-gu, Seoul 03080

[432] Date of Deposit: July 11, 2023

[433] Accession No.: KCLRF-BP-00525 b r I b            (kclb, une IV V 03129. AISAI CHStS 103        WB®! 7§ TEL : +-82-2-3668-7915, FAX : +82-2-742-0021, hltp*- / / cellbank.snu.ac.kr, kclb@kclh.kr 3^1 Korean Cell line Bank; Korean Cell Une Research Foundation, International Microorganism Depository Authority approved by WIPO Cancer Research Institute. Seoul National University College 0! Medicine. 103 Daehak-ro. Chongnogu, Seoul J10-799, KOREA BUDAPEST TREATY ON THE INTERNATIONAL RECOGNITION OF THE DEPOSIT OF MICROORGANISMS FOR THE PURPOSE OF PATENT PROCEDURE INTERNATIONAL FORM RECEPTION IN THE CASE OF AN ORIGINAL DEPOSIT issued pursuant to Rule 7.1 To: Genexine, Inc. 172 Magokjungang.ro, Genexine Progen Innovation Park, Gangseo-gu, Seoul, Korea I. IDENTIFICATION OF THE MICROORGANISM Identification reference given by the DEPOSITOR : CHO DG44 a-2,3 ST Accession number given by the INTERNATIONAL DEPOSITARY AUTHORITY: KCLRF-BP-00S2S H. SCIENTIFIC DESCRIPTION AND / OR PROPOSED TAXONOMIC DESIGNATION The microorganism identified under I above was accompanied by : [x] A scientific description [x] A proposed taxonomic designation (Mark with a cross where applicable) UI. RECEIPT AND ACCEPTANCE This International Depositary Authority accepts the microorganism identified under 1 above, which was received by it on July 11, 2023 IV. INTERNATIONAL DEPOSIT ARV AUTHORITY Name : Director Korean Cell Line Research Foundation Address : Cancer Research Institute Seoul National University College of Medicine 101, Daehak-ro, Jongno-gu, Seoul, 03080, Korea Signature(s) Date : 2023. 07. 21 Form BP / 4 (KCLRF Form 17)                                                                     Page sole

Claims

[CLAIMS]

1. An EPO-Fc fusion protein producing cell line transfected with an a-2,3-sialyltransferase gene.

2. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the a-2,3-sialyltransferase gene comprises the nucleotide sequence of SEQ ID NO: 1.

3. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the EPO-Fc fusion protein binds and activates the endogenous EPO receptor at about the same level or better than the EPO BRP.

4. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the EPO-Fc fusion protein induces the proliferation and differentiation of erythroid progenitor cells into reticulocytes at about the same level or better than EPO BRP.

5. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the EPO-Fc fusion protein induces the proliferation of EPO-dependent cell line such as UT-7 at about the same level or better than EPO BRP.

6. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the Fc lacks the immune effector activities comprising, but not limited to, ADCC (antibody-dependent cellular cytotoxicity), ADCP (antibody-dependent cellular phagocytosis), and CDC (complementdependent cytotoxicity) activities that harm the EPO receptor-expressing cells.

7. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the Fc extends the half-life of the Fc fusion protein by binding FcRn only at pH range that is representative of the endosomal compartment in mammalian cells and dissociating from FcRn at neutral pH, thus avoiding the cellular degradation.

8. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the Fc extends the half-life of the Fc fusion protein by increasing its molecular weight to avoid renal clearance.

9. The EPO-Fc fusion protein producing cell line according to claim 1, wherein the EPO-Fc fusion protein comprises an EPO sequence and hybrid Fc sequence corresponding to the amino acid residues 1-193 and amino acid residues 194-438 of amino acid sequence SEQ ID NO: 2, respectively.[Claim10]The EPO-Fc fusion protein producing cell line according to claim 1, wherein the Fc comprises 30 amino acid residues of the hinge region and 8 amino acid residues of CH2 domain N-terminus of human IgD (amino acid residues 194-231 of SEQ ID NO: 2), and 100 amino acid residues of CH2domain C-terminus and 107 amino acid residues of CH3 domain of human IgG4 (amino acid residues 232-438 of SEQ ID NO: 2).

11. A method for producing an EPO-Fc fusion protein, comprising: (a) fed-batch culturing an EPO-Fc fusion protein producing cell line according to any one of claims 1 to 10; and (b) isolating and purifying an EPO-Fc fusion protein produced in the cell line.

12. The method according to claim 11, wherein the EPO of the EPO-Fc fusion protein further comprises modification for additional glycosylation.

13. The method according to claim 11, wherein the EPO-Fc fusion protein further comprises PEG.

14. The method according to claim 11, wherein the fed-batch culture is performed in the presence of N-acetyl-D-mannosamine (ManNAc) and galactose.

15. The method according to claim 14, wherein the N-acetyl-D-mannosamine is present at a concentration of 1 mM to 20 mM, and the galactose is present at a concentration of 1 mM to 40 mM.

16. The method according to claim 11, wherein the fed-batch culture is performed in temperature of 34°C to 38°C and pH of 6.7 to 7.5.

17. The method according to claim 11, wherein the fed-batch culture comprises culturing the cell line for 9 to 11 days.

18. The method according to claim 16, wherein the fed-batch culture is cultured at a temperature lowered by 2 to 5°C on day 3 or 4 days after inoculation of the cell line

19. The method according to claim 11, wherein the EPO-Fc fusion protein isolation and purification is performed at room temperature.

20. The method according to claim 19, wherein the EPO-Fc fusion protein isolation and purification, except for viral inactivation, is performed at room temperature.

21. The method according to claim 11, wherein EPO-Fc fusion protein isolation and purification is performed at 17°C to 27°C.

22. The method according to claim 11, wherein the EPO-Fc fusion protein has specific quality attributes comprising at least one of the following: C-terminal lysine content, N- and O-glycan profiles, sialic acid content, and pI range.[Claim 231The method according to claim 22, wherein the C-terminal lysine content of the EPO-Fc fusion protein is about 1% to about 10%.[Claim 241The method according to claim 22, wherein the C-terminal lysine content of the EPO-Fc fusion protein is about 2% to about 9%.[Claim 251The method according to claim 22, wherein the C-terminal lysine content of the EPO-Fc fusion protein is about 3% to about 7%.[Claim 261The method according to claim 22, wherein the sialic acid content of the EPO-Fc fusion protein is about 17 mol / mol to about 27 mol / mol.[Claim 271The method according to claim 22, wherein the sialic acid content of the EPO-Fc fusion protein is about 20 mol / mol to about 25 mol / mol.[Claim 281The method according to claim 22, wherein the pI range of the EPO-Fc fusion protein is about 4.0 to about 6.0.[Claim 291The method according to claim 22, wherein the pI range of the EPO-Fc fusion protein is about 4.5 to about 5.3.[Claim 301The method according to claim 22, wherein the N- and O-glycan profiles comprise the nonglycosylated content on position N24 about 0-20%, on position N38 about 0-5%, on position N83 about 10-17%, on position N261 about 0-10%, and on position S126 about 50-70%.[Claim 311The method according to claim 22, wherein the N- and O-glycan profiles comprise the glycan containing sialic acid on position N24 about 50-90%, on position N38 about 55-95%, on position N83 about 40-80%, on position N261 about 0-30%, and on position S126 about 15-40%.[Claim 321The method according to claim 11, wherein the EPO-Fc fusion protein isolation and purification is performed using affinity chromatography or ion exchange chromatography.[Claim 331The method according to claim 32, wherein the ion exchange chromatography comprises an anion exchange chromatography process using equilibration buffers with pH of about 6.00 to about 6.50 to facilitate the binding of acidic species of EPO-Fc fusion protein to the resin.[Claim 341The method according to claim 32, wherein the ion exchange chromatography comprises an anion exchange chromatography process using equilibration buffers with pH of about 6.10 to about 6.30 to facilitate the binding of acidic species of EPO-Fc fusion protein to the resin.[Claim 351The method according to claim 32, wherein the ion exchange chromatography comprises an anion exchange chromatography process using equilibration buffer with conductivity of about 6.00 to about 10.50 mS / cm, and elution buffer with conductivity of about 11.50 to about 16.00 mS / cm to facilitate the isolation of EPO-Fc fusion protein with the quality attributes thereof.[Claim 361The method according to claim 32, wherein the ion exchange chromatography comprises an anion exchange chromatography process using equilibration buffer with conductivity of about 8.00 to about 10.00 mS / cm, and elution buffer with conductivity of about 12.00 to about 15.00 mS / cm to facilitate the isolation of EPO-Fc fusion protein with the quality attributes thereof.

Images