SG192303A1 - Cho-gmt recombinant protein expression - Google Patents

Abstract

57CHO-gmt Recombinant Protein ExpressionAbstract 5 A glycoprotein that is expressed by a mammaan cell, wherein the mammalian cellcomprises:a mutated GnT I gene and a mutated CST gene;a mutated GnT I gene and a mutated UGT gene; ora mutated CST gene; or 10 d) a mutated UGT gene; ore) a mutated GFT gene and a mutated CST gene.Figure 1

Description

CHO-gmt Recombinant Protein Expression

Field of the Invention

The invention relates to the fields of biotechnology and molecular biology. The invention in particular relates to mammalian cell lines and their use in protein expression. The invention further relates to the use of such proteins in medicine.

Background to the invention

Mucins are a family of heavily O-glycosylated secreted or membrane-associated large proteins produced by simple and glandular epithelia. A common feature shared by all mucin glycoproteins is the variable number of tandem repeat (VNTR) located in the central part of the molecule. These tandem repeats (TRs) are rich in serine, threonine and proline (for reviews see Thomton et al., 2008; Hattrup and Gendler, 2008). Human mucin 1 (MUC1) was the first mucin to be cloned and the best studied mucin fo date. The MUC1 gene encodes a type | transmembrane protein with a single transmembrane domain and a C-terminal cytosolic tail. During maturation,

MUC1 is cleaved into the N-terminal subunit and the C-terminal subunit. The C- terminal subunit contains the C-terminal 158-amino acids, which forms an extracellular domain, the transmembrane domain and a cytoplasmic tail. The size of the N-terminal subunit varies depending on the number of the TRs. After cleavage, the N-terminal subunit remains associated with the C-terminal subunit by interacting with the extraceliutar domain of the C-terminal subunit. The N-terminal 23-amino acids of MUC1 function as a signal peptide to target MUC1 to the cell surface and are removed during the process. The number of VNTR of the N-terminal subunit can vary between 25~100. Each TR contains identical twenty amino acids (HGVTSAPDTRPAPGSTAPPA). The two S and three T residues in each repeat represent the five potential O-glycosylation sites. In addition to O-giycosylation, there are 5 potential N-glycosylation sites in the MUC1 molecule, 4 are located at the C- terminal end of the N-terminal subunit and one is located at the extraceliular domain the of the C-terminal subunit T hornton et al., 2008; Hattrup and Gendler, 2008).

Cell surface carbohydrates are characteristic of different stages of normal development and differentiation; distinct carbohydrates are expressed in tissue- and cell-specific patterns during those processes. The O-glycosylation pattern change of

MUC1 is one of the best examples to illustrate this fact. In normal epithelial cells,

MUC1 is expressed at low levels and restricted to the apical membranes of the epithelium to hydrate, protect and lubricate the surface. The TRs of MUC1 are O- glycosylated by highly branched complex carbohydrates. In comparison, MUC1 is highly overexpressed in adenocarcinomas, such as breast, ovarian and pancreatic cancers. In these cells, MUC1 is expressed over the entire cell surface and is no longer restricted to the apical surface. In addition, MUC1 is aberrantly glycosylated

Co (for reviews see Taylor-Papadimitriou, 1999; Hollingsworth and Swanson, 2004: Tarp and Clausen, 2008; Kufe, 2009; Rachagani et al., 2009; Bafha et al., 2010). Instead of the highly branched complex carbohydrates, MUC1 expressed in cancer cells carries short immature O-glycans. Some of these short O-glycans are the Tn antigen (GalNAca-O-Ser/Thr), Sialyl-Tn (STn) antigen (NeuAca2-6GalNAca-0O-Ser/Thr) and

T antigen (GalB1-3GalNAca-O-Ser/Thr) (Hull et al., 1989; Lioyd et al., 1996).

Therefore, these short O-glycans are referred to as tumor-associated antigens and have been widely used as therapeutic targets and diagnostic markers (Vlad and Finn, 2004; Dube and Bertozzi, 2005; Brockhausen, 20086; Potapenko et al., 2010). In fact, the Tn, STn and T antigens were recognized as tumor-associated antigens jong before the mucins were cloned (Springer, 1984; Springer, 1997). Other short O- : glycans attached to the overexpressed MUC1 are sialyl-T (ST) antigen (NeuAca2- 3Galf1-3GalNAca-O-Ser/Thr) and disialyl-T (diST) antigen (NeuAco2-3GalB1- 3[NeuAca2-6]GalNAca-O-Ser/Thr). They are not considered as tumor-associated antigens as they are also found in normal celis. As the precursors of more complex

O-glycans, T and Tn epitopes are masked with other sugars in healthy adults.

Therefore, they are un-detectable in healthy and benign-diseased tissues except in early embryonic stages. But they are expressed in about 90% of all carcinomas.

Although T and Tn antigens are absent in normal adult tissues, all humans have anti-

T and anti-Tn antibodies in our blood. This is likely the result of prior exposure to the commonly found Enterobacteriaceae that contain T and Tn epitopes. Chicks raised under germ-free conditions do not have anti-T and anti-Tn antibodies whereas chicks from the same hatch raised under ordinary conditions have these antibodies. Human infants produced high. titers of anti-T and anti-Tn antibodies after being exposed to certain bacteria. Thus, T and Tn epitopes are “tumor-associated antigens” not “tumor-specific antigens”. Most anti-T antibodies in humans are igh. in fact, anti-T

IgM constitutes 7 to 14% of total IgM (Springer, 1984; Springer, 1997). However, ha tgG antibodies to aberrantly O-glycosylated VNTR derived from MUC 1 was detected in healthy individuals. igG antibodies to Tn-VNTR were only detected in patients vaccinated with Tn-KLH or in sera from breast, ovarian, and prostate cancer patients ~ (Wandall et al., 2010).

in contrast to the anti-T and anti-Tn antibodies which all of us have due to the humoral immune response to Enterobacteriaceae, there is no pre-existing cellular immune response to T and Tn antigens. Delayed-type hypersensitivity (DTH) is an inflammatory response that develops 24 to 48 hours after exposure to an antigen that the immune system recognizes as a foreign antigen. When the T epitope is used as the antigen in the assay, the DTH reaction to T antigen (DTHR-T) can be used to detect the cellular immune response to T antigen. Healthy individuals do not have a positive DTHR-T. Interestingly, DTHR-T was positive in most of the carcinoma patients. Carcinoma patients have both humoral and cellular immune responseto T antigen whereas healthy people have only the humoral immune repose, but no cellular immune response, to T antigen (Springer, 1997). in a recent report, the presence of auto-antibodies to aberrantly glycosylated MUCH in early stage breast cancer has been associated with a better prognosis (Blixt et al, 2011). In that study, sera were collected in the 1970s and 1980s from a large cohort of breast cancer patients, patients with benign breast disease and ‘healthy controls. In the clinical follow-up investigation, the presence and level of anti- aberrantiy glycosylated MUC1 antibodies was found significantly higher in the sera from cancer patients compared with the controls. High levels of a subset of autoantibodies to the coreBMUC1 (GIcNAcB1-3GalNAc-MUC1) and STn MUC1 (NeuAco2,6GalNAc-

MUC1) glycoforms were significantly associated with reduced incidence and increased time to metastasis. Therefore, the association of strong antibody response with reduced rate and delay in metastases suggests that autoantibodies can affect disease progression (Blixt et al., 2011).

Abemrantly O-glycosylated MUCT has been a target for cancer immunotherapy in nurnerous clinical trials

MUCH1 is highly immunogenic as most monoclonal antibodies raised against human cancer cell lines react with the TR region of MUC1 (Xing et al., 2001; Tang et al., 2008). The aberrant O-glycosylation of MUC1 results in the generation of cancer- associated antigens in two ways: (1) by exposure of protein backbone epitopes that are normally masked by the highly branched complex carbohydrates; (2) by changing the structure of the carbohydrate side chains attached to MUC1 to the well-known tumor-associated O-glycans, such as T, Tn and STn antigens. The loss of polarity of the epithelial cancer cells allows the deposition of these antigens on the whole surface of the cell which makes them more accessible to the immune system. In the last decade, numerous clinical trails have been carried out to test the anti-cancer : properties of many MUC1-based vaccines (Richardson and Macmillan, 2008; Tang et al., 2008; Beatson et al., 2010). A few well-studied examples are discussed here.

STn-KLH (Theratope): STn is chemically linked to keyhole limpet haemocyanin (KLH)

The STn antigen is expressed in about 30% of breast cancers (Julien and Delannoy, 2003; Miles and Papazisis (2003). As the expression of STn is highly restricted in normal tissues (Julien and Delannoy, 2003), this antigen has been considered a target for the development of anti-cancer vaccine (Tang et al., 2008). A synthetic

STn-keyhole limpet haemocyanin (KLH) vaccine (Theratope), which consists of 3000 mol of the STn disaccharide conjugated to 1 mol of KLH, has been designed by the biotech company Biomira of Canada (now Oncothyreon, Seattle, USA; Ragupathi et al., 1999). Preclinical studies in mice showed that immunization with Theratope can induce STn-specific IgG. A phase Il clinical study in breast cancer patients showed that potent STn-specific humoral responses can be induced following immunization with STn-KLH. However, in the follow-up phase Ill trial no benefit for Theratope- immunized patients compared to control group was shown (Holmberg et al., 2000;

Finke et al., 2007). STn epitope alone without the MUC1 VNTR peptide backbone may not be specific enough to elicit strong immune response against aberrantly glycosylated MUC1 expressed on cancer cells. Nevertheless, studies using STn-KLH as a vaccine in animal models or in breast cancer patients are still on going in several labs (Gilewski et al., 2007; Julien et al., 2009).

BLP25 liposome vaccine (Stimuvax): Peptides derived from the TR of MUC1 mixed with lipids

Another type of vaccines was based on unglycosylated peptides derived from MUC1

VNTR. In the early days, several groups have immunized mice with various synthetic peptides of different sizes derived from the VNTR of MUC1. While some peptides triggered humoral response others elicited T cell responses. A few triggered both humoral and T cell responses (Ding et al., 1993; Zhang et al., 1996; Soares et al., 2001). Recently, a vaccine that contains a 25-mer synthetic peptide based on the

VNTR of MUCH in a liposomal formulation, or BPL25 liposome vaccine (Stimuvax) has been developed and studied in clinical trials by Oncothyreon Inc. The BLP25 lipopeptide consists of a 25-amino acid sequence

(STAPPAHGVTSAPDTRPAPGSTAPP) that provides MUC1 specificity. If contains a palmitoyl lysine residue at the carboxy terminal to enhance the incorporation of the lipopeptide into the liposome particle. The vaccine consists of BLP25 lipopeptide, three lipids (cholesterol, dimyristoyl phosphatidylglycerol, and dipalmitoyl 5 phosphatidylcholine) and immunoadjuvant monophosphoryl lipid A. The vaccine was granted fast-track status in September 2004 by the US FDA. At the end of 2008,

Merck obtained the exclusive worldwide licensing rights from Oncothyreon Inc. for 13 million USD. Merck has been conducting Phase Ill trials of Stimuvax to treat non- small cell lung cancer and breast cancer (Goldman and DeFrancesco, 2009).

Mannosylated MUC1 peptides

Antigen mannosytation is an effective approach to potentiate antigen immunogenicity, due to the enhanced antigen uptake and presentation by antigen presenting cells (APCs), particularly dendritic cells (DCs). Apostolopoulos and colleagues have been interested in targeting MUC1 to mannose-binding receptors on

DCs. A peptide (five VNTRs) of MUC1 has been conjugated fo yeast mannan to increase the uptake by DCs. They have shown that T1-type response which induces high cytotoxic T lymphocytes (CTLs) can be induced by conjugation of the antigen to : the carbohydrate polymer mannan under oxidizing conditions (Apostolopoulos et al., 1995). The MUCH peptide that was fused to GST conjugated to oxidized mannan was studied in several clinical trials (Loveland et al., 2008; Apostolopoulos et al., 2006; Tang et al., 2008). However, conventional chemical methods used to mannosylate antigens are not consistently reliable, due to the possibility of : irreversible modification of immunodominant epitopes. Expression of soluble recombinant mannosylated proteins in yeast requires the introduction of glycosylation sites on the protein, as well as complicated characterization or optimization procedures. These issues need to be resolved in order to produce reliable mannosylated antigens.

Tn-glycosylated MUC1 glycopeptide

The main problem of the above mentioned strategies is that the tumor-associated glycans and the MUC1-derived peptides are separated. When using unglycosylated

MUC1-derived peptide as vaccines, the immune response are directed against the unglycosylated MUC1 only and not to the aberrantly glycosylated MUC1 which is expressed on the cancer cells. Attempts to generate strong humoral immunity to

MUC1 by immunization with these unglycosylated peptides have generally failed partly because of {olerance. To overcome this problem, MUC1 TR glycopeptides with tumor-associated antigens were chemoenzymatically synthesized using a panel of recombinant human glycosyliransferases (Sgrensen et al, 2006). MUC1 glycopeptides with different densities of Tn and STn glycoforms conjugated to KLH elicited the strongest antibody response reacting with MUC1 expressed in breast cancer cell lines. The elicited humoral immune response showed strong specificity for cancer cells suggesting that the glycopeptide design hoids promise as a cancer vaccine. The elicited immune responses were directed against combined glycopeptide epitopes, and both peptide sequence and carbohydrate structures were important for the antigen.

Recombinant MUCH produced by wild type CHO-K1 cells

Recombinant MUC1 with 16 TRs has been produced and reported (Backstrom et al. 2003; Link et al. 2004). As this MUC1 was produced in wild type CHO-K1 cells, the ~ 15 O-glycans are ST and diST which are different from the tumor-associated antigen. i

Therefore, recombinant MUC1 produced by wild type CHO cells is chemically different from the MUC1 expressed on cancer cell surface.

Limitations of above-described studies

Each of the studies described above has its limitations. STn-KLH (Theratope) contains only the O-glycan but no MUC1 peptide. On the contrary, BL.P25 liposome vaccine (Stimuvax) contains only the peptide and no glycans. MUC1 peptide conjugated to mannan seems to enhance antigen uptake by DCs. In vivo targeting of

C-type lectin receptors on DCs is an effective strategy to increase efficacy of vaccines. However, this method has the same problem because it lacks the tumor- associated O-glycans attached to the MUC1 peptide. It has been shown that when humoral responses were elicited by unglycosyiaied MUC1 peptide, the antibodies were unable to recognize glycosylated MUC1 (von Mensdorff-Pouilly et al., 2000).

The GSTA motif of the MUC1 20-amino acid tandem repeat has been shown to be a highly immunodominant epitope only when presented with immature short glycans (Tn/STn) (Tarp et al., 2007). Vlad et al. (2002) have shown that the O-glycans are not removed during processing of tumor antigen MUC1 glycopeptides by DCs. DCs endocytose glycopeptides, process them into smaller peptides and present them on

MHC class Il molecules without removing the carbohydrates. Resulting glycopeptide in the binding grove of MHC I! triggers glycopeptide-specific CD4 T cells that can also respond to aberrantly glycosylated MUC1. Furthermore, O-glycosylation also affects intracellular processing of MUC1by DC. One of the intracellular cleavage sites in MUCH is between the Thr-Ser peptide bond within the 20-amino acid tandem repeat. However, if either amino acid residue is glycosylated, cleavage at this site is prevented (Hanisch et al., 2003). All these data suggest that glycosylated and : 5 unglycosylated MUC1 peptides represent different antigens and therefore MUC1 peptide used in the vaccines shouid be properly glycosylated. Chemoenzymatically synthesized Tn/STn MUC1 glycopeptides represent a step forward as they were able to elicit cancer-specific anti-MUC1 antibody responses (Serensen et al., 2006).

However, it is difficult to reach identical glycosylation patterns from batch to batch.

New approaches are needed to produce the same vaccines consistently that closely mimic the structure of MUC1, both in peptide sequence and the pattern of glycosylation, in a more cost-effective manner.

Mannose-binding C-type lectins on dendritic cells (DCs) and macrophages can dramatically enhance antigen uptake and processing :

DCs are the most potent antigen-presenting cells (APCs) and have a central role in directing the adaptive immune response. DCs are positioned at the external and internal surfaces of the body to specifically bind antigens and present them to lymphocytes in lymphoid organs. The uptake receptors on DCs can dramatically increase the efficiency for antigen capture and processing. Many of these uptake receptors are mannose-specific C-type lectins, such as langerin (CD207), DC-SIGN (CD209) and mannose receptor {CD206) (McGreal et al., 2005; Gazi et al., 2009;

Keppler-Ross et al., 2010). Another type of APCs, macrophages, also uses mannose-specific C-type lectins as uptake receptors (Gazi et al., 2009). Therefore, mannosylation of vaccines has been an effective method to specifically target the antigens to DCs and macrophages for enhanced vaccine efficacy. As discussed above, yeast mannan has been conjugated to MUC1 peptides for enhanced immunogenicity. In addition fo mannan, mannose-terminated N-glycans on recombinant proteins has also been shown fo be recognized by mannose-specific C- type lectins and endocytosed by macrophages. One successful example comes from the treatment of the patients with Gaucher's disease with recombinant glucocerebrosidase that breaks down GlcCer within the macrophages. The recombinant glucocerebrosidase produced by wild-type CHO cells can only be captured efficiently by the macrophages when treated with sequential exoglycosidases until the mannose residues of the N-gtycan (MansGleNAG,) are exposed (Friedman et al., 1999). Recombinant glucocerebrosidase produced by a

CHO glycosylation mutant that has a dysfunctional GnT | gene contains mannose- terminated M-glycans (MansGicNAc,} (Goh et al., 2010). This product can also be : specifically picked up by the C-type lectins on the surface of macrophages (Van

Patten et al., 2007). We aim to produce MUC1 in a GnT | mutant CHO cell line that will result

MUC1 has been a target for cancer immunotherapy in several clinical trials.

Theratope was developed by Biomira, Inc. as treatment for breast cancer. It is a synthetic vaccine in which STn disaccharides are chemically linked to a protein carrier, keyhole limpet haemocyanin (KLH). Theratope failed in phase Ili trials because STn epitope alone without the peptide backbone of the TRs was not specific enough to elicit strong immune response against aberrantly glycosylated MUC1 on cancer cells. BLP25 liposome vaccine (Stimuvax) contains unglycosylated peptides derived from TRs mixed with three lipids that act as adjuvants. Currently Merck is conducting a phase Ill trial on Stimuvax. A potential problem with Stimuvax is that unglycosylated MUC1 peptides are chemically different from glycosylated MUC1 expressed on cancer cells. The antibodies that bind the unglycosylated TRs will not be able to recognize the tumor-associated MUC1 with high affinity. Mannosylated

MUC1 pepiides have also been investigated as vaccines.

A peptide containing five TRs has been conjugated to yeast mannan to increase the ~ uptake by dendritic cells (DCs). Mannose-binding C-type lectins on DCs and macrophages can dramatically enhance antigen uptake and processing. However, ‘this MUC1 peptide also lacks the tumor-associated O-glycans. Gene therapy - attempts have also been made by delivering the MUC1 cDNA into the patients. As the transgene will not be expressed in the cancer cells of the patients, the O-glycans on the MUC1 produced will be same as that of the normal cells. Therefore, this

MUC1 vaccine will very likely be tolerated by the immune system. In summary, the problem with all these strategies is the failure of glycosylating the MUC1-derived peptides with tumor-associated short O-glycans. Chemoenzymatically synthesized multimeric Tn/STn MUCH glycopeptides has elicited cancer-specific anti-MUC1 antibody responses and overridden tolerance, demonstrating the importance of O- glycans. However, chemical synthesis of Tn/STn MUC1 is very expensive and difficult fo produce consistent products in large quantities.

Summary of the Invention

We have isolated two CHO glycosylated mutants: CHO-gmt1 (formerly MAR-11) and

CHO-gmt2 (formerly MAR-1). CHO-gmt1 tacks functional CMP-sialic acid transporter and therefore the only O-glycan it can synthesize is the T antigen. CHO-gmt2 has a dysfunctional UDP-galactose transporter. Consequently, the O-glycans that CHO- gmt2 cells can produce are the Tn and STn antigens. We have also shown that most of CHO cells, if not all, that survive the cytotoxic RCA-I treatment have a mutated N- acetylglucosaminyltransferase | (GnT I} gene. As a result, the N-glycans produced by these mutants have the structure of MansGlcNAc,, which can be specifically recognized by the mannose-binding C-type lectins on the surface of antigen presenting cells such as dendritic cells and macrophages. We have isolated cells that : have mutated GnT | gene from CHO-gmt1 and CHO-gmt2 cells using RCA-l. The newly isolated mutants will produce proteins that are O-glycosylated with T or Tn and

STn antigens and N-glycosylated with the MansGicNAG, glycan. MUC1 produced in these double mutant lines possess enormous potentials as anti-cancer vaccines due to the presence of tumor-associated antigens attached to the VNTR region as well as mannose-terminated N-glycans which will allow for enhanced uptake by dendritic cells and macrophages. These recombinant N-terminal subunits of MUC1 will be investigated as anti-cancer vaccines in mouse models. Their efficacy in eliciting humoral and cellular immune responses will be investigated. Following positive results from the mouse models the next step is to produce the same recombinant proteins under GMP conditions for analysis in human cytotoxicity fest.

The present invention relates fo a glycoprotein that is expressed by a mammalian cell which has a mutated gene. The mammalian cell may have one or more of a mutated

GnT| gene, a mutated CST gene, a mutated UGT gene and a mutated GFT gene.

Preferably, the mammalian cell comprises: a) a mutated GnT | gene and a mutated CST gene; b) a mutated GnT | gene and a mutated UGT gene: or

Cc) a mutated CST gene; or d) a mutated UGT gene; or e} a mutated GFT gene and a mutated CST gene.

The glycoprotein may comprise or consist of a Mucin, such as a Mucin which is aberrantly glycosylated in cancer cells. The Mucin may be a transmembrane Mucin.

In particular, the recombinant protein may comprise, or consist of Mucin 1 (MUCT), or fragments thereof. Fragments of MUC1 may include the N terminal subunit of MUCH, and may include one or more, for example 1-10, 3-10, 11-20, 21-30, 3140, 41-50, 51-60, 61-70, 71-80, 81-90, 91-100, 100 or more, 150 or more, or 200 or more fandem repeats.

The glycoprotein may be an antibody, antibody variant or antibody binding fragment.

In such cases the mammalian cell preferably comprises a mutated GFT gene and a mutated CST gene.

The glycoproteins of the invention may have altered patterns of glycosylation. The altered glycosylation may be different to the glycosylation pattern associated with expression of the protein in a non-mutated mammalian cell. For example, a glycoprotein expressed in CHO cell with a mutation may have an altered pattern of glycosylation as compared to a glycoprotein expressed in a CHO cell which does not have that mutation.

The glycoprotein may have an altered pattern of glycosylation of N-glycans. The glycoprotein may comprise a mannose terminated glycan structure. The glycoprotein may comprise multiple mannose terminated glycan structures.

Additionally or alternatively, the glycoprotein may have an altered pattern of glycosylation of O-glycans. For example, the glycoprotein may have an altered pattern of sialylation and/or galactosylation. The glycoprotein may partially or entirely lack sialic acid, and/or may partially or entirely lack galactose.

In some embodiments the glycoprotein may comprise MUC1 or a fragment thereof that has T, Tn or Stn antigen giycans. ~The mammalian cell may be a CHO cell, BHK cell, Vero cell, 293 cell, NSO cell, 3T3 cell, COS-7 cell or PER C8 cell.

The invention also provides antibodies raised against the glycoproteins according to the invention.

: The invention also provides glycoproteins (including antibodies), and antibodies there to, for use in medicine. The glycoprotein or antibody may be used in the treatment of: cancer. The cancer may be an adenocarcinoma, such as breast cancer. The treatment may comprise administering to the patient a glycoprotein according to the invention, or an antibody there to. The treatment may comprise exposing cells (such : as dendritic cells) isolated from the patient in need of treatment to a glycoprotein according to the invention, or an antibody thereto.

The invention further provides methods of medical treatment. The method may comprise administering a glycoprotein according to the invention, or an antibody thereto to a patient in need of such treatment. The patient may be suffering from cancer. The cancer may be an adenocarcinoma, such as breast cancer. ’ Alternatively, the method may comprise obtaining cells from the patient and exposing * 15 those cells to a glycoprotein according to the invention, or an antibody thereto. The cells may be immune cells, such as dendritic cells.

The invention further provides glycoproteins, or antibodies thereto for use in the manufacture of a medicament for the treatment of cancer. The cancer may be an adenocarcinoma, such as breast cancer.

The invention further provides a mammalian cell. The mammalian cell may have one or more of a mutated GnT | gene, a mutated CST gene, a mutated UGT gene and a mutated GFT gene. :

Preferably, the mammalian cell comprises: a) a mutated GnT | gene and a mutated CST gene; b) a mutated GnT | gene and a mutated UGT gene; or c) a mutated UGT gene; or . d) a mutated GFT gene and a mutated CST gene.

The mammalian cell may be a CHO cell, BHK cell, Vero cell, 293 cell, NSO cell, 3T3 cell, COS-7 cell or PER C8 cell.

The mammalian cell may express a recombinant protein. The glycoprotein may comprise or consist of a Mucin, such as a Mucin which is aberrantly glycosylated in cancer cells. The Mucin may be a transmembrane Mucin. In particular, the recombinant protein may comprise, or consist of Mucin 1 (MUCH), or fragments thereof. Fragments of MUC1 may include the N terminal subunit of MUC1, and may include one or more, for example 3-10, or 5-6 tandem repeats.

The invention also provides a culture of cells according to the invention.

The invention further provides a method of expressing a recombinant glycoprotein. in the method, the glycoprotein is expressed in a mammalian cell according to the invention. The glycoprotein may be encoded in the genome of the mammalian cell, or may be expressed from a vector in the cell.

Description of Preferred Embodiments

The invention includes the combination of the aspects and preferred features described except where such a combination is clearly impermissible or expressly avoided.

The section headings used herein are for organizational purposes only and are not to be construed as limiting the subject matter described.

Aspects and embodiments of the present invention will now be illustrated, by way of example, with reference to the accompanying figures. Further aspects and embodiments will be apparent to those skilled in the art. All documents mentioned in + this text are incorporated herein by reference.

GnT I

The mammalian cells according to the invention may have a mutation in N- acetylglucoaminyltransferase |. For example modified cells may have a mutation in a gene that is, or is homologous fo GenBank: AF343963 GI:14388980.

We demonstrate that recombinant proteins of interest may be produced in cells lacking a functional GnT | gene and may comprise mannose terminated glycan structures. These results indicate that such cells have the potential to become a host cell line for producing proteins, comprising mannose-terminated N-glycans, including glycoprotein drugs and vaccines.

The mutation in GnT | in the mammalian cells according to the invention results in a loss of GnT | function. The GnT [ function may be reduced or entirely absent.

Function of the GnT | gene in the cells according to the invention may be reduced as compared with the function of wild type GnT I, such as wild type human GnT I. In some cases the function is reduced relative to GnT | found in wild type CHO cells.

Expression of glycoproteins in such cells may be beneficial for expression of polypeptides which are to be targeted to celis with lectin receptors, for example, macrophages and dendritic cells. GnT I deficient cells made according to the methods described herein may be used for the production of recombinant proteins which have mannose-terminated glycan structures and are hence able to be taken up by macrophages. i5

This is in contrast to the normal complex-type N-glycans that also contain sialic acid and galactose residues using expression from other cells.

The use of GnT | deficient cells as host cells has a number of advantages. For example, as the recombinant expressed proteins of interest have mannose- : terminated glycan structures, there is no need for enzymatic treatment to expose these.

Mammalian cells with a mutated GnT | have been previously described in © 25 W02010/033085 and WO2011/119115.

The mutant GnT | cells according to the invention may have been identified, at least in part, by selection with RCA-I (Ricinus communis agglutinin 1). The identification or selection protocol may be as set out in WO2010/033085 or WO2011/119115.

Alternatively, the mutant gene may have been introduced into the mammalian cell by recombinant gene technology. The mammalian cell may be derived from, or otherwise comprise the same GnT | mutation as CHO cell line JW152.

The mutation in GnT | may comprise a substitution, deletion or addition. It may be a point mutation. The mutation may result in a premature stop codon.

csT

The mammalian cells according to the invention may have a mutation in CMP-sialic acid transporter (CST). The CST may be, or be homologous fo, NP_001233684.1

G1:350537765.

We demonstrate that recombinant proteins of interest may be produced in cells lacking a functional CST gene and may comprise different O-glycans to a recombinant glycoprotein produced in cells with an unmutated CST. These results indicate that such cells have the poteniial to become a host cell line for producing proteins with reduced sialylation, including glycoprotein drugs and vaccines.

The mutation in CST in the mammalian cells according fo the invention results in a loss of CST function. The CST function may be reduced or entirely absent. +15 Function of the CST gene in the cells according to the invention may be reduced as compared with the function of wild type CST, such as wild type human CST. In some cases the function is reduced relative to CST found in wild type CHO cells.

Expression of giycoproteins in such cells may be beneficial for producing polypeptides with a similar or identical pattern of glycosylation to polypeptides with the same or very similar polypeptides produced by cancerous cells, for example adenocarcinoma cells. Cells with a mutated CST as described herein may be used for the production of proteins which have altered O-glycans compared to those produced by cells without a mutated CST. Such proteins may have reduced sialylation. The resulting proteins may comprise N-antigen type glycans, or be N- antigen glycoproteins.

UGT

The mammalian cells according to the invention may have a mutation in UDP- galactose transporter (UGT). The UGT may be, or be homologous to, CBL95110.1

G1:296173022. :

We demonstrate that recombinant proteins of interest may be produced in cells lacking a functional UGT gene and may comprise different O-glycans to a glycoprotein produced in a cell with an unmutated UGT. These results indicate that such cells have the potential to become a host cell line for producing proteins with wr reduced galactosylation, including glycoprotein drugs and vaccines.

The mutation in UGT in the mammalian cells according to the invention results in a loss of UGT function. The UGT function may be reduced or entirely absent.

Function of the UGT gene in the cells according to the invention may be reduced as compared with the function of wild type UGT, such as wild type human UGT. In some cases the function is reduced relative to UGT found in wild type CHO cells.

Expression of glycoproteins in such cells may be beneficial for expression of polypeptides with a similar or identical pattern of glycosylation to polypeptides with the same or very similar polypeptides produced by cancerous cells, for example adenocarcinoma cells. Cells with a mutated UGT as described herein may be used for the production of proteins which have altered O-glycans compared to those produced by cells without a mutated CST. Such proteins may have reduced sialylation. The resulting proteins may comprise Tn-antigen or STn-antigen type glycans, or be Tn-antigen or STn-antigen glycoproteins.

GFT

The mammalian cells according to the invention may have a mutation in GDP-fucose transporter (GFT). The FGT may be, or be homologous to, NP_001233737.1

G1:350538845.

The GDP-fucose transporter (abbreviated here as GFT, also known as Sic35¢1) was first identified by genetic complementation analyses based on samples from CDG-lIc patients (8,9). Defects in this gene have been associated with leukocyte adhesion deficiency II (LAD-II), alias CDG-llc. The amino acid sequence of GFT shows a high level of conservation with CMP sialic acid transporter (CST) and UDP-galactose transporter (UGT). It was predicted to have 10 transmembrane helices with both N- and C-termini in the cytosolic side, similar to the topology of CST which was experimentally established. However, elements that are critical for the localization and activity of GFT remain poorly understood.

We demonstrate that recombinant proteins of interest may be produced in cells lacking a functional GFT gene and may comprise reduced fucose in the N-glycans as compared to a recombinant protein produced in the presence of an unmutated GFT. : These results indicate that such cells have the potential to become a host cell line for producing proteins with reduced fucose, including antibodies, glycoprotein drugs and : vaccines. > .

The mutation in GFT in the mammalian cells according to the invention results in a loss of GFT function. The GFT function may be reduced or entirely absent.

Function of the GFT gene in the cells according fo the invention may be reduced as compared with the function of wild type GFT, such as wild type human GFT. in some cases the function is reduced relative to GFT found in wild type CHO cells.

Expression of glycoproteins in such cells may be beneficial for expression of polypeptides with a similar or identical pattern of giycosylation to polypeptides with the same or very similar polypeptides produced by cancerous cells, for example adenocarcinoma cells. Cells with a mutated GFT as described herein may be used for the production of proteins which have altered N-glycans that have reduced fucose as compared to those produced by cells without a mutated GFT. Such proteins may have reduced sialylation. i

Mucins :

Mucins are a family of high molecular weight heavily glycosylated proteins produced by epithelial tissues. Overexpression of mucin proteins, especially MUC1 is associated with many types of cancer. Although some mucins are membrane bound due to the presence of a hydrophobic membrane spanning domain that favours retention in the plasma membrane, most mucins are secreted onto mucosal surfaces.

MUC1 (Genbank Accession number 15941.3, G1.296439295)

MUC1 is also known as Mucin-1, MUC-1, Breast carcinoma-associated antigen DF3,

Carcinoma-associated mucin, Episialin, H23AG, PEMT, Peanut-reactive urinary mucin (PUM), Polymorphic epithelial mucin (PEM), Tumor-associated epithelial membrane antigen (EMA), Tumor-associated mucin and CD227.

Representative nucleic acid sequences of MUC1 include: Human pancreatic mucin mRNA (GenBank Accession Number J05582.1), Homo sapiens mucin 1, cell surface associated (MUC1), transcript variant 1, mRNA (NCBI Reference Sequence:

MP_002456.4).

Representative amino acid sequences of MUC1 include: MUC1 (Homo sapiens) (GenBank Accession Number: CAA56734.1), mucin-1 isoform 1 precursor (Homo sapiens) (NCBI Reference Sequence NP_002447 4), mucin- isoform 2 precursor (Homo sapiens} (NCBI Reference Sequence NP_00101801.6.1), mucin-1 isoform 3 precursor (Homo sapiens) (NCBI Reference Sequence NP_001018017.1)..

Human mucin 1 (MUCH) is a heavily glycosylated transmembrane protein expressed on the apical surface of most epithelial tissues.

During maturation, MUC1 is cleaved into two subunits, an extracelluiar N-terminal subunit and a C-terminal transmembrane subunit. A unique feature of MUC1 is the variable-number of tandem repeats (VNTR) that is located in the middle of the

Nterminal subunit. Each tandem repeat (TR) contains identical 20 amino acids (HGVTSAPDTRPAPGSTAPPA). The Ser and Thr residues in the TR are O- glycosylation sites. .

The N-terminal subunit also contains 4 N-glycosylation sites near its C-terminus. In normal epithelium, MUC1 is expressed at low levels and glycosylated with highly’ branched complex O-glycans.

However, in adenocarcinomas, such as breast, ovarian and pancreatic cancers,

MUC1 is overexpressed in 80% of the cases. in addition, MUC1 expressed on cancer cells contains shortened O-glycans such as T antigen (Galp1-3GalNAca-O-

Ser/Thr), Tn antigen (GalNAco-OSer/Thr) and sialyl-Tn (STn) antigen (NeuAca2-

BGalNAca-O-Ser/Thr) which are collectively called tumor-associated antigens. These glycans are not expressed in any adult tissues. MUC1 was recently recognized by the National Cancer Institute (USA) as one of the three most important tumor proteins for vaccine development. MUC1 is also overexpressed in 90% of the subset of breast cancer patients who are not responsive to tamoxifen or aromatase inhibitors, or the drug Herceptin. These so-called friple-negative tumors are extremely aggressive and difficult to treat. A vaccine directed against MUC1 has fremendous potential as a prophylactic vaccine or as a therapeutic vaccine for fighting breast cancers and other adenocarcinomas.

Antibodies

Antibodies may be raised against glycoproteins produced by the cells according to the invention. Such antibodies may be useful in medicine, for example in the treatment of cancer. Such antibodies may be monoclonal or polyclonal. The antibodies may be specific to glycoproteins produced by the cells of the invention, such that they are able to distinguish between glycoproteins produced by the cells of the invention, and glycoproteins produced in wild type, or unmodified cells.

Suitable monoclonal antibodies can be prepared using methods well known in the art (e.g. see Kbhler, G.; Milstein, C. (1975). "Continuous cultures of fused cells secreting : antibody of predefined specificity". Nature 256 (5517): 495; Siegel DL (2002). "Recombinant monoclonal antibody technology”. Schmitz U, Versmold A, Kaufmann

P, Frank HG (2000); "Phage display: a molecular tool for the generation of antibodies—a review". Placenta. 21 Suppl A: S106—12. Helen E. Chadd and Steven

M. Chamow; “Therapeutic antibody expression technology,” Current Opinion in

Biotechnology 12, no. 2 (April 1, 2001): 188-194; McCafferty, J.; Griffiths, A.; Winter,

G.; Chiswell, D. (1990). "Phage antibodies: filamentous phage displaying antibody variable domains". Nature 348 (6301): 652-554; "Monocional Antibodies: A manual of techniques ", H Zola (CRC Press, 1988) and in "Monoclonal Hybridoma

Antibodies: Techniques and Applications ", J G R Hurrell {CRC Press, 1982).

Chimaeric antibodies are discussed by Neuberger et al (1988, 8th International . Biotechnology Symposium Part 2, 792-789) .

In certain embodiments of the invention the glycoprotein expressed by the cells of the invention is an antibody or antibody fragment. In these embodiments, the cell preferably comprises a mutated GFT.

Fragments of antibodies, such as Fab and Fab, fragments may also be used as can genetically engineered antibodies and antibody fragments. The variable heavy (Vy) and variable light (V) domains of the antibody are involved in antigen recognition, a fact first recognised by early protease digestion experiments. Further confirmation was found by "humanisation" of rodent antibodies. Variable domains of rodent origin may be fused to constant domains of human origin such that the resultant antibody retains the antigenic specificity of the rodent parented antibody (Morrison et al (1984)

Proc. Nati. Acad. Sd. USA 81, 6851-6855).

That antigenic specificity is conferred by variable domains and is independent of the constant domains is known from experiments involving the bacterial expression of antibody fragments, all containing one or more variable domains. These molecules include Fab-like molecules (Better et al (1988) Science 240, 1041); Fv molecules (Skerra et al (1988) Science 240, 1038); single-chain Fv (ScFv) molecules where the

Vy and Vi partner domains are linked via a flexible oligopeptide (Bird et al (1888)

Science 242, 423; Huston et al (1988) Proc. Natl. Acad. Sd. USA 85, 5879) and ’ single domain antibodies (dAbs) comprising isolated V domains (Ward et al (1989)

Nature 341, 544). A general review of the techniques involved in the synthesis of antibody fragments which retain their specific binding sites is to be found in Winter &

Milstein (1991) Nature 349, 293- 299.

By "ScFv molecules" we mean molecules wherein the Vy and V, partner domains are covalently linked, e.g. directly, by a peptide or by a fiexibie oligopeptide.

Fab, Fv, ScFv and dAb antibody fragments can all be expressed in and secreted from E. coli, thus allowing the facile production of large amounts of the said fragments.

Whole antibodies, and F(ab"), fragments are "bivalent". By "bivalent" we mean that ’ the said antibodies and F(ab"), fragments have two antigen combining sites. In contrast, Fab, Fv, ScFv and dAb fragments are monovalent, having only one antigen combining site. Synthetic antibodies which bind to the glycoproteins of the invention may also be made using phage display technology as is well known in the art (e.g. see "Phage display: a molecular tool for the generation of antibodies--a review".

Placenta. 21 Suppl A: 5106-12. Helen E. Chadd and Steven M. Chamow; "Phage . antibodies: filamentous phage displaying antibody variable domains". Nature 348 (6301): 552-554).

Therapy .

Glycoproteins of the present invention may be used in medicine. In certain embodiments, the glycoproteins of the invention are useful in the treatment of tumours and cancer in animals in need of treatment thereof. Preferably, the animal undergoing treatment is a human patient in need of such treatment. Preferably the : cells are mammalian cells, more preferably human cells.

Glycoproteins of the invention may be formulated as pharmaceutical compositions for clinical use and may comprise a pharmaceutically acceptable carrier, diluent or adjuvant. The composition may be formulated for topical, parenteral, intravenous, intramuscular, intrathecal, intraocular, subcutaneous, oral, inhalational or transdermal routes of administration which may include injection. Injectable formulations may comprise the selected compound in a sterile or isotonic medium.

Administration is preferably in a "therapeutically effective amount”, this being sufficient to show benefit to the individual. The actual amount administered, and rate and time-course of administration, will depend on the nature and severity of the disease being treated. Prescription of treatment, e.g. decisions on dosage etc, is within the responsibility of general practitioners and other medical doctors, and typically takes account of the disorder to be treated, the condition of the individual patient, the site of delivery, the method of administration and other factors known to practitioners. Examples of the techniques and protocols mentioned above can be found in Remington's Pharmaceutical Sciences, 20th Edition, 2000, pub. Lippincoit,

Williams & Wilkins.

The term “treatment” as used herein pertains generally to treatment and therapy, in which some desired therapeutic effect is achieved, for example, the inhibition of the progress of the condition, and includes a reduction in the rate of progress, a halt in the rate of progress, amelioration of the condition, and the cure of the condition.

Treatment as a prophylactic measure (i.e. prophylaxis) is also included.

Alternatively, targeting therapies may be used fo deliver the active agent more specifically to certain types of cell, by the use of targeting systems such as antibody or cell specific ligands. Targeting may be desirable for a variety of reasons; for example if the agent is unacceptably toxic, or if it would otherwise require too high a dosage, or if it would not otherwise be able to enter the target cells.

A composition may be administered alone or in combination with other treatments, either simultaneously or sequentially dependent upon the condition to be treated.

In some embodiments of the invention, cells that have been obtained from the patient in need of treatment are exposed to a glycoprotein expressed by a cell according to . the invention outside of the body, i.e. ex vivo. For example, the cells may be pulsed with the glycoprotein. The cells obtained from the patient may be immune cells, for example dendritic cells or monocytes. Following exposure to the glycoprotein, the cells may be returned io the patient in need of treatment thereof. The cells isolated from the patient are expanded ex vivo, before and/or after exposure to the glycoprotein.

Patient

The patient to be treated may be any animal or human. The patient is preferably a non-human mammal, more preferably a human patient. The patient may be male or female.

Proteins, Fragments and Derivatives

Whilst components used in the methods of the present invention may comprise full- length protein sequences, this is not always necessary. As an alternative, i5 homologues, mutants, derivatives or fragments of the full-length polypeptide may be used.

Derivatives include variants of a given full length protein sequence and include naturally occurring allelic variants and synthetic variants which have substantial amino acid sequence identity to the full length protein.

Protein fragments may be up to 10, 20, 30, 40, 50, 60, 70, 80, 90, 100 or 150 amino acid residues tong. Minimum fragment length may be 3,4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 or 30 amino acids or a number of amino acids between 3 and 30.

Mutants may comprise at least one modification (e.g. addition, substitution, inversion and/or deletion) compared to the corresponding wild-type polypeptide. The mutant may display an altered activity or property, e.g. binding.

Mutations may occur in any of SEQ ID No.s 1-18 and components containing such fragments may serve the purpose of modulating the activity of the mutant to restore, completely or partially the activity of the wild type polypeptide.

Derivatives may also comprise natural variations or polymorphisms which may exist between individuals or between members of a family. All such derivatives are included within the scope of the invention. Purely as examples, conservative replacements which may be found in such polymorphisms may be between amino acids within the following groups:

Co (i alanine, serine, threonine; (ii) glutamic acid and aspartic acid; (iii) arginine and leucine; (iv) asparagine and giutamine; (v) isoleucine, leucine and valine; (vi) phenylalanine, tyrosine and tryptophan.

Derivatives may also be in the form of a fusion protein where the protein, fragment, homologue or mutant is fused to another polypeptide, by standard cloning : techniques, which may contain a DNA-binding domain, transcriptional activation domain or a ligand suitable for affinity purification (e.g. glutathione-S-transferase or six consecutive histidine residues.

Methods for expressing a protein of interest from a mammalian cell are known in the art. A gene encoding the glycoprotein of interest may be inserted into the genome of a mammalian cell, or introduced on a vector. The DNA of the gene is transcribed to messenger RNA which is then translated into polypeptide chains which are ultimately folded inio proteins.

Following culture of cells that express glycoprotein, the glycoprotein is preferably isolated. Any suitable method for separating proteins from cell culture known in the art may be used. In order {o isolate a protein of interest from a culture, it may be necessary to first separate the cultured cells from media containing the protein of interest. If the protein of interest is secreted from the cells, the cells may be separated from the culture media that contains the secreted protein by centrifugation.

If the protein of interest collects within the cell, it will be necessary to disrupt the cells prior to centrifugation, for example using sonification, rapid freeze-thaw or osmotic lysis. Centrifugation will produce a pellet containing the cultured cells, or cell debris of the cultured cells, and a supernatant containing culture medium and the protein of interest.

It may then be desirable to isolate the protein of interest from the supernatant or culture medium, which may contain other protein and non-protein components. A common approach to separating protein components from a supernatant or culture medium is by precipitation. Proteins of different solubilities are precipitated at different concentrations of precipitating agent such as ammonium sulfate. For example, at iow concentrations of precipitating agent, water soluble proteins are extracted. Thus, by adding different increasing concentrations of precipitating agent, proteins of different solubilities may be distinguished. Dialysis may be subsequently used to remove ammonium sulfate from the separated proteins.

Brief Description of the Figures

Embodiments and experiments illustrating the principles of the invention will now be discussed with reference to the accompanying figures in which:

Figure 1. N- and O-glycans attached to MUC1 produced by wild type and mutant

CHO cells. The mutant cells produce mannose-terminated N-glycans to enhance uptake by dendritic cells and tumor-associated O-glycans to trigger cancer specific immune responses. A, glycans produced by different CHO cells. B, N-terminal subunit of MUC 1produced by different CHO cell.

Figure 2. Genetic analyses of CHO-gmt1 (A) and CHO-gmit2 (B). CHO-gmt1 lacks

CMP-sialic acid transporter (CST) activity. CHO-gmt2 lacks UDP-galactose transporter (UGT) activity.

Figure 3. Genetic analyses of CHO-gmt6 (A) and CHO-gmit7 (B). CHO-gmi6 lacks both CMP-sialic acid transporter (CST) activity and N-acetylglucosaminyltransferase (GnT I) activity. CHO-gmt7 lacks both UDP-galactose transporter (UGT) activity and

GnT 1 activity.

Figure 4. Recombinant MUC1-Fc produced by CHO-K1 and different CHO glycosylation mutant cells were analyzed by Western blot using anti-MUC1 3

Figure 5. Design of ZFNs targeting the GFT gene in CHO cells. (A) A DNA sequence in the coding region of Chinese hamster GFT gene that is ideal for targeting by ZFNs.

Note that the left finger binding site and the right finger binding site are separated by 6 bps. (B) Zinc fingers were designed based on published literature. Each ZFN oo contains 4 fingers. (C) Mutations at the fargeted site were identified in different clones, including deletion and insertion mutations. Both alleles of the GFT gene in clone E were mutated by the ZFNs. It has been named CHO-gmt5.

Figure 6. Lectin staining and MALDI-TOF characterization of glycan structures of wild-type and mutant CHO cells. (A) CHO-K1, CHO-gmt1, and CHO-gmt5 cells were seeded on glass coverslips, cultured overnight before being fixed and permeabilized.

Terminal galactose residues were detected using FITC-conjugated PNA (colored green). Fucose residues were detected using biotinylated AAL and AlexaFluor 647- conjugated streptavidin (pseudo-colored red). Nuclei were stained by Hoechst 33342 : and colored blue. (B) N-glycans isolated from recombinant EPO-Fc produced in

CHO-gmt1 and CHO-gmt5 cells were analyzed by MALDI-TOF. The N-glycans produced by CHO-gmt1 were mostly fucosylated, and the dominant species were the asialo, core-fucosylated galactosyl biantennary, triantennary and tetraantennary glycans. However, the N-glycans produced by CHO-gmt5 were the asialo, afucosylated galactosyl bi-antennary, tri-antennary and tetra-antennary glycans.

Figure 7. Localization of GFT within the Golgi. HA-tagged human GFT was transiently transfected into Hela cells and GFT was detected by a monoclonal anti-

HA antibody (colored green). Golgi compartments were stained by antibodies specific “for defined Golgi markers, GM 130 (cis-Golgi), Manli (medial-Golgi), B4GalT1 (trans-

Golgi) and TGN48 (trans Golgi network, or TGN). All Golgi markers were colored red.

Nuclei were stained by Hoechst 33342 and colored blue. The boxed areas in the merged images were enlarged and shown on the right. Scale bar: 20 ym.

Figure 8. The cytosolic C-terminal tail sequence, and more specifically aa 341 — 355, is essential for GFT transport activity but not its localization to the Golgi. (A)

Schematic representation of GFT constructs with different tail sequences. FL, full- length GFT; AC-tail, GFT with a deletion from aa 341 — 364; CST-{ail, replacement of aa 341 — 364 with the C-terminal sequence (aa 317 — 336 according to CST numbering) of CST; ACT1, GFT with a deletion of aa 356 — 364; ACT2, GFT with a deletion of aa 341 ~ 355. All GFT constructs were N-terminally tagged with an HA epitope (shown as black boxes). Predicted transmembrane regions are shown as gray boxes. B. Immunoflucorescence and lectin staining analyses for localizations and : activities of GFT variants. CHO-gmt5 cells seeded on glass coverslips were transfected with a GFT construct. GFT was detected by a monoclonal HA antibody and colored green. Golgi was marked by antibody specific for giantin and colored red. Fucose residues were detected by biotinylated AAL and AlexaFluor 647- conjugated streptavidin. Nuclei were stained by Hoechst 33342 and colored blue.

Scale bar: 20 ym. (C) FACS analysis of cell surface fucosyiation. CHO-gmt5 cells - were transfected with a GFT construct and fixed without a permeabilization step two days after transfection. Cell surface fucose residues was labeled by biotinylated AAL and Cy3-conjugated streptavidin. FACS analysis was performed in duplicates on a

BD LSR Il analyzer and representative results were shown. In each histogram plof, shaded area shows negative control (untransfected CHO-gmt5 incubated with biotinylated AAL and Cy3-conjugated streptavidin). Red line shows the profile of

CHO-gmt5 transfected with the indicated GFT construct. Horizontal line indicates the gate for defining the AAL-positive population. (D) Quantification of FACS results.

Percentage of cells in the gated region (as shown in C) was taken and normalized to that of FL-transfected cells fo indicate the relative transport activity of different C-tail variants. Average values are shown. Error bar indicates standard deviation.

Figure 8. Two elements in the C-terminal tail, a cluster of Lys residues and the EM motif, are independently required for the transport activity, but not localization of

GFT. (A) Schematic representation of GFT constructs with different tail sequences.

FL, full-length GFT; 3K/3G, a mutant canstruct with Lys346, Lys347, and Lys355 mutated to Gly (underlined); EM/GG, a mutant construct with 344EM345 mutated to

GG (underlined); EM-CST-tail, replacement of aa 347 — 364 with the corresponding : C-terminal sequence (aa 320 — 336 according to CST numbering) of CST (CST sequence underlined). Note the presence of 344EM345 in the EM-CST-tail construct.

All GFT constructs were N-terminally tagged with an HA epitope (shown as black boxes). Predicted transmembrane regions are shown as gray boxes. (B)

Immunoflucrescence and lectin staining analysis for localization and activity of 3K/3G, EM and EM-CST-tail variants in CHO-gmt5 cells. CHO-gmt5 cells seeded on glass coverslips were transfected with a GFT construct. GFT was detected by a monoclonal HA antibody and colored green. Golgi was marked by antibody specific for giantin and colored red. Fucose residues were detected by biotinylated AAL and

AlexaFluor 647-conjugated streptavidin. Nuclei were stained by Hoechst 33342 and colored blue. Scale bar: 20 ym. (C) FACS analysis of cell surface fucosylation. CHO- gmtd cells were transfected with a GFT construct and fixed without a: permeabilization step two days after transfection. Cell surface fucose residues was labeled by biotinylated AAL and Cy3-conjugated streptavidin. FACS analysis was performed in duplicates and representative results were shown. in the FACS histogram plot, shaded area shows negative control (untransfected CHO-gmt5 incubated with biotinylated AAL and Cy3- conjugated streptavidin). Red line shows the profile of CHO-gmt5 transfected with the indicated GFT construct. Horizontal line indicates the gate for defining the AAL-positive population. (D) Quantification of

FACS results. Percentage of cells in the gated region was taken and normalized to that of FLtransfected cells to indicate the relative transport activity of different C-tail variant. Average values are shown in the bar chart. Error bar indicates standard deviation.

Figure 10. Differential involvement of three Gly residues located in the transmembrane helical regions on GFT activity. (A) Schematic representation of GET and CST. Gray boxes show predicted transmembrane regions in GFT and CST.

Three conserved Gly residues in transmembrane helices, 5 ,6 and 8, are indicated.

The black boxes indicate an HA tag. (B) Immunofluorescence and lectin staining analyses for localization and activity of a GFT variant containing Gly->lle or Gly->Tyr substitution at positions 180, 198, and 277. GFT was detected by a monoclonal HA antibody and colored green. Golgi was marked by antibody specific for giantin and colored red. Fucose residues were detected by biotinylated AAL and AlexaFluor 647- conjugated streptavidin. Nuclei were stained by Hoechst 33342 and colored biue.

Scale bar: 20 ym. (C) FACS analysis of cell surface fucosylation. CHO-gmt5 cells were transfected with GFT constructs and fixed without a permeabilization step two days after transfection. Cell surface fucose residues was labeled by biotinylated AAL and Cy3-conjugated streptavidin. FACS analysis was performed in duplicates and representative results were shown. In each histogram plot, shaded area shows negative control (untransfected CHO-gmt5 incubated with biotinylated AAL and Cy3- conjugated streptavidin). Red line shows the profile of CHO-gmt5 transfected with the indicated GFT construct. Horizontal fine indicates the gate for defining the AAL- positive population. (D) Quantification of FACS results. Percentage of cells in the gated region was taken and normalized to that of FL transfected cells io indicate the relative transport activity of different C-tail variant. Average values are shown. Error bar indicates standard deviation.

Detailed Description of the Invention

The details of one or more embodiments of the invention are set forth in the accompanying description below including specific details of the best mode contemplated by the inventors for carrying out the invention, by way of example. It will be apparent to one skilled in the art that the present invention may be practiced without limitation to these specific details.

Example 1

CHO glycosylation mutants can be used to produce recombinant MUC1 with tumor- associated O-glycans

Chinese hamster ovary (CHO) celis have been the main host cells for producing recombinant glycoproteins in the biotechnology industry. The N- and O-glycosylation in CHO cells has been a major research focus for many years. The O-glycosyiation pathways can be complex and cell type specific in different mammalian cells (Tarp - and Clausen, 2008). However, the O-glycosylation in CHO cells is very simple with a core 2 to core 1 shift. The O-glycans on the recombinant erythropoietin (EPO) produced by CHO cells only existed in two forms: ST and diST (Hokke et al., 1995).

Recombinant MUC1 produced by CHO cells also contained mainly ST and diST (Backstrom et al., 2003; Olson et al., 2005; Rughetti et at., 2005). A detailed analysis revealed that the O-glycans attached to recombinant MUC1 produced in CHO cells contained 85% ST, 13% diST and only 2% T (Rughetti et al., 2005). The total O- glycans derived from CHO cells grown in suspension and CHO cells grown in : monolayer both contained ST and diST (North et al., 2010).

Numerous CHO glycosylation mutants have been isolated and characterized by several labs. Stanley and colleagues reported the Lec (for lectin resistant) series of

CHO glycosylation mutants (Patnaik and Stanley P, 2006). Unfortunately, these cells do not grow in serum-free medium in suspension cultures. As a result, although the

Lec mutants played important roles in elucidating the biological functions of protein glycosylation and possible applications of these mutants in biotechnology industry, they have not been utilized in large scale production of any recombinant proteins.

Using cytotoxic lectins, mainly Maackia amurensis agglutinin (MAA) and Ricinus communis agglutinin-1 (RCA-l, or RCA-120), we have isolated several CHO glycosylation mutants from the wild-type CHO-K1 cells (Lim et al., 2008; Goh et al, 2010). These mutants can easily be adapted in serum-free medium and cultured in suspension in bioreactors. The growth rate and final viable cell density reached by these mutants are comparable to wild-type CHO-K1 cells. Therefore, our CHO mutants have the potential to be the host cells for large scale production of recombinant proteins. The following three existing CHO glycosyiation mutants characterized in this fab are relevant to this proposal and will be described in more details.

CHO-gmt1

CHO-gmt1 (formerly MAR-11} has been characterised (Lim et al., 2008). The total cell surface sialic acid on CHO-gmt1 is lower than that on a previously reported CHO mutant fine, Lec2. Biochemical analyses indicate that recombinant glycoproteins produced by these cells lack sialic acid. Genetic test suggested that CHO-gmt1 cells lacked CMP-sialic acid transporter (CMP-SAT) activity (Fig. 2A). Molecular cloning of

CMP-SAT cDNA from CHO-gmt1 cells revealed a C to T point mutation which results in a premature stop codon. As a result, CHO-gmt1 cells express a truncated version: of CMP-SAT which contains only 100 amino acids, in comparison to the normal

CMP-SAT which contains 336 amino acids (Lim et al., 2008). The predicted O- glycans produced in CHO-gmi1 cells will be mainly T antigen which is tumor- associated antigens.

CHO-gmt2

CHO-gmt2 was also isolated by MAA. CHO-gmt2 has a mutated UDP-galactose transporter (UDP-GalT) gene (Fig. 2B). The G538C mutation in the open reading frame results in a A180P mutation in the UDP-GalT protein. With a dysfunctional

UDP-GalT, CHO-gmt2 cannot transport UDP-Gal into the Golgi apparatus. As a result, O-glycans produced in CHO-gmt2 cells will be Tn and STn which are also tumor-associated antigens. It has been shown recently that Tn-based vaccines may directly target DCs and develop potent antibody response in the absence of adjuvant (Freire et al., 2010).

CHO-gmt4

CHO-gmt4 (formerly JW152) was isolated by RCA-| and has been (Goh et al., 2010).

Complementation tests revealed that all the RCA-l-resistant mutants possessed a dysfunctional GnT | gene (Goh et al., 2010). Protein O-glycosylation in these cells should not be affected. However, the N-glycosylation is prematurely terminated at the

MansGIcNAC, stage due to the lack of GnT I. We have shown that recombinant EPQ produced in these cells carries MansGlcNAc, glycan and some fucosyiated

MansGlcNAc, glycan. A small amount of Man,GlcNAc, was also attached to the recombinant EPO. As described earlier, proteins that carry mannose-terminated N- “glycans can be targeted to DCs and macrophages via their mannose-binding C-type lectins. A surprising finding during the course of this work was that all the CHO cells that survived the RCA-| treatment carry a mutated GnT | gene. A possible explanation for this finding is that RCA-l may bind all different N-glycans, although with different affinity, except MansGIcNAc. N-glycan. Therefore, all the celis survived

RCA-I must have a dysfunctional GnT 1. In our lab, treating CHO cells with RCA-I has become a simple method to isolate cells with mutated GnT | gene. We have filed two international patent applications for this discovery (Patent I: Pub. No.: US 2011/0177555 A1. Title: GnT | Mutant CHO Cell Lines; Patent II: United States : Patent Application 61/317,369 Title: Method of Producing Mannose-terminated

Recombinant Proteins).

CHO-gmt6 and CHO-gmi7

We have isolated two novel CHO glycosylated mutants that have been named CHO- gmt6 and CHO-gmi7 respectively. CHO-gmi6 cells lack functional CMP-sialic acid transporter (CST) and GnT | (Fig. 3A). CHO-gmt7 cells lack UDP-galactose transporter (UGT) and GnT | (Fig. 3B). As a result, the N- and O-glycans produced in

CHO-gmt6 and CHO-gmt7 cells are different from that produced in wild type CHO cells. The N-glycans produced by these two mutants are MansGicNAc,. This mannose-terminated N-glycan can be specifically recognized by the mannose- binding receptors on DCs and macrophages and therefore enhances the antigenicity of the protein. The O-glycans produced by CHO-gmt6 are T antigens whereas the O- glycans produced by CHO-gmt7 are Tn and STn antigens.

CHO-gmit6 and CHO-gmt7 will be used as host cells to produce a smalier version of the N-terminal subunit of MUC1 as anti-cancer vaccine. This recombinant N-terminal subunit of MUC1 will have 5~10 TRs. The T, Tn and STn O-glycans attached to the

MUC1 produced by CHO-gmt6 and CHO-gmt7 can trigger tumor-specific immune response in the host as their O-glycans are the same as those attached to the MUC1 ) on tumor cells. The mannose-terminated N-glycan on this recombinant MUC1 will increase it immunogenicity by enhancing the uptake by DCs and macrophages.

Expression of recombinant MUC1 N-terminal subunit in newly isolated CHO glycosylation mutants, CHO-gmt6 and CHO-gmt7.

Expression constructs will be generated to express a modified N-terminal subunit of

MUC1. The amino acid sequence of the N-terminal subunit.of MUC1 will remain the same except the number of the VNTR will be reduced to 5~10. The cDNA will encode 497 amino acids. During maturation, a signal peptide of 23 amino acids at “the N-terminus will be removed. The secreted mature proteins will contain 474 amino acids. There will be 50 potential O-glycosylation sites in the 10 TR region and 4 N- glycosylation sites at the C-terminal region. in the future, the number of TRs in the construct may increase or decrease. The protein will be transiently expressed in

CHO-K1 cells and CHO-gmt6 and CHO-gmt7 cells. The presence of the recombinant protein in conditioned media will be probed by anti-MUC1 antibodies. Stably transfected clones that express high levels of recombinant MUC1 N-terminal subunit will be isolated and banked for large scale production.

Purification of recombinant MUC1 N-terminal subunit expressed in CHO-gmt6 and

CHO-gmi7 cells.

Recombinant MUC1 N-terminal subunit will be purified by a Concanavalin A (Con A) affinity chromatography followed by size-exclusion gel filtration chromatography. The

U-mannose residues at the non-reducing end of the Man5GIcNAc2 N-glycan on the recombinant MUC1 subunit will specifically bind Con A linked to the matrix. The

MUC1 subunit bound to the column can be eluted by an O-mannoside or an {J- glucoside, or by changing pH. The presence of the recombinant MUC1 subunit will be followed by anti-MUC1 antibodies during purification.

Evaluation of recombinant MUC1 N-terminal subunits as potential vaccines in mice.

Mice will be used as the model animals for testing the efficacy of these recombinant vaccines. B-cell response elicited by the vaccine will be evaluated by measuring the titer of the antibody in immunized mice. The T-cell response will be measured by the activity of CTL in immunized mice.

Results

A vector that expresses the N-terminus of MUC1 that contains 5 TRs fused to the Fc of igG1 (MUC1-Fc¢) has been generated. Recombinant MUC1-Fe¢ has been produced by wild type CHO-K1 cells and CHO-gmt1, CHO-gmt2, CHO-gmt6 and CHO-gmt7 = - celis. The results are shown in Fig. 4. :

Example 2

The MUC1 glycopeptide with tumor-associated O-glycans will trigger cancer-specific anti- MUC1 antibody responses. Mannose-terminated N-glycans on the MUC1 will enhance its efficacy by enhancing its uptake by DCs and macrophages.

To investigate the efficacy of recombinant N-terminal subunit of MUC1 as therapeutic vaccines in mouse models and breast cancer patients the following experimental procedure is used: 1. Production of MUC1-Fc fusion proteins. a. Generation of an expression construct to express MUC1-Fc fusion protein. MUCA (N terminal subunit) and Fc are linked by a cleavable linker. b. Production of MUC1-F¢ in CHO-gmt5 and CHO-gmt6 ceils. For pilot human study, the production will be done by our industry collaborator that has GMP facilities. 2. Mouse Studies a. Tumor kinetic studies. MUC1 mammary tumors will be established in MUC1 transgenic mice. These mice will be immunized with liposomal preparations of the N- terminal subunit of recombinant MUC1 with empty liposomes as controls. b. Cytolytic activity. In vaccinated and controf mice, CD8 positive cells will be isolated from tumor draining lymph nodes (without any ex-vivo stimulation) and used as effector cells. 51Cr labeled DCs pulsed with MUC1 peptides will be used as target cells. Percentage killing will be measure with 51Cr-release-assay. ¢. Determination of ADCC. Serum from vaccinated mice will be incubated with 51Cr labelled MUC1 expressing tumor cells. NK cells (effector) will be co-incubated with tumor cells and the release of Cr will be determined. d. Interferon-gamma Ellispot Assays. CD8+ T cells from the draining lymph nodes of vaccinated mice will be isolated and the level of MUCH antigen-specific responses will be determined by Inf-g Ellispot Assays.

co e. Antibody titres. Lavels of anti-MUC1 IgG, 1gG1, IgG2, IgG3 and igM will be determined. 3. Human ex vivo studies a. DCs cultured and expanded from PBMCs from healthy donors and breast cancer patients will be pulsed with recombinant MUC1 N-terminal subunit produced by CHO- gmtS and CHO-gmt6 cells, and matured ex-vivo. Autologous mononuclear cells will be co-cultured with these DCs in the presence of cytokines. b. The phenotype, activation and maturation markers of DCs will be determined by flow cytometry. ¢. Supernatant will be analysed for IL-12 and INF-gamma d. Tumor cell killing. Mononuclear cells from the co-culture will be incubated with Cr 51Cr labelled tumor cells and lytic activity measured. MHC class 1 blocking antibody will be used to determine MHC restriction. 4. Pilot human study. Following animal toxicology studies, a phase |, first-in-man, i5 dose escalation pilot study wili be undertaken to explore the safety and tolerability of the vaccine and evaluate the ability of the vaccine to induce MUC1 antigen specific responses in vivo.

References for Examples 1 and 2

Apostolopoulos V, Pietersz GA, Loveland BE, Sandrin MS, McKenzie IF. (1995)

Oxidative/reductive conjugation of mannan to antigen selects for T1 or T2 immune responses. Proc Natl Acad Sci U S A. 92(22):10128-32.

Apostolopouios V, Pietersz GA, Tsibanis A, Tsikkinis A, Drakaki H, Loveland BE,

Piddlesden SJ, Plebanski M, Pouniotis DS, Alexis MN, McKenzie IF, Vassilaros S. (2006) Pilot phase lil immunotherapy study in early-stage breast cancer patients using oxidized mannan-MUC1 [ISRCTN7171 1835]. Breast Cancer Res. 8(3):R27.

Bafna S, Kaur S, Batra SK. (2010) Membrane-bound mucins: the mechanistic basis for alterations in the growth and survival of cancer cells. Oncogene. 29(20):2893-904.

Backstrom M, Link T, Olson FJ, Karlsson H, Graham R, Picco G, Burchell J, Taylor-

Papadimitriou J, Noll T, Hansson GC. (2003) Recombinant MUC1 mucin with a breast cancer-like O-glycosylation produced in large amounts in Chinese-hamster ovary. Biochem J. 376:677-86.

Beatson RE, Taylor-Papadimitriou J, Burchell JM. (2010) MUC1 immunotherapy.

Immunotherapy. 2(3):305-27. :

Blixt O, Bueti D, Burford B, Allen D, Julien S, Hollingsworth M, Gammerman A,

Fentiman |, Taylor-Papadimitriou J, Burchell JM. (2011) Autoantibodies to aberrantly glycosylated MUCH in early stage breast cancer are associated with a better prognosis. Breast Cancer Res. 13(2):R25.

Brockhausen i. (2006) Mucin-type O-glycans in human colon and breast cancer: glycodynamics and functions. EMBO Rep. 7(6):599-604.

Ding L, Lalani EN, Reddish M, Koganty R, Wong T, Samuel J, Yacyshyn MB, Meikle

A, Fung PY, Taylor-Papadimitriou J, et al. (1993) Immunogenicity of synthetic peptides related to the core peptide sequence encoded by the human MUC1 mucin gene: effect of immunization on the growth of murine mammary adenocarcinoma cells transfected with the human MUC1 gene. Cancer Immunol Immunother. 36(1):9- 17.

Dube DH, Bertozzi CR. (2005) Glycans in cancer and inflammation--potential for therapeutics and diagnostics. Nat Rev Drug Discov. 4(6):477-88.

Finke LH, Wentworth K, Blumenstein B, Rudolph NS, Levitsky H, Hoos A. (2007) : Lessons from randomized phase Ill studies with active cancer immunotherapies— outcomes from the 2006 meeting of the Cancer Vaccine Consortium (CVC). Vaccine. 27,25 Suppl 2:B97-B109.

Freire T, Zhang X, Dériaud E, Ganneau C, Vichier-Guerre S, Azria E, Launay O, Lo-

Man R, Bay S, Leclerc C. (2010) Glycosidic Tn-based vaccines targeting dermal dendritic cells favor germinal center B-cell development and potent antibody response in the absence of adjuvant. Blood. 116(18):3526-36.

Friedman B, Vaddi K, Preston C, Mahon E, Cataldo JR, McPherson JM. (1999) A comparison of the pharmacological properties of carbohydrate remodeled recombinant and placental-derived beta-glucocerebrosidase: implications for clinical efficacy in treatment of Gaucher disease. Blood. 93(9):2807-16.

Gazi U, Martinez-Pomares L. (2009) Influence of the mannose receptor in host immune responses. Immunobiology. 214(7):554-61.

Gilewski TA, Ragupathi G, Dickler M, Powell S, Bhuta S, Panageas K, Koganty RR,

Chin-Eng J, Hudis C, Norton L, Houghton AN, Livingston PO. (2007) Immunization of high-risk breast cancer patients with clustered sTn-KLH conjugate plus the immunologic adjuvant QS-21. Clin Cancer Res. 13(10):2977-85.

Goh J, Zhang P, Chan KF, Lee MM, Lim SF, and Song Z. (2010) RCA-I-resistant

CHO mutant cells contain dysfunctional GnT | and expression of GnT i in these mutants enhances sialylation of recombinant erythropoietin. Metab. Eng. 12(4):360- 368.

Goldman B., DeFrancesco L. (2009) The cancer vaccine roller coaster. Nat.

Biotechnol. 27(2):129-39.

Hanisch FG, Schwientek T, Von Bergweli-Baildon MS, Schultze JL, Finn O. (2003)

O-Linked glycans control glycoprotein processing by antigen-presenting cells: a biochemical approach to the molecular aspects ‘of MUC1 processing by dendritic cells. Eur J Immunol. 33(12):3242-54.

Hattrup CL., Gendler SJ. (2008) Structure and function of the cell surface (tethered) mucins. Annu Rev Physiol. 70:431-57.

Hokke CH, Bergwerff AA, Van Dedem GW, Kamerling JP, Vliegenthart JF. {1995)

Structural analysis of the sialylated N- and C-linked carbohydrate chains of recombinant human erythropoietin expressed in Chinese hamster ovary cells. Eur J

Biochem. 228(3):981-1008.

Hollingsworth MA, Swanson BJ. (2004) Mucins in cancer: protection and control of the cell surface. Nat Rev Cancer. 4(1):45-60. : Holmberg LA, Oparin DV, Gooley T, Lilleby K, Bensinger W, Reddish MA, MacLean

GD, Longenecker BM, Sandmaier BM. (2000) Clinical outcome of breast and ovarian cancer patients treated with high-dose chemotherapy, autologous stem cell rescue and THERATOPE 8Tn-KLH cancer vaccine. Bone Marrow Transplant. 25(12):1233- 41,

Hull SR, Bright A, Carraway KL, Abe M, Hayes DF, Kufe DW. (1989) Oligosaccharide differences in the DF3 sialomucin antigen from normal human milk and the BT-20 human breast carcinoma cell line. Cancer Commun. 1(4):261-7.

Julien 8, Picco G, Sewell R, Vercoutter-Edouart AS, Tarp M, Miles D, Clausen H,

Taylor-Papadimitriou J, Burchell JM. (2009) Sialyl-Tn vaccine induces antibody- mediated tumour protection in a relevant murine model. Br J Cancer. 100(11):1746- . 54.

Julien S, Delannoy P. (2003) Sialyl-Tn antigen in cancer: from diagnosis to therapy

Recent Research Developments in Cancer Kerala: Transworld Reserach Network: 185—199.Pandalai SG (ed) Vol. 5, pp.

Keppler-Ross S, Douglas L, Konopka JB, Dean N. (2010) Recognition of yeast by murine macrophages requires mannan but not glucan. Eukaryot Cell. 9(11):1776-87.

Kufe DW. (2009) Mucins in cancer: function, prognosis and therapy. Nat Rev Cancer. 9(12):874-85. :

Lim SF, Lee MM, Zhang P and Song Z. (2008) The Golgi CMP-sialic acid transporter:

A new CHO mutant provides functional insights. Glycobiology 18(11):851-60.

Link T, Backstrom M, Graham R, Essers R, Zdérner K, Gétgens J, Burchell J, Taylor-

Papadimitriou J, Hansson GC, Noll T. (2004) Bioprocess development for the production of a recombinant MUC1 fusion protein expressed by CHO-K1 cells in protein-free medium. J Biotechnol. 110(1):51-62. - Lloyd KO, Burchell J, Kudryashov V, Yin BW, Taylor-Papadimitriou J. (1996)

Comparison of O-linked carbohydrate chains in MUC-1 mucin from normal breast epithelial cell lines and breast carcinoma cell lines. Demonstration of simpler and fewer glycan chains in tumor cells. J Biol Chem. 271(52):33325-34.

Loveland BE, Zhao A, White S, Gan H, Hamilton K, Xing PX, Pietersz GA,

Apostoiopoulos VV, Vaughan H, Karanikas V, Kyriakou P, McKenzie IF, Mitchell PL. (2006) Mannan-MUC1-pulsed dendritic cell immunotherapy: a phase | trial in patients with adenocarcinoma. Clin Cancer Res. 12(3 Pt 1):869-77.

McGreal EP, Miller JL, Gordon S. (2005) Ligand recognition by antigen-presenting cell C-type lectin receptors. Curr Opin Immunol. 17(1):18-24.

Miles D, Papazisis K. (2003) Rationale for the clinical development of STn-KLH {Theratope) and anti-MUC-1 vaccines in breast cancer. Clin Breast Cancer. 3 (Suppl 4):5134-S138.

North SJ, Huang HH, Sundaram 3, Jang-Lee J, Etienne AT, Trollope A, Chalabi S,

Dell A, Stanley P, Haslam SM. (2010) Glycomics profiling of Chinese hamster ovary cell glycosylation mutants reveals N-glycans of a novel size and complexity. J Biol

Chem. 285(8):5759-75.

Olson FJ, Béckstrédm M, Karlsson H, Burchell J, Hansson GC. (2005) A MUC1 tandem repeat reporter protein produced in CHO-K1 cells has sialylated core 1 O- glycans and becomes more densely glycosylated if coexpressed with polypeptide-

GalNAc-T4 transferase. Glycobiology, 15(2):177-91.

Patnaik SK, Stanley P. (2008) Lectin-resistant CHO glycosylation mutants. Methods

Enzymol. 416:159-82.

Potapenko iO, Haakensen VD, Liiders T, Helland A, Bukholm |, Serlie T, Kristensen

VN, Lingjaerde OC, Barresen-Dale AL. (2010) Glycan gene expression signatures in normal and malignant breast tissue; possible role in diagnosis and progression. Mol

Oncol. 4(2):98-118. 3C

Plebanski M, Katsara M, Sheng KC, Xiang SD, Apostolopoulos V. (2010) Methods to measure T-cell responses. Expert Rev Vaccines. 9(6):595-600.

Rachagani S, Torres MP, Moniaux N, Batra SK. (2009) Current status of mucins in the diagnosis and therapy of cancer. Biofactors. 35(6):509-27.

Ragupathi G, Howard L, Cappello S, Koganty RR, Qiu D, Longenecker BM, Reddish

MA, Lloyd KO, Livingston PO. (1999) Vaccines prepared with sialyl-Tn and sialyl-Tn trimers using the 4-(4-maleimidomethyl)cyclohexane-1-carboxyl hydrazide linker group result in optimal antibody titers against ovine submaxillary mucin and sialyl-Tn- positive tumor cells. Cancer Immunol Immunother. 48(1):1-8.

Richardson JP, Macmillan D. (2008) Mucin-Based Vaccines. Glycoscience. Vol 2,

Part 12, p2645-2698, DOI: 10.1007/978-3-540-30429-6_68 | Rughetti A, Peliicciotta |, Biffoni M, Backstrom M, Link T, Bennet EP, Clausen H, Noll

T, Hansson GC, Burchell JM, Frati L, Taylor-Papadimitriou J, Nuti M. (2005)

Recombinant tumor-associated MUC1 glycoprotein impairs the differentiation and function of dendritic cells. J Immunol. 174(12):7764-72.

Te 15 Serensen AL, Reis CA, Tarp MA, Mandel U, Ramachandran K, Sankaranarayanan

V, Schwientek T, Graham R, Taylor-Papadimitriou J, Hollingsworth MA, Burchell J,

Clausen H. (2006) Chemoenzymatically synthesized multimeric Tn/STn MUCH glycopeptides elicit cancer-specific anti-MUC1 antibody responses and override tolerance. Glycobiology. 16(2):96-107. 20 . i

Soares MM, Mehta V, Finn QJ. (2001) Three different vaccines based on the 140- amino acid MUC1 peptide with seven tandemly repeated tumor-specific epitopes ‘elicit distinct immune effector mechanisms in wild-type versus MUC1-transgenic mice with different potential for tumor rejection. J Immunol. 166(11):6555-63.

Springer GF. (1984) T and Tn, general carcinoma autoantigens. Science. 224(4654):1198-206.

Springer GF. (1997) Immunoreactive T and Tn epitopes in cancer diagnosis, prognosis, and immunotherapy. J Mol Med. 75(8):594-602.

Tang CK, Katsara M, Apostolopoulos V. (2008) Strategies used for MUC1 immunotherapy: human clinical studies. Expert Rev Vaccines. 7(7):963-75.

Tarp MA, Clausen H. (2008) Mucin- type O-glycosylation and its potential use in drug and vaccine development. Biochim Biophys Acta. 1780(3):546-63.

Tarp MA, Serensen AL, Mandel U, Paulsen H, Burchell J, Taylor-Papadimitriou J,

Clausen H. (2007) identification of a novel cancer-specific immunodominant glycopeptide epitope in the MUC1 tandem repeat. Glycobiclogy. 2007 Feb; 17(2):197- 209. :

Taylor-Papadimitriou J, Burchell J, Miles DW, Dalziel M. (1999) MUC1 and cancer

Biochim Biophys Acta. 1455(2-3):301-13.

Thornton DJ, Rousseau K, McGuckin MA. (2008) Structure and function of the polymeric mucins in airways mucus. Annu Rev Physiol. 70:459-86.

Van Patten SM, Hughes H, Huff MR, Piepenhagen PA, Waire J, Qiu H, Ganesa C,

Reczek D, Ward PV, Kuizko JP, Edmunds T. (2007) Effect of mannose chain length on targeting of glucocerebrosidase for enzyme replacement therapy of Gaucher disease. Glycobiology. 17(5):467-78.

Vlad AM, Finn OJ. (2004) Glycoprotein tumor antigens for immunotherapy of breast cancer. Breast Dis. 20:73-9. 20 .

Vlad AM, Muller S, Cudic M, Paulsen H, Otvos L Jr, Hanisch FG, Finn OJ. (2002)

Complex carbohydrates are not removed during processing of glycoproteins by dendritic cells: processing of tumor antigen MUC1 giycopeptides for presentation to major histocompatibility complex class ll-restricted T cells. J Exp Med. 196(11):1435- 46. von Mensdorff-Pouilly S, Petrakou E, Kenemans P, van Uffelen K, Verstraeten AA,

Snijdewint FG, van Kamp GJ, Schol DJ, Reis CA, Price MR, Livingston PO, Hilgers

J. (2000) Reactivity of natural and induced human antibodies fo MUC1 mucin with

MUCT peptides and n-acetylgalactosamine {(GalNAc) peptides. Int J Cancer. 86(5).702-12.

Wandall HH, Blixt Q, Tarp MA, Pedersen JW, Bennett EP, Mandel U, Ragupathi G,

Livingston PO, Hollingsworth MA, Taylor-Papadimifriou J, Burchell J, Clausen H. (2010) Cancer biomarkers defined by autoantibody signatures to aberrant O- glycopeptide epitopes. Cancer Res. 70(4):1306-13.

Xing PX, Apostolopoulos V, Pietersz G, McKenzie IF. (2001) Anti- mucin monoclonal antibodies. Front Biosci. 6:01284-95.

Zhang S, Graeber LA, Helling F, Ragupathi G, Adluri S, Lioyd KO, Livingston PO. (1996) Augmenting the immunogenicity of synthetic MUC1 peptide vaccines in mice.

Cancer Res. 56(14):3315-9.

Example 3

GDP-Fucose transporter mutant

It is now widely recognized that removal of the fucose from the N-glycans attached to

Asn297 of human IgG1 significantly enhances its binding to its receptor, Feyllla, and thereby dramatically improves antibody-dependent cellular cytotoxicity (ADCC) (12;13). Recent reports have shown that removal of sialic acid from IgG1 also enhanced ADCC in vivo (14,15). As disclosed above, we have generated a CHO mutant line that has a dysfunctional CMP-sialic acid transporter (16). As a resuli, glycoproteins produced by this cell line are completely free of sialic acid. The mutant line was originally named MAR-11 and has since been renamed as CHO-gmt1, for

CHO glycosylation mutant-1. We utilized the zinc finger nuclease technology and successfully knocked out the gene encoding GFT from the CHO-gmt1 cells. The resultant cell line, CHOgmt5, was shown to produce N-glycans essentially free of sialic acid and fucose. We ultimately aim to test whether an IgG1 lacking both fucose and sialic acid would perform even better in activating ADCC. However, in this study we have focused our efforts on making use of this mutant cell line to study the structurefunction relationship of GFT. The activity of GFT was indirectly reflected by the binding of a fucose-specific lectin, Aleuria Aurantia lectin (AAL), to the CHO-gmt5 cells. Using this system, we successfully.identified elements located in different : regions of GFT with significant impacts on its transport activity.

EXPERIMENTAL PROCEDURES

Materials — Monoclonal anti-hemagglutinin (HA) antibody (Clone HA-7), rabbit polycional anti- B4GalTI antibodies were purchased from Sigma (St. Louis, MO).

Rabbit polyclonal antibodies for giantin, GM130, mannosidase Il {Manll}, and TGN46 were purchased from Abcam (Cambridge, UK). Biotinylated AAL was purchased from

Vector Laboratories (Burlingame, CA). FITC-conjugated PNA was purchased from

EY Laboratories (San Mateo, CA). AlexaFluor dye-conjugated secondary probes, including goat anti-mouse IgG, goat anti-rabbit 1gG, and streptavidin, were purchased from Invitrogen (Eugene, OR). Cy3-conjugated streptavidin were purchased from

Jackson immunoResearch (West Grove, PA). Trypsin was purchased from Promega (Madison, WI). NGlycosidase F (PNGase F) was purchased from Calbiochem (San

Diego, CA). Hypercarb SPE cartridges (200 mg sorbent bed weight) were from

ThermoFisher Scientific (Waltham, MA). 2, 5-Dihydroxybenzoic acid (DHB) and Sep-

Pak Vac C18 cartridge were from Waters Corporation (Milford, MA). Acetic acid, ammonium bicarbonate, hydrochloric acid, methy iodide, sodium acetate and sodium : 10 hydroxide (NaOH) were all of analytical reagent grade from Merck (KGaA, Germany).

Acetonitrile, dimethyl sulfoxide and methanol were of HPLC grade from Merck.

Chloroform was of HPLC grade from Fisher Scientific (Pittsburgh, PA). Ultrapure water (Sartorius, Goettingen, Germany) was used throughout the analysis. "15 Cell culture and transfection of the ZFN constructs — CHO-gmt1 cells, formerly called

MAR-11 (16), were cultured in Dulbecco's Modified Eagle’s Medium supplemented with 10% fetal bovine serum (FBS) (both from Invitrogen, Auckland, CA), at 370C with 5% CO2. Prior to transfection, 5 x105 cells were seeded into each well of 6-well plates and cultured overnight. The constructs for expressing the pair of ZFNs were transiently transfected into CHO-gmt1 cells using Lipofectamine 2000 (Invitrogen,

USA) according to manufacturer's protocol. In each transfection, a total of 4 ug plasmid DNA (2 ug for each ZFN) and 10 pi Lipofectamine 2000 reagent in 500 pl serum-free DMEM were added to each well containing 2 ml of medium. After overnight incubation, the transfection reagent was replaced with normal culture medium and cultured for 2 days before single cell cloning into 96-cell plates.

Constructs for GFT variants — The human GFT (database accession number:

NM_018389) was synthesized by RT-PCR. Briefly, RNA was extracted from HEK293 cells using RNeasy Mini Kit (Qiagen, Hilden, Germany) according to manufacturers instructions. MRNA was reverse-transcribed into cDNA using ImProm-1i kit (Promega, Madison, WI). cDNA encoding GFT was amplified by PCR reaction with primer pair specific for 5- and 3'-ends of the coding region. The HA sequence was incorporated into the forward primer. The amplified GFT ¢DNA was digested with

Hindlll and Xhol (New England Biolabs, Ipswich, MA) and ligated into pcDNA3.1(+) vector. GFT variants with mutation in the C-terminal sequence (AC-tail, CST-tall,

ACT1, ACT2, EM/GG, 3K/3G and EM-CST-tail) were generated by standard PCR or overlap PCR. GFT variants with point mutations in the ACT2 deletion region or the transmembrane helical regions were generated using site-directed mutagenesis kit (Stratagene, Cedar Creek, TX) according to manufacturer's instructions. All constructs contain a Kozak sequence (GCCACC) ahead of the start codon and were cloned into pcDNA3.1 (+) vector using Hindlll and Xhol sites. Coding regions of all constructs were sequenced to rule out unwanted mutations.

Designing the ZFNs to target the GDPfucose transporter in CHO celfs — The “modular assembly” method was used to generate specific ZFNs for targeting the

GFT gene in CHO cells. The open reading frame of the Chinese hamster GDP- fucose transporter was analyzed using the web-based ZiFiT program provided by the

Zinc Finger Consortium (ZiFiT: software for engineering zinc finger proteins (V3.0) at: http://bindr.gdcb.iastate.edu/ZiFiT/ (17). The ZiFiT output located a DNA sequence in the first exon of the GDP-fucose transporter coding region (5'- tAACCTCTGCCTCAAGTACGTAGGGGTGG CCt-3) as a potential target site for

ZFNs (Fig. 5A). Two binding sites for ZFNs (underlined) are separated by 6 bps which is the optimal distance for cleavage by two ZFNs. This sequence perfectly matches the ideal target sequence for 4-fingered ZFNs which is: 5’-NNCNNCNNCNNCXXXXAXXGNNGNNGNN GNN-3'.

It allows each zinc finger in the left ZFN and the right ZFN fo bind to a 5-GNN-3’

DNA triplet (Fig. 5A). The fingers that bind the 5-GNN-3' triplets are the best studied and strong DNA-binding fingers (18-20). The structural scaffold for the ZFNs was adopted from previous publications (21;22). To eliminate unwanted homodimerization of Fokl cleavage domain, the high-fidelity Fokl-KK and FokI-EL variants were used (23). Each zinc finger in our ZFNs was designed based on the publicly available information described in the literature. The complete design of the left and right finger : is shown in Fig. 5B.

Isolation and characterization of CHO-gmt5 cells, CHO-gmt1 cells with inactivated

GDPfucose transporter gene — Two days after transfection single cells were seeded in 96-well plates for clone isolation. Genomic DNA from each single clone was isolated and the GDPfucose transporter target locus was amplified by PCR. The sequence of the forward PCR primer is 5-

GGCGCCTCTGAAGCGGTCCAGGATCC. The sequence for the reverse PCR primer is 5'- GCCACATGTGAGCAGGGCATAGAAGG. A 520-bp PCR product was generated from the wild type CHO genomic sequence which can be digested by a restriction enzyme SnaBl (TAC|GTA) around the ZFN restriction site (Fig. 5A) fo give rise to a smaller fragment of 114-bp and a larger fragment of 406-bp. In case a mutation has occurred at the targeted site by the ZFNs, the PCR product may become resistant to the restriction enzyme. However, for insertion mutants, the PCR products have to be sequenced in order fo identify the mutations.

Analysis of the oligosaccharides released from recombinant EPO-FC using matrixassisted laser desorption/ionization time-offlight mass spectrometry (MALDI-

TOF) — The recombinant EPO-Fc was produced in CHOgmt1 cells and CHO-gmt5 cells as described earlier (24) for N-glycan structure analyses. Briefly, the Fc region oo from human 1gG1 was fused to the C-terminus of EPO by overlap PCR. The PCR product that encodes the EPO-Fc fusion was cloned info pcDNA3.1 with a Kozak sequence placed upstream of the translation start codon ATG. The EPO-Fe construct was transiently transfected into CHO-gmt1 and CHO-gmi5 cells. Recombinant EPO-

Fe produced was purified with a Protein A column. An amount of 100 pg of purified

EPO-Fc was used for carbohydrate structure analysis. The carbohydrates liberated from the EPO-Fc by PNGase F were analyzed by matrix assisted laser desorptionfionization time-of-flight mass spectrometry (MALDI-TOF) analysis as described previously (16). Immunofiuorescence microscopy — CHOgmtb cells were piated on glass coverslips in 6- well plates and transfected next day using FuGene 6 reagent (Roche, Indianapolis, IN} with plasmid constructs for different GFT variants.

Two days after transfection, cells were washed with PBS, fixed with 4% paraformaldehyde (PFA) (diluted from 16% methanol-free PFA from Thermo

Scientific (Rockford, IL)) in PBS for 10 min and permeabilized by 0.1% Triton X- 100 (Sigma, St. Louis, MO) in PBS for 5 min. After washing with PBS twice, cells were blocked with Carbo-Free blocking solution (Vector Laboratories, Burlingame, CA), briefly rinsed with PBS and then incubated with PBS containing 1% BSA and a monoclonal anti-HA antibody (5 g/ml final concentration) as well as a Golgi marker antibody at an appropriate dilution for 1 hour. For experiments involving the analysis of GFT activity, biotinylated AAL (5 pg/ml final concentration) was included in the primary antibody mixture. The coverslips were then washed three fimes in PBS and incubated with secondary antibodies, i.e. AlexaFluor 488 conjugated goat anti-mouse 9G (for the detection of HA-GFT) and AlexaFluor 594 conjugated goat anti-rabbit

IgG (for the detection of Golgi marker), as well as AlexaFiuor 647 conjugated- streptavidin for experiments involving biotinylated AAL. in all experiments, Hoechst

33342 (Invitrogen, Eugene, OR) at 2 pg/ml (final concentration) was included in the secondary antibody mixture to stain the nuclei. Finally, the coverslips were washed four times in PBS and then mounted onto glass slides using ProlongGold antifade medium (Invitrogen, Eugene, OR). Fluorescence images were taken using a Carl -

Zeiss LSM 510 META confocal microscope with a 63X PLLANapochromat objective (1.40 NA) immersed in oil. All AAL images within the same set of experiments were acquired under constant imaging parameters. Representative areas were cropped with LSM Image Brower and used to produce the final figures. For localization analysis of GFT, images were taken at 10 successive optical slices in the Z-direction (0.41 um thickness per slice) and used to produce the final figure.

FACS analysis of cell surface fucosylation

CHO-gmt5 cells were transfected with different GFT constructs in 12-well plates using the method described above. Two days after transfection, cells were washed with PBS and harvested using PBS containing 2 mM EDTA. After washing with PBS, cells were fixed with 4% PFA (diluted from 16% methanol-free PFA. Thermo

Scientific) in PBS for 10 min. After washing with PBS twice, cells were incubated with

Carbo-Free blocking solution for 30 min, and then incubated with PBS containing 1%

BSA and 5 pg/mi biotinylated AAL for 30 min. After 2 washes with PBS, cells were incubated in PBS containing 1% BSA and 5 pg/ml Cy3- conjugated streptavidin for 30 min. Finally, cells were washed twice in PBS and analyzed by a BD LSR II FACS analyzer. For each run, 10,000 events were recorded. AAL-positive region was gated based on unfransfected CHO-gmt5 cells and cells transfected with a full-length GFT construct. Under such gating parameters, CHOgmt5 cells transfected with a T306R construct had <0.5% population in the AAL-positive region. To control for transfection efficiency, a separate set of transfections were done with each GFT construct mixed with a GFP construct at 10:1 ratio. Cells were stained by the same method described above and GFP fluorescence was analyzed by FACS. Similar GFP fluorescence profiles were obtained for all transfections, suggesting comparable transfection efficiencies for all constructs.

RESULTS

ZFNs generated by the combinatorial selection-based methods may have high

DNAbinding affinity and low toxicity. However, it requires large randomized libraries and the selection expertise that only a few labs possess. We engineered a pair of

ZFNs to target the GDP-fucose gene in CHO cells using a simplified “modular assembly” strategy. Each zinc finger in our ZFNs was designed based on the publicly available information described in the literature. We have demonstrated that this method can be a viable approach for researchers who do not have the expertise to perform a combinatorial selection.

Design of ZFNs to target the GDP-fucose transporter in CHO cells — The open reading frame of the Chinese hamster GDP-fucose transporier was analyzed using the web-based ZiFiT program provided by the Zinc Finger Consortium (ZiFiT: software for engineering zinc finger proteins (V3.0)) at: hitp://bindr.gdcb. iastate.edu/ZiFiT/ (17,25). One potential target site in the first exon of the GDPfucose transporter coding region (5'- tAACCTCTGCCTCAAGTACGTAGGGGTGG CCt-3°) was identified (Fig. 5A). Two potential binding sites for ZFNs (underlined) are separated by 6 bps which is the optimal distance for cleavage by two ZFNs. This sequence perfectly matches the ideal target sequence for 4-fingered ZFNs which is: 5'-

NNCNNCNNCNNCXXXXXXGNNGNNGNN GNN-3'. It allows each zine finger motif in the left ZFN and the right ZFN to bind to a 5-GNN- 3’ DNA triplets (Fig. 5A) as it has been shown that the 5-GNN-3’ triplets generally show high affinities for proper zinc fingers (18-20). The structural scaffold for the ZFNs was adopted from previous publications (21,22). To eliminate unwanted homodimerization of Fokl cleavage domain, the high-fidelity Foki-KK and Fokl-EL variants were used (23). The DNA- binding domains of the two ZFNs against this site were assembled using an archive of engineered zincfinger motifs collected from previous publications, mainly by

Barbas’s group and Sangamo BioSciences Inc.. The complete design of the left and right finger is shown in Fig. 5B.

Generation and identification of GDPfucose transporter-deficient CHO cells —

CHOgmt1 cells were transfected with the vectors that express the ZFNs to bind the

ZFN-L and ZFN-R shown in Fig.5A. Two days after transfection single cells were seeded in 96-well plates for clone isolation. Genomic DNA from each single clone was isolated and the GDP-fucose transporter target locus was amplified by specific

PCR primers. A 520-bp PCR product was generated from the wild-type CHO genomic sequence which can be digested by a restriction enzyme SnaBl (TAC|GTA) around the ZFN restriction site {Fig. 5A) to give rise to a smaller fragment of 114-bp and a larger fragment of 406-bp. In the event that a deletion mutation has occurred at the targeted site by the ZFNs, the PCR product may become resistant to the restriction enzyme. However, for insertion mutants, the PCR products have to be sequenced in order to identify the mutations. DNA sequencing results for the PCR products of the isolated clones showed several random deletion and insertion - mutations around the ZFN target site (Fig. 1C). C, D, E, F, G and M represent different single clones in which mutations have been identified. While clone D, F, G and M each has one mutated allele, they all still possessed one wild-type allele of the

GFT gene. As a result, these cells still express fucose-containing glycoproteins as suggested by the positive staining of a fucose-specific lectin, AAL (data not shown).

Two alleles of the GFT gene in the clone C were mutated, resulting in one deletion mutation (C1) and one insertion mutation (C2). Unfortunately, this clone had three alleles of the GFT gene and the third was not mutated. Therefore, clone C also possessed a wild type phenotype. Interestingly, the inserted DNA fragment shown in

C2 came from the expression vector used in this experiment. Clone E had two alleles of the GFT gene and both were mutated, each with a deletion mutation. E1 has a 3- bp deletion, resulting in mutation of K137-Y138 to one N residue. E2 has a 40-bp deletion which resulted in a prematurely truncated GFT product of 172 amino acids.

The PCR product of this clone was found to be resistant to the restriction enzyme

SnaBi because the recognition site (TAC|GTA) in both alleles had been abolished by the deletions. This clone, named CHO-gmt5, was used for further fucosylation analyses.

Endogenous and recombinant N-glycans produced by CHO-gmi5 cell are free of fucose — Lectin staining was performed to gain insights into the glycan terminal structures an CHOgmt5 cells (Fig. 6A). Parental and wild-type cell lines, CHO-gmt1 and CHO-K1, were also included in this analysis. Cells were labeled with two lectins,

PNA which recognizes terminal galactose residue, and AAL which recognizes fucose residue. CHO-K1 is negative for PNA binding due to the capping of galactose residues by sialic acids. It is positive for AAL binding because of the abundant fucosylation of Nglycans. in contrast, CHO-gmt1 is positive not only for AAL, but also for PNA due to the genetic defect in CST leading to production of galactose- terminated N-glycans. As expected, CHO-gmt5 was shown fo be positive for PNA.

However, it is negative for AAL binding, suggesting the lack of fucosylation due to knockout of the GFT gene (note the weak staining of CHO-gmt5 by AAL is the background, which is consistent with a previous publication) (26). Recombinant EPC- - 35 Fc fusion protein produced in CHO-gmt1 (MAR-11) and CHOgmt5 were purified and the N-linked glycans liberated by PNGase F were analyzed by MALDI-TOF as previously described (16) and the mass spectra are as shown in Fig. 6B. The glycans isolated from CHO-gmt1 were mostly fucosylated, and the dominant species were the asialo, core-fucosylated galactosyi biantennary, triantennary and tetraantennary glycans. in contrast, the dominant species isolated from CHO-gmt5 were the asialo, afucosylated galactosyl biantennary, triantennary and tetraantennary glycans, correspondingly. Minor peaks corresponding to fucosylated species were observed in

CHO-gmt5 sample at residual levels, but these peaks were also control in a control sample where CHO-gmt5 cells were not even transfected with an EPO-Fc construct (data not shown}, suggesting these residual amounts of fucosylated glycans might come from the culture medium. These data clearly demonstrated that glycoproteins expressed on the cell surface of, or secreted by, CHO-gmt5 cells are completely free of fucose due to a dysfunctional GTF gene.

The human GDP-Fucose transporter is localized to medial- and trans-Golgi — The C. elegans and human GFT have been shown to be localized to the Golgi apparatus (9).

To ascertain the distribution of GFT in the Golgi, we transiently expressed HA-tagged

GFT in Hela cells and analyzed the distribution of GFT by immunofluorescence microscopy and compared its localization pattern with defined Golgi cisternal markers. Cells expressing moderate level of GFT were analyzed and representative images were shown in Fig. 7. GFT did not show co-localization with cis-Golgi marker

GM130, or with TGN marker TGN46. Partial colocalization was detected with medial

Golgi marker Manll and high degree of co-localization was seen with frans-Golgi marker B4GalTl. This indicates that at steady state, the majority of GFT protein is localized to the trans-Golgi with a smaller fraction localized to medial-Goigi. This observation is consistent with the distribution pattern of CST in Hela cells (27).

The cytosolic C-terminal tail sequence of GFT is critical for its transport activity but not localization fo the Golgi — The cytosolic C terminal tail sequence of CST has been shown fo be involved in the localization and trafficking of this transporter (27). To determine if this feature is conserved in GFT, we made a series of constructs encoding HA-tagged GFT variants with deletions in the C-terminal tail region (Fig. 8A}. Constructs were transfected into CHOgmt5 cells. Two days after transfection, cells were fixed and stained with anti-HA antibody to determine the localization of the

GFT variants. Cells were counter-stained by biotinylated AAL followed by AlexaFluor 647-conjugated streptavidin as an indirect readout for GFT activity. Multiplex fluorescent images were taken at constant imaging conditions and shown in Fig. 8B.

Expression of a full-length GFT (FL) restored the fucosylation in transfected

CHOgmt5 cells as evident by the strong AAL fluorescence intensity, both in the Golgi and membrane regions. Deletion of aa. 341 — 364 (AC-tail construct) did not change the Golgi focalization of GFT. However, AAL fluorescence intensity, especially that outside Golgi, was greatly reduced. The diminished rescue of fucosylation suggests this sequence is critical for GFT transport activity. Replacing this sequence with the corresponding tail sequence CST (CST-tail) did not restore the fucosylation, suggesting that, unlike CST, the Cterminal tail of GFT is essential for its activity. To narrow down the critical amino acid residue(s), we made two shorter deletions within the GFT C-tail region, aa 356 — 364 (ACT1) and aa 341 — 355 (ACT?2). Both variants exhibited simitar Golgi-localization pattern but striking difference in AAL fluorescence intensity. Celis expressing ACT1 showed AAL intensity comparable to FL-transfected cells, suggesting aa. 356 — 364 are dispensable for GFT transport activity. However, cells expressing ACT2 were negative for AAL binding, except for a basal level of fluorescence in the Golgi area. This result indicates that aa. 341 - 355 are critical for the transport activity of GFT. in addition to analyzing the lectin-stained cells under the microscope, we also employed FACS to examine the cell surface fucosylation to quantitatively compare the relative activities of different GFT variants (Fig. 8C and 4D). Constructs encoding GFT variants were transfected into CHO-gmt5 cells. Two days after transfection, cells were fixed but not permeabilized to allow specific tabeling of cell surface fucosyl residues by AAL. GFT consiruct bearing a T308R, a point mutation previously reported in a CDG-lIc patient (8), was included as a negative control. Cells transfected with the FL construct typically show 15 — 20% in the AAL-positive region (Fig. 8C). This percentage was used to define 100% transport activity (Fig. 8D). Under such conditions, the activity of GFT T308R is close to zero. Consistent with the lectin staining results, AC-tail and CST-tail GFT variants failed to restore the cell surface fucosylation. ACT1 showed an activity comparable to

FL, whereas ACT2 mutant completely lost its activity. 344EM345 and a cluster of three Lys residues in the C-terminal tail of GFT have significant impacts on the activity of GFT — To pinpoint the exact amino acid residue critical for GFT activity in the ACT2 deletion region, a series of mutant GFT constructs were generated by replacing every amino acid in this stretch with a Gly residue. These constructs were then expressed in CHO-gmt5 cells and analyzed by immunofluorescence microscopy for their cefiular localization and AAL-staining for

GFT activity. All of these GFT mutants were localized to Golgi and shown to be fully functional as evident by strong AAL fluorescence intensity. Therefore, the impact of

ACT?2 could be contributed by a combination of amino acids instead of any individual residue within this region. Yoda and co-workers identified a series of Lys residues in the C-tail region of yeast GDPmannose transporter (GMT) involved in ER retrieval of this protein by binding to COP! coatomer (28). They postulated that NSTs with a cluster of at least three Lys residues near the membrane spanning region could utilize these Lys for ER-Golgi trafficking. GFT was one of such NSTs on their list.

Interestingly, GFT contains four Lys residues, three of which are located in the ACT2 deletion region. These three Lys were replaced by three Gly residues (3K/3G construct illustrated in Fig. 9A) to assess their involvement in GFT localization and activity. Immunofiuorescence results indicated a dispensable role of these Lys in

GFT localization (Fig. 9B). However, AAL binding was reduced, especially outside the Golgi area. Consistently, FACS analysis of cell surface fucosylation showed the activity of GFT with triple LysOGly mutations was reduced by approximately 50% (Fig. 9C and 9D). Sequence alignment analysis of the GFT proteins identified in all the species in the database shows the conservation of 344EM345 motif (data not shown). To test if this conserved EM motif is involved in the bioactivity of GFT, we made an EM/GG construct by replacing these two conserved amino acids with two

Gly residues (Fig. 9A). The EM/GG construct was expressed in CHO-gmt5 cells and found to be localized to Golgi (Fig. 9B). Activity of this variant was analyzed by lectin staining and FACS. Cells expressing EM/GG variant showed AAL fluorescence mainly in the Golgi area (Fig. 9B). Consistently, the cell surface fucosylation was found to be greatly reduced as compared to cells transfected with the FL construct (Fig. 9C and 9D). To test if the EM and Lys cluster are two inter-dependent elements for GFT activity, we made an additional mutant construct by replacing the C-terminal sequence of GFT downstream of the 344EMK346 motif with the corresponding region from the CST (Fig. 9A, shown as EM-CST-tail) and analyzed its localization and bioactivity in CHO-gmt5 cells. This GFT variant maintained its localization to the

Golgi and did not show any compromised activity in AAL binding (Fig. 9B —D).

Comparing the activity of mutant EM-CST-tail (Fig. 9) and that of mutant CST-tail (Fig. 8) also suggests an important role for the EM motif. Therefore, the C-terminal

Lys cluster and EM sequence are two independent elements required for the activity of GFT.

Three conserved Gly residues in the transmembrane helices have differential impacts on the activity of GFT — Previously, we identified four pairs of Gly in the transmembrane helices of CST critical for its transport activity (16). Multiple sequence alignment of GFT, CST and UGT proteins in human and Chinese hamster show that the first Gly in the first, third and fourth pairs are conserved (Fig. 10A and data not shown). To ascertain the involvement of these three conserved helical Gly residues in the transport activity of GFT, we mutated individual Gly to lie or Tyr.

Mutating any of the Gly to a bulky amino acid Ile or Tyr did not alter the Golgi localization of GFT. However, these mutants displayed dramatic difference in rendering AALbinding to the transfected cells (Fig. 10B): G180! mutation resulted in substantial decrease of AAL fluorescence intensity, whereas G180Y almost completely abolished AAL fluorescence outside Golgi; G198I and G198Y did not affect the AAL binding, whereas G277| and G277Y mutations render the GFT nonfunctional as evidenced by the loss of AAL fluorescence even in the Golgi area.

The results based on lectin-staining were verified by FACS (Fig. 10C and 10D).

Consistently, G180i mutation impaired the transport activity of GFT whereas a more bulky amino acid, Tyr, at this position almost completely dimished the activity of GFT,

Gly198 is dispensable from GFT activity as substitution with lie or Tyr did not make a significant change in cell surface fucosylation. Gly277 was most sensitive to mutation as a substitution with lle completely abolishes cell surface fucosylation.

DISCUSSION

NST proteins typically contain 6 — 10 membrane-spanning helices with both the amino- and carboxyl-termini facing the cytosol. The role of C-terminal! tail sequence in the trafficking and localization of NST proteins is well established in the context of mouse CMPsialic acid transporter (27), a larger splice variant of human UDP-fucose iransporter (29), as well as yeast GDP-mannose transporter (28). Several localization or trafficking-related motifs have been identified in the C-terminal tail region. These include ER-export motif (// and GV in the context of mouse CST), ER-retention signal {KXK motif in the context of human UGT2), and ER-retrieval sequence (Lys cluster in the Cterminal tail in the context of yeast GMT). The human GFT also ends with a GV sequence at its C-terminus. However, this GV sequence, even in combination with upstream C-terminal tail sequence, does not function as an ER export signal in the context of GFT as deletion of the C-terminal does not change the localization of GFT in CHO-gmt5 (Fig. 8B) and Hela cells (data not shown). Surprisingly, the C-tail deleted, Golgi-localized GFT mutant almost completely loses its transport activity. By contrast, the Cterminal of CST, although critical for its localization, was shown to be partially or completely dispensable for CST activity by lectin blotting and isoelectric focusing analysis of the recombinant product (16:27). In addition, biological function of human UGT was also shown to be independent of the C-terminal tail (4).

The cluster of Lys in the C-terminal tail region of yeast GMT was shown to interact with COPI coatomer and prevents GMT from accumulating in the non-Golgi vacuolar structures (28). However, we have shown that disruption of the conserved Lys cluster in human GFT did not alter its localization to the Golgi (Fig. 9B). But the fact that mutating all three Lys to Gly significantly compromised the GFT transport activity suggests these residues could be involved in other interactions, which among many possibilities may include the binding or recognition of nucleotide-sugar or other proteins.

Our results clearly demonstrated that despite the substantial sequence similarity ~ between different transporters, functions of the less conserved tail sequences have i5 diverged during evolution. impacts of the tail sequences on trafficking and function of transporters have to be analyzed in a context-dependent manner. On the other hand, although it is practically impossible to deduce the specificity of NST based on its primary sequence, residues that are conserved between transporters or species can provide hints for elements that are critical for the transport activity of NST. Here we showed the C-terminal conserved EM residues have dramatic impact on the activity : of GFT. Again, the potential interacting partners of the conserved sequence remain to be elucidated. oo

Membrane-spanning helices have also been shown to have dramatic impact on the activity and specificity of NST. Domain-swap experiments have clearly shown that different transmembrane helices are utilized for CST and UGT (5:6). lt was also noted that some areas of the helices are rich in Gly, which are largely conserved in many NST proteins by multiple sequence alignment (data not shown). One plausible hypothesis is that these Gly residues, due to the lack of side chain, contribute to the formation of transmembrane channels to allow for the passage of nucleotide-sugars.

This idea was tested in our previous work with CST (16). We showed that mutation of four pairs of Gly to Ala or lle dramatically impair the transport activity of CST. Our results from GFT further supported this notion. We noted that the first Gly residues in the first, third and four Gly pairs are conserved in GFT. These Gly residues have different impacts on the activity of GFT: the Gly located in helix 5 has partial effect,

Gly located in helix 6 is dispensable, whereas Gly located in helix 9 is essential (Fig.

6). interestingly, the Gly pair in helix 9 of CST was shown to have the strongest impact on CST activity (16). Therefore, the conserved Gly residues in the transmembrane helices are generally required for the transport activity of NSTS instead of determinants of their specificities.

Recently, another Slc35 protein, ER localized Efr in Drosophila (30) and ERGIC/cis-

Golgilocalized mammalian Sle35c¢2 (31), was shown to have transport activity for

GDP-fucose and regulate the O-fucosylation of Notch. Overexpression of Sic35¢2 was shown to compete with the canonical GFT and have an inhibitory effect on the formation of Lewis-X structure but imposes a marginal effect on core fucosylation.of

N-glycans (31). We analyzed the localization of GFT in Hela and found it to have * more co-localization with trans-Golgi marker and to a lesser extent with medial-Golgi marker (Fig. 7). Therefore, the spatial segregation of Sic35¢c2 and GFT explains why -Sle35¢2 regulates fucosylation of Notch, which takes place pre-Golgi, and does not affect core fucosylation of N-glycans. Thus, targeting GFT is necessary and sufficient for regulating core fucosylation of N-glycans at membrane transport level,

Absence of core fucose on Fc N-glycan has been shown to enhance the ADCC effect of I9G1 (12,13). CHO GFT-/- cell line has been generated by homologous recombination and used as a host cell line for the production of fucose-free recombinant antibody (32). Absence of sialic acid on Fc N-glycan is also favorable for

ADCC effect (14;15). It is tempting to speculate that the absence of both fucose and sialic acid will offer combinatorial enhancement fo the ADCC effect. Qur CHO-gmt5 with doubie genetic defects in GFT and CST is capable of producing recombinant glycoproteins free of fucose and sialic acid and thus, has the potential application for the production of recombinant antibodies with improved ADCC effect. In fact, our ongoing study has demonstrated that CHOgmt mutants (including CHO-gmt1 and 5) can be adapted to suspension culture in protein-free media with growth profiles similar to that of wild-type CHO-K1 cells. :

CHO glycosylation mutants with genetic defects in multiple NSTs (such as CST and

UGT) have been isolated and reported (7). Our CHOgmt5 is a new addition to this range of mutants and holds promise for further targeting of additional NSTs towards a mammalian cell line with minimal NST background for the analysis of NST activity and specificity. The abbreviations used are: AAL, Aleuria Aurantia lectin; CST, CMP- sialic acid transporter; FACS, fluorescence-activated cell sorting; GFT, GDP-fucose transporter; GMT, GDP-mannose transporter; NST, nucleotide-sugar transporter:

PNA, peanut agglutinin; UGT, UDP-galactose transporter; ZFN, zinc-finger nuclease.

References for Example 3 1. Hirschberg, C. B., Robbins, P. W., and Abeijon, C. (1998) Annu. Rev. Biochem. 67, 49-69 2. Jaeken, J. and Matthijs, G. (2007) Annu. Rev. Genomics Hum. Genet. 8, 261-278 3. Chan, K F., Zhang, P., and Song, Z. (2010) Glycobiology 20, 689-701 4. Aoki, K., Sun-Wada, G. H., Segawa, H., Yoshioka, S., Ishida, N., and Kawakita, M. (1999) J. Biochem. 126, 940-950 - 5. Aoki, K., Ishida, N., and Kawakita, M. (2001) J. Biol. Chem. 276, 21555-21561 6. Aoki, K., Ishida, N., and Kawakita, M. (2003) J. Biol. Chem. 278, 22887-22893 7. Patnaik, S. K. and Stanley, P. (2006) Methods Enzymol. 416, 159-182 8. Lubke, T., Marquardt, T., Etzioni, A., Hartmann, E., von Figura, K., and Korner, C. (2001) Naf Genet. 28, 73-76 9. Luhn, K., Wiid, M. K., Eckhardf, M., Gerardy-Schahn, R., and Vestweber, D. (2001) Nat Genet. 28, 69-72 10. Martinez-Duncker, |., Dupre, T., Piller, V., Piller, F_, Candelier, J. J., Trichet, C.,

Tchernia, G., Oriol, R., and Moliicone, R. (2005) Blood 105, 2671-2676 11. Eckhardt, M., Gotza, B., and Gerardy-Schahn, R. (1999) J. Biol. Chem. 274, . 8779-8787 12. Shields, R. L., Lai, J., Keck, R., O'Connell, L. Y., Hong, K., Meng, Y. G., Weikert,

S. H., and Presta, L. G. (2002) J. Biol. Chem. 277, 26733-26740 13. Shinkawa, T., Nakamura, K., Yamane, N., Shoji-Hosaka, E., Kanda, Y.,

Sakurada, M., Uchida, K., Anazawa, H., Satoh, M., Yamasaki, M., Hanai, N., and

Shitara, K. (2003) J. Biol. Chem. 278, 3466- 3473 14. Kaneko, Y., Nimmerjahn, F., and Raveich, J. V. (2006) Science 313, 670-673 : 15. Anthony, R. M., Nimmerjahn, F., Ashline, D. J., Reinhold, V. N., Paulson, J. C., and Ravetch, J. V. (2008) Science 320, 373-376 16. Lim, S. F., Lee, M. M,, Zhang, P., and Song, Z. (2008) Glycobiology 18, 851-860 17. Sander, J. D., Zaback, P., Joung, J. K., Voytas, D. F., and Dobbs, D. (2007)

Nucleic Acids Res. 35, W599-WB05 18. Segal, D. J., Dreier, B., Beerli, R. R., and Barbas, C. F., Ill (1999) Proc. Natl.

Acad. Sci. U.S.A 86, 2758-2763 19. Dreier, B., Segal, D. J, and Barbas, C. F., lll (2000) J. Mol. Biol. 303, 489-502 20. Liu, Q., Xia, Z., Zhong, X., and Case, C. C. (2002) J. Biol. Chem. 277, 3850-3856

21. Umov, F. D., Miller, J.C, Lee, Y. L,, Beausejour, C. M., Rock, J. M., Augustus,

S., Jamieson, A.

C., Porteus, M. H., Gregory, P. D., and Holmes, M. C. (2005) Nature 435, 646-651 22. Doyon, Y., McCammon, J. M., Miller, J. C., Faraji, F., Ngo, C., Katibah, G. E.,

Amora, R., Hocking, T. D., Zhang, L., Rebar, E. J., Gregory, P. D., Umov, F. D., and

Amacher, S. L. (2008) Nat Biotechnol. 26, 702-708 23. Milter, J. C., Holmes, M. C., Wang, J., Guschin, D. Y., Lee, Y. L., Rupniewski, |,

Beausejour, C M., Waite, A. J., Wang, N. S., Kim, K. A., Gregory, P. D., Pabo, C. O., and Rebar, E. J. (2007) Nat Biotechnol. 25, 778-785 24. Goh, J. S,, Zhang, P., Chan, K. F., Lee, M. M., Lim, S. F., and Song, Z. (2010)

Metab Eng 12, 360- 368 25. Sander, J. D., Maeder, M. L., Reyon, D., Voytas, D. F., Joung, J. K., and Dobbs,

D. (2010) Nucleic Acids Res. 38, W462-W468 26. Matsumura, K., Higashida, K., Ishida, H., Hata, Y., Yamamoto, K_, Shigeta, M.,

Mizuno-Horikawa, Y., Wang, X., Miyoshi, E., Gu, J., and Taniguchi, N. (2007) J. Biol.

Chem. 282, 15700-15708 27. Zhao, W., Chen, T. L., Vertel, B. M., and Colley, K. J. (2006) J. Biol. Chem. 281, 31106-31118 28. Abe, M., Noda, Y., Adachi, H., and Yoda, K. (2004) J. Cell Sci. 117, 5687-5696 © 28. Kabuss, R., Ashikov, A, Oelmann, S., Gerardy-Schahn, R., and Bakker, H. (2005) Glycobiology 15, 905-911 30. Ishikawa, H. O., Ayukawa, T., Nakayama, M., Higashi, S., Kamiyama, S.,

Nishihara, S., Aoki, K., Ishida, N., Sanai, Y., and Matsuno, K. (2010) J. Biol. Chem. 285, 4122-4129 31. Ly, L,, Hou, X,, Shi, S., Korner, C., and Stanley, P. (2010) J. Biol. Chem. 285, 36245-36254 32. Ishiguro, T., Kawai, S., Habu, K., Sugimoto, M., Shiraiwa, H., lijima, 8., Ozaki, S.,

Matsumoto, T.,and Yamada-Okabe, H. (2010) Cancer Sci. 101, 2227-2233

Claims (23)

1. Claims: . 1. A glycoprotein that is expressed by a mammalian cell, wherein the mammalian cell ~ comprises: a) a mutated GnT | gene and a mutated CST gene; b) a mutated GnT | gene and a mutated UGT gene; or c) a mutated CST gene; or d) a mutated UGT gene; or e) a mutated GFT gene and a mutated CST gene.
2. 2. The glycoprotein according to claim 1 which is a recombinant glycoprotein.
3. 3. The glycoprotein according to claim 1 or claim 2 wherein the mammalian cell is a CHO cell, BHK cell, Vero cell, 293 cell, NSO cell, 3T3 cell, COS-7 cell or PER C6 cell.
4. 4. The glycoprotein according to any one of claims 1, 2 or 3 which is MUC1, or a fragment thereof.
5. 5. The glycoprotein according to claim 4 which comprises the N terminal subunit of MUCH.
6. 6. The glycoprotein according to claim 5 which comprises 1-100 tandem repeats.
7. 7. The glycoprotein according to any one of the preceding claims which has an altered pattern of glycosylation.
8. 8. The glycoprotein according to claim 7 which comprises a mannose {erminated glycan structure.
9. 9. The glycoprotein according to claim 7 or 8 which comprises T, Tn or Stn antigen glycans.
10. 10. An antibody raised against the glycoprotein of any one of the preceding claims.
11. 11. An antibody expressed by a mammalian cell according to any one of claims 1 to
12. 12. The antibody of claim 11 wherein the mammalian cell has a mutated GFT gene and a mutated CST gene.
13. 13. The glycoprotein or antibody according to any preceding claim for use in medicine.
14. 14. The glycoprotein or antibody according to claim 13 for use in the treatment of cancer.
15. 15. A method of treatment comprising administering the glycoprotein or antibody according to any one of claims 1 to 10 to a patient in need thereof.
16. 16. The glycoprotein or antibody for use according io claims 14 or 15 wherein the freatment comprises exposing cells obtained from a patient in need of treatment to the glycoprotein according to any one of the preceding claims.
17. 17. A glycoprotein or antibody according to any one of claims 1 to 10 for use in the manufacture of a medicament for the treatment of cancer. :
18. 18. A mammalian cell which comprises: a) a mutated GnT | gene and a mutated CST gene; b) a mutated GnT | gene and a mutated UGT gene; or ©) a mutated UGT gene; or : d) a mutated GFT gene and a mutated CST gene.
19. 19. A mammalian cell according to claim 16 which is a CHO cell, BHK cell, Vero cell, 293 cell, NSO cell, 3T3 cell, COS-7 cell or PER C6 cell.
20. 20. The mammalian cell according to claim 16 or claim 17 which expresses a recombinant protein.
21. 21. A culture comprising mammalian cells according to claims 18, 19 or 20.
22. 22. A method for expressing a recombinant glycoprotein, comprising expressing the glycoprotein in a mammalian cell according to any one of claims 18 to 21.

    22. The method according to claim 22 wherein the glycoprotein is encoded by the genome of the mammalian cell. }
23. 23. The method according to claim 22 wherein the glycoprotein is encoded by a vector.